WhiteFiber WYFI NASDAQ

WhiteFiber's financing record covers about fourteen months and shows a company that moved from an equity IPO to repeated convertible issuance. The IPO priced at $17 (claim 4558). It was upsized to 9,375,000 shares (claim 25619), against an earlier 7,812,500-share plan (claim 4655), and reached $183.3M gross after the over-allotment (claim 24138). Bit Digital retained roughly 74% (claim 4535), so the public float started narrow. Since then, new capital has come mainly through convertibles. The January notes carried a $26 conversion price (claim 3885). The August notes, upsized to $270M (claim 7771), convert at $34 (claim 7489). The higher strike limits per-dollar dilution, but the August exchange adds 6.3M shares directly (claim 7770), and that issuance should be counted alongside the conversion overhang.

The mix looks more structured than opportunistic, though with a recurring pattern. Of the January notes' $222M net proceeds (claim 3887), $120M went into a zero-strike call option (claim 3888). This means more than half of that raise was used to manage share count rather than fund assets. Seven months later, the August offering is retiring $198.15M of those same 2031 notes for $118.5M cash plus shares (claim 7770). Refinancing a convertible this quickly requires explanation beyond the improved conversion terms. At the asset level, the RBC facility of $43.8M (claim 2166) and the $100M delayed-draw bridge (claim 8986) fit a project-finance strategy for building sites.

The timing of the NC-1 project financing is the clearest gap between announcement and execution. It was expected to close in early Q1 2026 (claim 18332). It then appeared as a term sheet in May (claim 1314) and had reached lender exclusivity by mid-August (claim 8602). That makes it delayed by about two quarters, and the bridge facility appears to fill the gap. The record does not show how proceeds were actually deployed. The stated use for the August notes (claim 182) is not detailed, so tracking cannot be confirmed. The record also contains internal inconsistencies, including January closing figures of $230M (claim 17581) versus $250M (claim 3883), and a greenshoe of $37.5M (claim 7829) versus $40M (claim 7491). No insider sales are on record. The only insider activity is $72,500 of officer IPO purchases (claim 4591).

The next items to watch are the final NC-1 financing terms and closing, the size of any draw on the bridge, whether the August greenshoe is exercised, and Q2's net loss of $14.98M (claim 201) against the contracted $865M Nscale revenue base (claim 3974).

WhiteFiber's record is short but busy. In roughly thirteen months the company moved from an upsized IPO to two convertible note offerings, a bank facility, and a bridge loan. Equity dilution has so far been front-loaded. The IPO priced at $17 (claim #4558) for $159.4M gross (claim #25620), and the fully exercised over-allotment raised that to $183.3M (claim #24138), leaving Bit Digital with roughly 74% (claim #4535). After that the company turned to convertibles. The first, $230M of 4.5% notes due 2031 at a $26 conversion price (claims #17581, #3885, #3884), routed $120M of its $222M net proceeds into a zero-strike call option (claims #3887, #3888). That structure effectively prepurchased shares and muted near-term dilution at the cost of cash on hand. Seven months later the company priced an upsized $270M convertible due 2032 (claim #7771) with a higher $34 conversion price (claim #1724). Part of it funds an exchange retiring $198.15M of the 2031 notes for $118.5M cash plus 6.3M shares (claim #7770). That exchange crystallises some of the dilution the call option had deferred, and it means the first instrument was largely unwound well before its first redemption date of February 2029 (claim #3886). The rising conversion price reads as constructive. The quick refinancing of a recent instrument requires explanation. The debt layer shows a steadily slipping timeline. A credit facility expected to close in early Q1 2026 (claim #18332) is marked delayed, and NC-1 project-level financing was delayed in January (claim #17576). It reached a term sheet in May (claim #1314) and lender exclusivity in August (claim #8602), but the record does not show it closed. In the interim, a $100M delayed-draw bridge expandable to $150M (claims #8986, #8987) and the smaller RBC facility of about $43.8M (claim #2166) have filled the gap. The mix therefore looks less like a pre-set plan than a sequence of bridges built around a permanent financing that has not yet arrived. The data does not let us test use of proceeds against deployment. A planned use for the 2032 notes is disclosed (claim #182), but its content and any subsequent spend are not visible, and the record is truncated. No insider sales or ATM activity appear. Readers should watch for the NC-1 permanent financing closing, the drawn balance on the bridge, and the post-exchange share count.

WhiteFiber's record describes a company that moved from equity to convertible debt within six months of listing, then turned to recycling its first convertible before the second year was out. The IPO priced at $17 (claim 4558). It raised $159.4M gross (claim 4562) and reached $183.3M once the over-allotment was exercised (claim 24138). Bit Digital retained a stake recorded as 74 (claim 4535), so the public float began narrow. Dilution since then has come mostly through conversion features rather than primary equity. The January convertible was announced at $210M (claim 17586) and closed at $230M (claim 17581), at 4.5% with a $26 conversion price (claims 3884, 3885). In August, a $250M proposal (claim 2718) was upsized to $270M (claim 7771) with a $34 conversion price (claim 1724). That deal was paired with an exchange of $198.15M of the 2031 notes for $118.5M in cash and 6.3M shares (claim 7770). The higher strike limits per-dollar dilution, but the exchange converts part of the earlier debt into shares now rather than at maturity. The mix reads as broadly coherent rather than improvised. Corporate-level convertibles funded the equity layer, while asset-level debt was added at the Montreal sites. The RBC facility of $43.8M (claim 2166) split into an equipment lease and an MTL-2 term loan (claims 2167, 2168). The less settled element is NC-1. A credit facility expected in early Q1 2026 is marked delayed (claim 18332), and project-level financing was described as delayed in January and again in May (claims 17576, 9070). In the interim, a $100M delayed-draw bridge (claim 8986) expandable to $150M (claim 8987) filled the gap. Lender exclusivity was reported in August (claim 8602), but no close is on record. The $120M of January proceeds directed to a zero-strike call (claim 3888) also requires explanation alongside a stated $150M equity commitment to NC-1 (claim 18331). Use-of-proceeds detail for the August deal is referenced (claim 182) but not specified in the record, so tracking deployment against intent is not yet possible. Insider activity is limited to indicated officer IPO purchases of $72,500 (claim 4591), and no insider sales appear. A sophisticated reader should watch for the NC-1 permanent financing close, how the bridge is repaid, and whether the $865M Nscale contract (claim 3974) converts into revenue fast enough to narrow a Q2 net loss of $14.98M (claim 201). The record supplied here ends mid-entry, so later claims may modify this read.

WhiteFiber's first fourteen months as a public company show one equity raise followed by an accelerating reliance on convertible and secured debt. The August 2025 IPO priced at $17 (claim #4558) and reached $183.3M gross after the over-allotment was fully exercised (claim #24138). Bit Digital kept roughly 74% ownership (claim #4535), so the public float began narrow. Later restatements of $144.5M and $22.2M (claims #4123, #4124) appear to be net figures and should be reconciled against the gross total before anyone relies on them. The only insider activity on record is an officer's indicated $72,500 IPO purchase (claim #4591). No insider sales are disclosed.

The mix since then reads as a stack of bridges toward a single objective: permanent financing for NC-1. That facility was first expected in early Q1 2026 (claim #18332). It was then described as delayed project-level financing in January (claim #17576), reached a term sheet in May (claim #1314, with claim #9070 marking it delayed), and had reached only lender exclusivity by August (claim #8602). In the interim, the company issued $230M of 4.5% convertibles (claims #17581, #3884), added a $43.8M RBC facility tied to MTL-2 (claim #2166), and took a $100M delayed-draw bridge expandable to $150M at a $3M original issue discount (claims #1287, #8987, #1292).

Use of proceeds tracks only partly. The IPO funding broadly matches the $150M NC-1 equity commitment (claim #18331). However, $120M of the January convertible's $222M net proceeds went to a zero-strike call option rather than capex (claims #3887, #3888). The August 2026 offering, upsized from $250M to $270M (claims #7828, #174), allocates $118.5M of cash plus about 6.3M shares to retire $198.15M of the January notes (claim #7770). A meaningful share of the new capital therefore refinances the previous instrument within roughly seven months, while the conversion price moved from $26 to $34 (claims #3885, #178). Against this, Q2 2026 showed a $14.98M net loss (claim #201) on a reported -47% net margin (claim #6906).

A careful reader should watch four things. First, whether NC-1 permanent financing actually closes, and how much of the bridge is drawn before it does. Second, whether the August offering's closing, listed as delayed (claim #1721), completes as priced. Third, the cumulative dilution from the exchange shares on top of the narrow float. Fourth, whether project financing begins to displace corporate-level convertibles or simply sits alongside them. The record supplied here appears truncated at its final entries, so later developments are not reflected.

WhiteFiber's capital history over roughly fourteen months moves from equity to debt. The IPO priced at $17 (claim #4558) and was upsized from 7,812,500 shares (claim #4590) to 9,375,000 (claim #4559). With the fully exercised greenshoe (claim #4420), gross proceeds reached about $183 million (claim #19576), and Bit Digital retained roughly a 74% stake (claim #4564). Every raise since then has been debt or equity-linked debt. The January convertible was proposed at $210 million (claim #17586) and closed at $230 million (claim #17581). The August convertible was proposed at $250 million (claim #7828) and priced upsized at $270 million (claim #7486). Its conversion price is about $34 (claim #7489), compared with $26 on the earlier notes (claim #17589), so each new share the company commits to issue is priced higher. Near-term share issuance still comes mostly through the note exchange, which delivers 6.3 million shares alongside $118.5 million of cash for $198.15 million of the 2031 notes (claim #7770).

The mix is more patched than planned. $120 million of the January proceeds went to a zero-strike call (claim #3888), leaving about $102 million of the $222 million net (claim #3887) for operations. Seven months later, about $118.5 million of the new net proceeds of $259.8 million (claim #7490) is going to retire most of those same notes. The company is refinancing its own recent financing. Asset-level debt is modest by comparison: the RBC facility is $43.8 million (claim #2166), and the delayed-draw bridge is $100 million with a $3 million original issue discount (claims #8986, #1292).

Use of proceeds tracks most weakly at NC-1. WhiteFiber committed $150 million of equity to the project (claim #18331) and said a credit facility would close in early Q1 2026 (claim #18332). That timeline then slipped several times. It was listed as delayed in January (claim #17576), sat at term-sheet stage in May (claim #9070), and reached lender exclusivity in August (claim #8602). The bridge and the second convertible appear to have covered the gap left by the project financing that had not closed. Losses continue alongside all of this, with a Q2 net loss of $14.98 million (claim #201) on about $29 million of revenue (claim #6903).

A reader should watch four things next. The first is whether NC-1 permanent financing closes and on what terms. The second is how much of the bridge has been drawn. The third is the treatment of the roughly $32 million of 2031 notes not covered by the exchange. The fourth is whether the $40 million option on the new notes (claim #7773) is exercised. Revenue conversion against the $865 million Nscale contract (claim #3974) will show whether this debt load is sized to contracted cash flow. The record provided here ends partway through the August 2026 entries, and it contains no insider sales, so this read cannot cover either.

WhiteFiber's record describes a company that moved quickly from a single equity raise to a layered, largely debt-based capital structure, and then began refinancing that structure within months. The IPO was upsized from 7,812,500 to 9,375,000 shares (claims #4655, #4559) and priced at $17 (claim #4558), with the over-allotment later reported as fully exercised (claim #4420). After the IPO, Bit Digital's stake was reported to fall from 80 to 74 percent (claims #4719, #4564), so the float available to outside holders started narrow. Every raise since then has been convertible or secured debt. The disclosed history contains no follow-on common equity and no ATM.

The convertible sequence deserves close attention. The January notes carried a 4.500% coupon and a $26 conversion price (claims #3884, #3885). Claims give the closed size inconsistently: $230M in claim #17581, $250M in claim #3883. They agree on $222M of net proceeds (claim #17582), of which $120M went to a zero-strike call option (claim #3888). Seven months later, the August $270M offering at a $34 conversion price (claims #7771, #7489) committed $118.5M of cash and 6.3M shares to exchange $198.15M of those same 2031 notes (claim #7770). Taken together, roughly $238M of about $482M in combined net proceeds (claims #17582, #7490) went to share-count management and liability exchange rather than to capacity. The higher conversion strike suggests the company chose this sequence deliberately, but it reads more as serial repair of capital structure than as a pre-planned program.

The stated use of proceeds tracks actual deployment only loosely. A credit facility was expected to close in early Q1 2026 (claim #18332). What closed in March was a $43.8M RBC facility directed at MTL-2 (claims #2166, #2168). NC-1 project financing was first marked delayed (claim #17576), then advanced to a term sheet (claim #1314) and to lender exclusivity (claim #8602). In the interim, a $100M bridge with a $3M original issue discount (claims #1287, #1292) filled the gap, while the $150M NC-1 equity commitment (claim #18331) and the $865M Nscale contract (claim #3974) remain the underlying rationale. Against a Q2 net loss of $14.98M (claim #201), the company is financing well ahead of cash generation.

On insider activity, the record shows only officers indicating $72,500 of IPO purchases (claim #4591) and no disclosed sales. The input also ends mid-record, so later claims may be missing. The items to watch are whether NC-1 permanent financing actually closes and on what terms, whether the $40M greenshoe is exercised (claim #7491), how much of the bridge facility is drawn, and the status of the MTL-3 purchase option, which is currently marked delayed (claim #2181).

WhiteFiber's capital history since its August 2025 listing shows a company that has leaned on debt rather than repeated equity raises. It priced its IPO at $17 per share (claim #4558) for roughly $159.4M gross (claim #4562), with Bit Digital retaining about a 74% stake afterward (claim #4535). Since then, direct share issuance has been limited. The notable exception is the roughly 6.3M shares used in the August 2026 note exchange (claim #7770). The more consequential dilution sits in convertible instruments. The conversion price rose from $26 on the January 2026 notes (claim #17589) to $34 on the August 2026 notes (claim #1724). That rise means each dollar raised carries less potential dilution than before, even as the total convertible balance has grown.

The mix looks deliberate, though the pace of refinancing requires explanation. The January convertible was sized at $210M (claim #17586), upsized to $230M at close (claim #17581), and used $120M of proceeds for a zero-strike call option (claim #17591). That structure effectively pre-purchased shares to offset conversion. Only seven months later, the August $270M offering (claim #1719) directed $118.5M in cash plus stock toward exchanging about $198M of those same 2031 notes (claim #7770). A convertible that was reworked this quickly deserves scrutiny of its terms and cost. Alongside the convertibles sit asset-level borrowings: the RBC facility of $43.8M tied to MTL-2 equipment and real estate (claim #2166) and a $100M delayed-draw bridge term loan expandable to $150M (claims #8986, #8987). The bridge carries an original issue discount (claim #1292).

The stated destination of much of this capital is NC-1. That includes a $150M equity commitment (claim #18331) supporting the $865M Nscale contract (claim #3974). NC-1's permanent financing has slipped. It was expected to close in early Q1 2026 (claim #18332), moved to a term sheet in May (claims #1314, #9070), and had reached only lender exclusivity by August (claim #8602). The bridge loan and repeated convertibles appear to be filling that gap. Q2 2026 revenue of roughly $29M (claim #6903) against a $14.98M net loss (claim #201) confirms the business is not yet self-funding.

Several items are worth watching next: whether NC-1 permanent financing closes and retires the bridge, how much of the delayed-draw facility is drawn, whether the MTL-3 option (claim #2181, marked delayed) completes, and what happens to the residual 2031 notes. Some claimed figures conflict, such as $200M, $230M and $250M for the January notes (claims #17578, #17581, #3883). The record supplied was also truncated, so this read should be treated as provisional.

WhiteFiber has gone to market roughly every four months since listing, and each trip has been larger than the last. The August 2025 IPO raised $159.4M gross at $17 (claims 4562, 4558); the January 2026 convertible priced at $210M and upsized to $230M (17586, 17581); the August 2026 convertible priced at $250M and upsized to $270M (7832, 1707). Secured bank debt filled the gaps between them — the RBC package of a $43.8M facility, an $18.5M equipment lease and a $19.6M MTL-2 term loan (2166, 2167, 2168) in March, then a $100M delayed-draw bridge upsizable to $150M carrying $3M of original issue discount (8986, 8987, 1292) in May. The cadence is accelerating and the instrument mix is migrating from common equity toward structured and asset-level debt.

The escalation is at least nominally demand-backed, with the Nscale agreement carrying $865M of contracted revenue (3974). The January convertible's internal economics still warrant explanation: of $222M net (3887), $120M went to a zero-strike call option (3888), leaving roughly $102M of deployable cash against a $230M headline. Seven months later, $198.15M of those same 4.500% 2031 notes were exchanged for $118.5M cash and 6.3M shares (7770), funded by the new 5.000% 2032 notes struck at $34 (1722, 1724). The conversion price rose from $26 (3885), but so did the coupon, and net new money from August is nearer $141M once the cash leg is netted (7490, 7492). That is refinancing more than expansion capital.

Use of proceeds tracks where it is verifiable at the asset level — MTL-2 closed against the RBC real estate loan (2178, 2168). Corporate-level NC-1 permanent financing has been discussed since January (17576), sat at term sheet in May (1314, 9070, marked delayed), and had only reached lender exclusivity by mid-August (8602). It has not closed. The MTL-3 option closing (2181) is likewise delayed.

Watch whether NC-1 closes and at what advance rate, since the bridge's repayment path plausibly runs through it; whether the $150M upsize is drawn; the size of the 2031 stub after the exchange; and share count absorbing 6.3M exchange shares against $29M quarterly revenue and a $15.0M net loss (6903, 201). Insider data is thin — $72,500 of officer IPO purchases (4591) — and Bit Digital's 74% (4564) has not been revisited on the record.

WhiteFiber has raised capital on roughly a six-month clock since listing, and each round has been larger than the last: the $159.375M IPO at $17 per share (claims 4562, 4558) in August 2025, the $250M 4.500% convertible in January 2026 (claim 3883), a cluster of RBC secured facilities in March (claims 2166, 2168, 2193) totaling under $85M, a $100M bridge and delayed-draw structure in May (claims 1287, 8986), and the $310M 5.00% convertible that closed in August 2026 (claim 20). The trajectory is not simply more dilution but faster and heavier recourse to the convertible market specifically, with the coupon widening from 4.500% to 5.00% even as the conversion price rose from $26 to $34 (claims 3885, 7489).

The mix is mostly coherent at the layer level — equity at inception, asset-level secured debt against MTL-2 real estate (claim 2168), convertibles for corporate liquidity — but two items sit outside that logic. The May bridge carried a $3M original issue discount and a $20M assignment to B. Riley (claims 1292, 1293), which is stopgap pricing rather than planned structure. And the August offering moved from $250M proposed to $270M priced to $310M closed inside 48 hours (claims 7758, 2719, 7498), which reads as taking what the window offered.

Use of proceeds tracks less cleanly than the headline amounts suggest. Of $222M net in January, $120M went to a call option (claims 3887, 3888). Of $298M net in August, $118.5M plus 6.3M shares retired $198.15M of the 2031 notes (claims 21, 7770). In both cases roughly half the raise serviced the capital structure rather than the buildout, and the 6.3M shares converted a contingent claim into realized dilution. Against that, Q2 2026 revenue of $29M and a net loss of $14.98M (claims 6903, 201) remain distant from the $865M Nscale backlog (claim 3974).

Watch the NC-1 permanent project financing, described as progressing and at lender exclusivity across three August filings (claims 248, 8602, 8483) without a close, and whether it takes out the bridge. Watch also the delayed MTL-3 option (claim 2181), the undrawn $150M accordion (claim 8987), and the unreconciled $230M initial convertible figure (claim 210). No ATM and no insider sales appear on record; officer IPO purchases totaled $72,500 (claim 4591).

The cadence is the most informative feature here. WhiteFiber sat quiet for roughly five months after its August 2025 IPO at $17 (claim #4558) for 9,375,000 shares and $159.375M gross (claims #4559, #4562), then moved to a roughly two-month financing rhythm from January 2026 onward: the $250M 4.500% convertible closing 2026-01-25 (claims #3883, #3884), the RBC package in March comprising a $43.8M credit facility, a $18.5M equipment lease and a $19.6M MTL-2 real estate term loan (claims #2166, #2167, #2168), a $100M delayed-draw term loan with an optional step-up to $150M alongside a $100M bridge in May (claims #8986, #8987, #1287), and then the August convertible that was proposed at $250M, priced upsized at $270M and closed at $310M (claims #7758, #7720, #7498). That is acceleration, not steady state.

The mix reads as two coherent programs plus one patch. Secured, asset-level debt is matched to hard assets — MTL-2 real estate and equipment against RBC paper. Unsecured converts carry corporate liquidity, and the strike progression from roughly $26 to roughly $34 (claims #3885, #7489), with $120M of the January net proceeds of $222M spent on a call option (claims #3887, #3888), shows deliberate dilution management rather than indifference to it. The May bridge, with $3M of original issue discount and a $20M B. Riley assignment (claims #1292, #1293), sits outside that logic and looks like timing relief. No ATM appears on the record, and the only insider datum is $72,500 of indicated officer purchases at IPO (claim #4591) — a thin basis for any insider read.

Use of proceeds requires explanation in one place. The August convertible's largest identified application was retiring $198.15M of the January notes for $118.5M cash and 6.3M shares (claim #7499), meaning seven-month-old paper was refinanced at a higher coupon, and 6.3M shares entered the count at the exchange rather than at conversion. Separately, NC-1 permanent project financing has been described as progressing, at lender exclusivity, and anticipated across 2026-05-20 through 2026-08-20 (claims #1314, #8602, #7500) without a closing on record; the MTL-3 option exercise (claim #2181) is likewise marked delayed.

Watch whether NC-1 project debt actually closes or the convert has quietly become its substitute, whether the $865M Nscale contract (claim #3974) converts against a $29M quarterly revenue base (claim #6903), and whether debt-to-equity near 1.0 with a current ratio near 1.0 (claims #6907, #6908) leaves covenant room for another secured layer.

WhiteFiber has moved from an equity-funded spinout to a mostly debt-financed issuer in about thirteen months. After the August 2025 IPO at $17 (claim #4558), which raised $159.4M gross (claim #4562) and reduced Bit Digital's stake from roughly 80% to 74% (claims #4719, #4564), there has been no further primary equity offering. Dilution since then has come through convertibles instead. The January 2026 4.500% notes carried a conversion price near $26 (claims #3883, #3885), and $120M of the $222M net went to a call-option structure (claims #3887, #3888). The August 2026 exchange then retired $198.15M of those notes for $118.5M in cash plus 6.3M shares (claim #7770). That share count is about two-thirds of the IPO's 9.375M shares (claim #4559), which makes it the largest single dilutive event on record, even though it was not a capital raise.

The mix looks more like sequential patching than a planned structure. The 2031 notes were largely refinanced within about seven months into 5.00% notes due 2032 (claim #20), with a higher conversion price near $34 (claim #178). The deal was upsized from a proposed $250M (claim #1706) to $270M (claim #174), and $310M is reported as closed (claim #7498). The March RBC facility of $43.8M (claim #2166) and the May $100M delayed-draw term loan with a $3M discount (claims #8986, #1292) fill gaps at the asset level.

The announced funding path for NC-1 has not matched what has actually happened. A project-level term sheet was cited in May (claim #1314) and is marked delayed (claim #9070). By August it had reached lender exclusivity (claim #8602) but was still described as anticipated (claim #7500). In the meantime, corporate convertibles and bridge debt have carried the build, against $865M of contracted Nscale revenue (claim #3974) and a Q2 net loss of $14.98M (claim #201).

Several things are worth watching next:

- Whether the NC-1 permanent financing actually closes, and whether it retires the delayed-draw facility. - The treatment of the roughly $52M of 2031 notes that appear to remain outstanding. - Whether the January call-option position was unwound, which the record does not address. - Share-count and Bit Digital ownership disclosures after the exchange.

Some reported figures also need reconciliation. The January note size appears as $230M in claim #210 and $250M in claim #3883. The IPO proceeds appear as $144.5M in claim #4123 and $159.4M in claim #4562. The record contains no insider sales, and it breaks off mid-entry at claim #22.

WhiteFiber's capital history splits into two phases. The first was a single equity raise. The second, beginning within five months of listing, relies on convertible debt, secured bank facilities and project-level borrowing. The IPO was upsized from the 7.8M shares first indicated (claim #4655) to 9.375M at $17 (claims #4559, #4558). The greenshoe was fully exercised (claim #4420). Bit Digital kept roughly a 74% stake (claim #4564), so the public float at listing was narrow.

The convertibles set the dilution trajectory more than the IPO did. The January 2026 notes carried a 4.5% coupon and an initial conversion price near $26 (claims #3884, #3885). By August, $198.15M of those notes were being exchanged for $118.5M in cash plus 6.3M shares (claim #7770). That share count is about two-thirds of the IPO's entire share offering. The replacement notes priced at a higher conversion price of about $34 (claim #178) but a higher 5% coupon (claim #179). They were upsized from a proposed $250M (claim #7828) to a $310M close (claim #7498). Retiring notes seven months after issuance, largely with new convertible proceeds (claim #1726), looks more like refinancing on better equity terms than a pre-planned sequence. The record gives no rationale for it.

The secured debt is more clearly tied to specific assets. The RBC facility of $43.8M (claim #2166) is split between a $19.6M real estate loan for MTL-2 (claim #2168) and an $18.5M equipment lease (claim #2167). The $100M delayed-draw facility (claim #8986) is labeled a bridge (claim #1287), which implies it expects a takeout. The NC-1 term sheet is marked delayed (claim #9070), though it later reached lender exclusivity (claim #8602).

Use of proceeds is only partly traceable. The $120M of January proceeds directed to a call option (claim #3888) is large against $222M of net proceeds (claim #3887) and requires explanation. The August use-of-proceeds disclosure (claim #182) is not itemized in the record shown, so deployment beyond the note exchange cannot be verified.

Readers should watch four things. First, whether NC-1 permanent financing closes and repays the bridge, or whether the bridge is drawn and extended. Second, how the roughly $52M of 2031 notes still outstanding, if claim #3883's $250M is the full issuance, gets handled. Third, whether the $865M Nscale contract (claim #3974) shows up in revenue that currently runs about $29M a quarter at a negative 47% net margin (claims #6903, #6906). Fourth, any further exchanges that settle in shares. The claim list supplied here is truncated, so later disclosures may change this reading.

WhiteFiber's record describes a company that went from IPO to its second convertible in about thirteen months, with debt now doing most of the work. The August 2025 IPO priced at $17 (claim 4558) for roughly $159.4M gross (claim 4534), upsized from an earlier 7.8M-share filing (claim 4655), and Bit Digital's stake fell from about 80% to 74% (claims 4719, 4535). Five months later came the 4.500% convertible due 2031, recorded at $230M in one place (claim 210) and $250M completed in another (claim 3883), with a $26 initial conversion price (claim 3885). By March 2026 the company had added a $43.8M RBC facility (claim 2166), and in May a $100M delayed-draw bridge expandable to $150M (claims 8986, 8987). In August it moved to a 5% convertible due 2032 at a $34 conversion price (claims 178, 179), sized at $250M proposed (claim 1706), $270M priced (claim 174), and $310M closed (claim 7498). The mix reads as a sequenced approach rather than patching. Asset-level bank debt sits under specific Montreal sites, a bridge carries NC-1, and converts fund the holding company. The August deal partly refinanced the January one. Holders of $198.15M of the 2031 notes are exchanging for $118.5M cash plus 6.3M shares (claim 7770). That trades a lower strike for a higher one, but the 6.3M shares are dilution realized now rather than later. Use of proceeds requires explanation in two places. About $120M of the January proceeds went to a call option (claim 3888), and $118.5M of the August proceeds goes to the exchange (claim 1726). In both deals, a large share of headline size never reached data-center construction. The record does not show what claim 182's stated use covers beyond that. NC-1 permanent financing has moved from a term sheet in May (claim 1314), later marked delayed (claim 9070), to lender exclusivity in August (claim 8602), and is still not closed. Meanwhile, Q2 2026 shows revenue of about $29M (claim 6903) against a $14.98M net loss (claim 201). The next things to watch are whether the NC-1 financing closes and on what terms, how much of the bridge has been drawn, and whether the Duke Energy milestones that raise the NC-1 purchase price are met (claims 2185, 2186). Also worth watching is whether the $865M Nscale contract (claim 3974) turns into revenue fast enough to service two convertible layers. The record supplied here is truncated after the August 20 exchange entries and contains no insider sale or ATM data, so those absences should not be read as confirmed.

WhiteFiber has raised equity once and has since funded itself through convertibles, bank debt and a bridge facility, with a capital event roughly every two to three months. The August 2025 IPO priced at $17 (claim 4558) for $159.4M gross (claim 4562) and took Bit Digital's stake from roughly 80% to 74% (claims 4719, 4564). Everything after it has been debt-like: the $250M 4.500% notes due 2031 (claims 3883, 3884), the $43.8M RBC facility (claim 2166), the $100M delayed-draw term loan expandable to $150M (claims 8986, 8987), and the $310M 5.00% notes due 2032 (claim 20). The record also leaves unreconciled the $230M figure (claim 210) against the $250M completed amount for the first convertible.

The mix is only partly coherent. The RBC facility is asset-matched, but the 2031 notes were seven months old when $198.15M of them was exchanged for $118.5M in cash and 6.3M shares (claim 7770). In effect, a large part of the new issue refinances the old one at a higher coupon. It also turns dilution that depended on a $26 conversion price (claim 3885) into shares issued now, equal to about two-thirds of the 9.375M IPO shares (claim 4559). The $34 conversion price on the new notes (claim 178) limits future dilution only if the shares trade above it. After the cash leg, the $298M in net proceeds (claim 21) leaves considerably less new capital than the headline suggests.

Deployment is hard to verify from this record. The RBC term loan and lease map cleanly to the MTL-2 purchase and build-out (claims 2167, 2168, 2178, 2179). The planned use of the 2032 notes is not itemized here (claim 182), and nothing on record shows how the January proceeds were deployed, including the $120M directed to a call option (claim 3888). Insider activity is limited to officer indications in the IPO (claim 4591), and no sales appear.

The item that requires explanation is NC-1 project-level financing. A term sheet was reported in May (claim 1314) and later marked delayed (claim 9070). By August it had reached only lender exclusivity (claim 8602), while the company still cited additional project financing needs (claim 7754). With a Q2 net loss of $14.98M (claim 201), the next things to watch are whether NC-1 permanent financing closes, whether it replaces the bridge facility, and whether the $865M Nscale contract (claim 3974) turns into revenue on its stated timeline.

WhiteFiber's capital trajectory from IPO through mid-2026 reflects neither aggressive dilution nor a single coherent strategy, but rather a company managing through execution gaps. The company raised $159.4 million gross in its July 2025 IPO (claim 4534), then returned to markets nine months later with a $230 million convertible issuance in January 2026 (claim 210), with $120 million of those proceeds immediately committed to a call option (claim 3888)—a defensive move suggesting limited confidence in near-term revenue generation. Six months after that, WhiteFiber upsized a $250 million convertible offering to $310 million (claim 7498), closing on August 20, 2026, the third major capital raise in fourteen months.

The mix tells a story of project-level capex needs rather than operating flexibility. The January convertible bore 4.5 percent interest at a $26 conversion price (claims 3884, 3885); the August offering, upsized and closed at a $34 conversion price with 5 percent interest (claims 7489, 7488), shows conversion terms improving as the company returns to market. Yet this same offering included an exchange of $198 million of the prior 2031 convertible notes for $118.5 million in cash plus 6.3 million shares (claim 7770)—a refinancing at higher dilution that suggests either holders accepting risk or the company facing compressed alternatives.

Deployment tracking becomes opaque at the contract level. The company announced an $865 million revenue contract with Nscale Services in December 2025 (claim 3974), yet reported only $29 million in quarterly revenue by Q2 2026 (claims 6903, 9216). March 2026 funding—$43.8 million RBC credit facility, $19.6 million real estate term loan, and project-level financing (claims 2166, 2168, 2178)—targeted specific assets like MTL-2 and NC-1. Multiple NC-1 financing milestones show delays (claims 2181, 8602), and the May bridge facility remained delayed-draw (claim 1287). Use of proceeds tracked intent but not certainty of execution.

Sophisticated readers should track two metrics: the timing of NC-1 and MTL-2 revenue contribution against the $865 million contract runway, and whether Q4 2026 results show revenue acceleration sufficient to justify a $310 million convertible at $34 conversion price. The cadence of three raises in fourteen months, combined with improving conversion prices offset by note exchanges at higher dilution, suggests management betting on project ramp while capital markets demand evidence. The 30-to-1 gap between contracted and realized revenue is the proving ground.

WhiteFiber pursued a capital-light IPO followed by rapid debt and convertible financing—a sequence that reveals operational urgency more than strategic coherence. The July 2025 IPO priced at $17 per share (claim 4558) raised $159.375 million gross (claim 4562) while Bit Digital retained 74 percent ownership (claim 4564). Within six months, the company issued $230 million in convertible senior notes at a $26 conversion price (claims 210, 3885), hedging the exposure with $120 million in call options (claim 3888). By August, it replaced that convertible with a larger $310 million offering at a $34 conversion price (claim 7489), while simultaneously exchanging $198.15 million of older notes for $118.5 million in cash and 6.3 million dilutive shares (claim 7770). The pattern is refinancing at progressively worse terms rather than organic capital deployment.

Use-of-proceeds statements diverged sharply from actual capital flow. The company disclosed $865 million in contracted revenue from Nscale Services in December 2025 (claim 3974), yet Q2 2026 delivered only $29 million in revenue (claim 6903) against a $14.98 million net loss (claim 201), yielding minus-47 percent net margins (claim 6906). Project-level financing structures announced in March 2026—RBC credit facilities ($43.8 million), equipment leases ($18.5 million), and real estate term loans ($19.6 million)—addressed specific asset acquisitions (claims 2166, 2167, 2168), but the convertible raises clearly funded working-capital losses rather than growth. Multiple delayed reporting flags on convertible-related claims (7687, 7486, 7770, 7771) obscured execution timing.

Readers should monitor whether NC-1 project-level financing (claim 7500) closes or remains a term sheet (claims 1314, 7754), whether quarterly cash burn depletes the August $310 million proceeds before maturity risks emerge, and whether Bit Digital sustains its 74 percent ownership through future dilution or steps back.

WhiteFiber raised approximately $476 million across three equity events within fourteen months: an $166 million IPO closing in November 2025 at $17 per share (claim 4558, 4420), a $250 million convertible in January 2026 priced at $26 (claim 3883, 3887), and a $310 million convertible in August 2026 priced at $34 (claim 7498). The velocity and pricing suggest escalating dilution expectations: the conversion price climbed steadily from IPO reference through both convertibles, implying near-term stock appreciation baked into institutional demand. Bit Digital's ownership stake declined from 80 percent pre-IPO (claim 4719) to 74 percent post-IPO (claim 4535), with additional dilution embedded in both convertible issuances.

The announced deployment of proceeds, however, does not track the timing or scale of capital raised. An $865 million Nscale Services agreement cited in December 2025 (claim 3974) represents contracted future revenue, not capex deployed. The January convertible's $222 million net proceeds (claim 3887) directed $120 million toward a call option—a share repurchase facility (claim 3888)—rather than project build-out. Most tellingly, August 2026's $310 million convertible included a concurrent exchange of $198.15 million in existing 2031 notes (claim 7499) for just $118.5 million in cash plus 6.3 million shares (claim 22), a material concession to note holders suggesting liquidity pressure and refinancing rather than fresh capex. Meanwhile, RBC facilities (claim 2166) bundled with conditional price escalators tied to Duke Energy milestones (claims 2185, 2186, 2187) indicate reliance on project-level secured debt structures rather than consolidated capital.

Operational results complicate the narrative. Q2 2026 revenue reached only $29 million against a $14.98 million net loss (claim 201), yielding a negative 47 percent net margin (claim 6906) and a 1.0 debt-to-equity ratio (claim 6907). This financial profile suggests capex has not yet translated to profitable revenue production. NC-1 project financing achieved lender exclusivity in August 2026 (claim 8602), indicating permanent capital structures for major contracted revenue remain incomplete. A sophisticated reader should track: actual project deployment and revenue ramp timelines for MTL-2 and NC-1, unit-level economics as projects complete, quarterly cash burn relative to capital availability, and whether the company's next capital raise occurs at higher or lower terms. The financing mix to date—layered debt, rapid equity issuances, note exchanges—indicates opportunistic patching to fund a capital-intensive build-out ahead of operational proof points, not a coherent strategy.

WhiteFiber's capital structure reveals a coherent core strategy with opportunistic refinancing overlays. The company went public in mid-2025 with approximately $166M gross proceeds (IPO share offering per claim 4655 plus greenshoe per claim 4561), then deployed capital into MTL-2 real estate ($23M per claim 2178) and development ($24M per claim 2179), accessing $43.8M in secured credit facilities (per claim 2166) to fund infrastructure. This progression—equity base, then secured lines for real assets, then convertibles for development—tracks a conventional infrastructure-stage sequence.

The convertible issuances accelerated opportunistically. An initial $230–$250M tranche priced in January 2026 at $26 conversion (claims 210, 3883) was followed by a $310M convertible in August 2026 at $34 conversion (claims 20, 7498, 7489), paired with a concurrent exchange of $198.15M existing notes for $118.5M cash plus 6.3M shares (claim 22). The rising conversion price signals stock appreciation and willingness to monetize it; the note exchange indicates pressure to reduce interest-bearing debt. Net proceeds of $298M (claim 21) vastly exceeded quarterly revenue of $29M in Q2 2026 (claim 6903), pointing to lumpy, project-dependent deployment rather than operational cash generation.

The critical tension emerges in use-of-proceeds tracking. The company disclosed $865M in contracted Nscale revenue (claim 3974) alongside project-level financing discussions (claims 1314, 9070), yet actual quarterly revenue remained $29M by August 2026. Net margin of –47% and Q2 net loss of $14.98M (claims 6906, 201) confirm the raised capital funds infrastructure buildout, not yet backing revenue. Debt-to-equity of 1.0 and current and quick ratios of 1.0 (claims 6907–6909) reveal moderate leverage but tight liquidity margins despite $600M+ raised.

NC-1 permanent financing reached lender exclusivity August 21 (claim 466), likely gating the next deployment phase. A sophisticated reader should monitor three sequences: whether contracted revenue begins converting to actual quarterly runs, whether the company services the $310M convertible's 5% coupon (claim 7488) through the revenue ramp, and whether the $34 conversion price holds if project milestones delay. The pattern shows disciplined infrastructure financing, not opportunistic equity dilution—but revenue execution remains years ahead of capital deployment, requiring sustained access to capital markets.

WhiteFiber is pursuing a capital intensity model that requires continuous external funding, with heavy reliance on convertible instruments that show timing optimization but fragmented deployment strategy. The IPO raised $159.375 million at $17 per share in August 2025 (claim 4562), with founder-backed Bit Digital retaining 74 percent post-close (claim 4535). Two large convertible offerings followed within seven months: $250 million in January 2026 with a $26 conversion price (claims 3883, 3885), and $310 million in August 2026 with a $34 conversion price (claims 20, 1724). The rising conversion prices suggest stock appreciation between offerings, allowing the company to meter dilution by pacing raises to stock performance rather than operating milestones.

The capital mix reveals opportunistic patching over coherent strategy. The January convertible earmarked $120 million specifically for purchasing call options (claim 3888), a defensive hedge inconsistent with aggressive capex deployment. March 2026 brought multiple RBC credit facilities totaling roughly $40-45 million in smaller tranches (claims 2166, 2167, 2168, 2194), suggesting project-level financing constraints. May's delayed-draw term facility of $100-150 million (claims 8986, 8987) remains undrawn as of the latest data, indicating capital was raised ahead of deployment clarity. The August convertible included a concurrent exchange of $198.15 million of existing notes for $118.5 million cash plus 6.3 million shares (claim 22), a refinancing maneuver that converts lower-priced debt into higher-priced equity.

Use-of-proceeds tracking is problematic. The December 2025 announcement of an $865 million Nscale Services contract (claim 3974) justified the capital raises, but Q2 2026 revenue registered only $29 million (claim 6903) against a $14.98 million net loss (claim 201), yielding a negative 47 percent net margin (claim 6906). This gap between contract size and realized revenue requires explanation; either deployment is behind forecast or contract terms are substantially back-loaded. Financial metrics show limited cushion: current ratio 1.0, quick ratio 1.0, debt-to-equity 1.0 (claims 6908, 6909, 6907). NC-1 secured financing entered lender exclusivity in August (claim 466), indicating that project-level funding is critical to delevering the balance sheet. Sophisticated readers should closely monitor whether Nscale revenue accelerates in coming quarters and whether NC-1 financing closes on the implied timeline; any slippage creates immediate refinancing pressure on a balance sheet already stretched from multiple convertible raises.

WhiteFiber's capital architecture reveals a company that moved from orderly equity financing to heavy convertible reliance in under a year. The IPO in July-August 2025 generated approximately $159.4M in gross proceeds (per claim 4534), with additional over-allotment proceeds of $22.2M (claim 4124), positioning the company as a standalone public entity with Bit Digital retaining a $74M post-IPO stake (claim 4535). This was then layered with a $230M convertible issuance in January 2026 at a $26 conversion price (claim 210), followed by a $310M convertible offering in August 2026 at $34 (claim 20)—an upsized bid from the originally proposed $250M (claim 307). The trajectory suggests capital sufficiency eroded faster than expected.

The capital mix shows opportunistic patching rather than disciplined strategy. Project-level secured financing through RBC for approximately $44M in equipment and real estate facilities (claims 2192, 2167, 2168) demonstrates selective asset-based lending, yet this remained insufficient. The August 2026 offering included a concurrent exchange of $198M in existing 2031 convertible notes for $118.5M in cash and 6.3M shares (claim 22)—dilutive refinancing of debt maturing in five years into longer-dated instruments. This exchange suggests either accelerating capital needs or market-window opportunism.

Use of proceeds mismatch is notable. The December 2025 announcement of $865M in Nscale Services contracted revenue (claim 3974) implied substantial deployment opportunity, yet Q2 2026 revenue totaled only $29M (claim 6903) against a quarterly net loss of $14.98M (claim 201). The January 2026 convertible allocated $120M toward a call option (claim 3888), suggesting refinancing of another debt instrument rather than capex. August 2026 net proceeds of approximately $298M (claim 21) were directed into the NC-1 project financing structure, per lender exclusivity entered on August 21 (claim 466)—indicating project-level capital starvation.

Sophisticated investors should monitor three vectors. First, the NC-1 permanent financing closure (currently in exclusivity, per claim 8602): project-level debt typically assumes operating cash flow that WhiteFiber has not yet demonstrated at scale. Second, stock price relative to the $34 conversion price on the August convertible: trading above that level accelerates dilution and signifies market confidence in the Nscale revenue thesis. Third, the cash burn rate—a $15M quarterly loss on $29M revenue at negative 47% net margin (claim 6906) is unsustainable without either revenue acceleration or capital discipline. The company has now raised over $600M in public markets within 13 months; further capital raises will face skepticism until project economics demonstrate operational viability.

WhiteFiber went public in mid-2025 and immediately pivoted toward convertible financing, signaling both confidence in equity value and urgency around capital needs. The IPO closed in November with $144.5 million in gross proceeds per claim 4123, augmented by $22.2 million from the fully exercised greenshoe option per claim 4124, for roughly $166 million of IPO equity capital. Within two months the company filed a $250 million convertible offering per claim 3883, issued at 4.5% per claim 3884 with a $26 conversion price per claim 3885 and $222 million net proceeds per claim 3887. By August 2026 it refinanced that instrument with a larger $310 million convertible due 2032 at 5.0% per claim 179 with a $33.84 conversion price per claim 452, upsized from an initially proposed $250 million per claim 7758. That concurrent exchange saw WhiteFiber pay $118.5 million cash and issue 6.3 million shares to retire $198 million of the original convertible per claim 453, leaving $31.85 million outstanding per claim 454.

The mix reveals opportunistic refinancing rather than purposeful capital architecture. The January convertible, issued barely two months post-IPO, was refinanced nine months later at materially worse terms: 5% versus 4.5% coupon, and a $34 conversion price versus $26, suggesting either equity weakness relative to initial expectations or acute near-term cash pressure. Claim 3888 noted that $120 million of the first convertible's proceeds funded a call option, an opaque deployment. March 2026 brought $82 million in RBC facilities per claims 2166-2168 transparently tied to MTL-2 real estate acquisition and development per claims 2178-2179, alongside conditional purchase escalators for NC-1 totaling up to $18 million per claims 2185-2187. A May bridge facility of $100 million per claim 1287 remained unresolved as of August. MTL-3 option exercise, per claim 2181, remains delayed.

The stark disconnect between capital and operations invites scrutiny. Q2 2026 revenue reached only $29 million per claim 6903, generating a $15 million net loss per claim 201 and a -47% net margin per claim 6906. Claimed contracted revenue of $865 million from an Nscale Services agreement per claim 3974 has not materialized in reported results, warranting explanation of timing, conditions, or scope. The NC-1 project entered exclusivity for permanent financing per claim 466, but that anticipated project-level financing per claim 7500 has yet to translate into disclosed terms or deployment. Monitoring cash burn against announced runway and tracking whether the Nscale contract flows into Q3+ results will clarify whether WhiteFiber faces a growth inflection or persistent execution risk.

WhiteFiber's capital trajectory from IPO in July 2025 through August 2026 reflects a company continuously reaching for additional funding despite sequential major raises, suggesting either aggressive expansion targets or underestimated capital needs at public market entry. The IPO at $17 per share generated approximately $159.4 million, per claims 4534 and 4562, yet within six months the company had announced a $230 million convertible at an effective conversion price of $26, per claims 210 and 3885. By August 2026, a second convertible round was upsized to $310 million at a $34 conversion price, per claims 20 and 7489, alongside a concurrent exchange of $198.15 million of older 2031-maturity notes for $118.5 million in cash and 6.3 million shares, per claim 22. This exchange mechanics—taking existing convertibles at cheaper conversion levels and rolling them into equity plus new debt at higher prices—is a coherent strategy for managing dilution while extending maturity risk, though it accelerates near-term equity issuance.

The stated uses of proceeds diverge from observable deployment. The January convertible explicitly allocated $120 million to a call option hedge, per claim 3888, a defensive posture at odds with a growth narrative. The August convertible claimed intended use toward data center expansion and anticipated project-level financing for NC-1, per claims 6875 and 6876, yet revealed project-level funding was segmented separately—RBC credit facilities and term loans totaling $43.8 to $44 million for MTL-2 and ancillary facilities, closed in March, per claims 2166 and 2215. The announced $865 million Nscale Services contract, per claim 3974, appeared in November 2025 post-IPO but Q2 2026 revenue reported only $29 million, per claim 6903. At that revenue level against a Q2 net loss of $14.98 million, per claim 201, the company faced a −47 percent net margin, per claim 6906.

A sophisticated reader should scrutinize three items. First, the timing gap between the Nscale contract announcement and revenue recognition requires explanation—either the project is ramping slower than public filings implied or accounting treatment differs materially. Second, current and quick ratios both at 1.0, per claims 6908 and 6909, alongside a debt-to-equity ratio of 1.0, per claim 6907, suggest limited liquidity cushion despite three major capital events in 14 months. Third, the NC-1 secured financing announced in August as anticipated, per claim 248, still lacks closing confirmation in the available record; facility-level financing gaps often precede covenant pressure or capital calls. The pattern suggests a company aggressively building ahead of realized revenue.

WhiteFiber's capital raising pattern from July 2025 through August 2026 reads as tactical refinancing masking operational stress rather than disciplined deployment. The IPO in mid-2025 raised approximately $144.5 million (per claim 4123, inclusive of the $22.2 million over-allotment exercise per claim 4124), with Bit Digital retaining 74 percent ownership (claim 4564), leaving minimal public float. By March 2026, the company had layered in a $43.8 million RBC credit facility (claim 2166) alongside targeted project-level debt for MTL-2 and MTL-3 development, suggesting capital intensity consistent with infrastructure deployment.

The divergence emerges sharply between announced opportunity and demonstrated economics. In December 2025, the company disclosed an $865 million contracted revenue agreement with Nscale Services (claim 3974), yet Q2 2026 results reported only $29 million in quarterly revenue (claim 6903) against a net loss of $14.98 million (claim 201), yielding a negative 47 percent net margin (claim 6906). This gap requires explanation—whether the contract remains unfulfilled, requires substantial capex before revenue recognition, or signals demand risk.

The convertible issuance cadence reveals opportunistic stacking. The initial $250 million convertible (claim 3883, closed January 2026) at 4.5 percent interest and $26 conversion price (claim 3885) was followed by a $222 million convertible in May (claim 1341) and a $100 million bridge facility (claim 1287). By August, the company priced a $310 million convertible offering (claim 20) at a $34 conversion price (claim 1724), concurrently exchanging $198.15 million of the earlier 2031-maturity convertible for $118.5 million cash and 6.3 million shares (claim 22, noted delayed). This structure—rolling forward higher-priced equity conversion while taking new debt—is defensive refinancing, not growth capital allocation.

The $310 million offering with net proceeds of approximately $298 million (claim 21) is material relative to the IPO base, implying either the revenue ramp from the Nscale contract materializes rapidly or the company faces funding pressure. Sophisticated readers should monitor whether the delayed note exchange (claim 22) closes as announced, as slippage may signal lender hesitation about equity dilution or valuation defensibility. The RBC facilities remain project-secured and tied to specific assets, limiting flexibility. Until Q2 or Q3 revenue inflects toward the Nscale contract scale, the capital structure remains defensive.

WhiteFiber's capital structure over the past 13 months reflects a company running simultaneous refinancing and expansion playbooks. The IPO at $17/share (claim 4558) in July 2025 raised $159.4M gross (claim 4562), but within six months the company had already sold $250M of convertible senior notes at a $26 conversion price (claim 3883, 3885). That first convertible closed with $222M net proceeds (claim 3887), of which $120M was committed to a call option (claim 3888)—likely an aggressive share purchase or hedge transaction that immediately deployed half the capital raise. The timing suggests management either wanted to support the stock price or lock in early-stage leverage before additional refinancing.

By August 2026, WhiteFiber upsized and repriced a second convertible to $310M with a $34 conversion price (claim 7498, 178), more than 30 percent higher than the first note despite an operating loss of $14.98M in Q2 (claim 201) and a negative 47 percent net margin (claim 6906). The most material detail: the company simultaneously exchanged $198.15M of its earlier 4.5% convertibles due 2031 (claim 7770) into $118.5M cash plus 6.3M shares—accepting an $79.65M cash reduction to replace lower-priced future dilution with current equity issuance. That trade signals management prefers immediate dilution at $34 over contingent dilution at $26.

The capital stack reveals opportunistic layering. Project-level financing from RBC ($43.8M credit facility plus $18.5M equipment leases, claims 2166–2167) sits alongside convertible debt and a $100M bridge facility (claim 1287). Each funding stream targeted specific needs—MTL-2 and MTL-3 real estate ($23M and $17M, claims 2178, 2181), NC-1 data center project finance—but the corporate convertibles lack such granularity or declared use beyond the first offering's call option.

The material gap is the Nscale contract. The company announced $865M in contracted revenue (claim 3974) in December 2025, yet Q2 2026 reported only $29M revenue (claim 6903). Either the contract is steeply back-end loaded or deployment has stalled. With a debt-to-equity ratio now at 1.0 (claim 6907) and no visible profitability path, watch whether Q3 2026 revenue accelerates materially or whether the company returns to the convertible markets before year-end. The rising conversion prices despite operational losses require explanation—either hidden asset value or a capital market that expects sharp near-term inflection.

WhiteFiber's capital trajectory reveals escalating reliance on convertible instruments in lieu of equity or traditional debt, a pattern that has accelerated markedly since the company's public listing. The July 2025 IPO yielded $144.5M in gross proceeds (claim 4123), with Bit Digital maintaining 74% ownership post-close (claim 4564). Six months later, in January 2026, the company issued a $250M convertible (claim 3883) at a 4.5% coupon (claim 3884), net proceeds of $222M (claim 3887). Notably, $120M of that capital was earmarked for a call option (claim 3888), a mechanism typically used to manage shareholder dilution—signaling early concern about ownership deterioration. In parallel, WhiteFiber drew approximately $44M in RBC credit facilities (claim 2166) and equipment leases (claim 2167), all marked for specific project assets: MTL-2 real estate acquisition and development (claims 2178–2179), and facility buildout.

By August 2026, the company's capital needs had shifted dramatically. It announced a $270M convertible offering (claim 7742) that was subsequently upsized to $310M (claim 20), with net proceeds of $298M (claim 21). Concurrently, it executed a note exchange of $198.15M of the earlier convertibles for $118.5M in cash and 6.3M shares (claim 7770). This refinancing reprices old debt at $26 conversion (claim 3885) into new obligations at $34 (claim 6)—economically unfavorable to prior holders. The pattern signals growing urgency: the announced $865M Nscale Services contract (claim 3974) remains substantially unmonetized. Q2 2026 revenue reached only $29M (claim 6903) against a $14.98M net loss (claim 201), with debt-to-equity at 1.0 (claim 6907). A sophisticated reader should examine whether the escalating convertible issuance—and the willingness to refinance prior debt at worse terms—reflects confidence in the revenue ramp, or instead reflects a tightening liquidity position that necessitates increasingly dilutive capital raises. The call option mechanism from January has no follow-up disclosure; whether it was exercised or abandoned requires explanation.

WhiteFiber's eighteen-month capital strategy from IPO through August 2026 splits between infrastructure funding and opportunistic refinancing. The IPO generated $144.5 million net (claim 4123), augmented by $22.2 million from the over-allotment exercise (claim 4124). In January 2026, the company issued a $250 million convertible at 4.5 percent interest with a $26 conversion price (claim 3883, 3885), netting $222 million (claim 3887). This coincided with a $865 million service contract announcement with Nscale (claim 3974), positioning the company for substantial revenue scaling. A second convertible of $222 million closed in May 2026 (claim 1341). Then in August, the capital activity shifted: management initially proposed a $250 million convertible offering, which resized to $270 million, then to $310 million at pricing (claim 5). The company captured approximately $298 million net (claim 21), while announcing a concurrent exchange of $198.15 million in January convertibles for $118.5 million in fresh cash and 6.3 million shares (claim 7770). The new convertible carries 5 percent interest and a $34 conversion price—a higher coupon offset by reduced dilution per bond.

The three-day arc from $250 million to $310 million proposal signals sizing in response to market appetite rather than pre-planned requirements. The concurrent refinancing—retiring lower-priced convertibles ($26) to issue higher-priced ones ($34) while extracting $118.5 million in cash—is consistent with either improved valuation confidence or balance-sheet repair under tightening conditions. Q2 2026 results complicate interpretation: the company reported a $14.98 million net loss (claim 201) on $29 million revenue (claim 6903), a negative 47 percent net margin (claim 6906). Against the $865 million Nscale contract, this signals either long-duration revenue recognition or slower ramp than contracted. The current ratio of 1.0 (claim 6908) remains tight despite $1 billion in raised capital since IPO, indicating rapid capex deployment into data center assets (NC-1, MTL facilities via RBC financing). Debt-to-equity stands at 1.0 (claim 6907), concentrated in convertibles maturing 2031–2032.

Readers should monitor: (1) timing and magnitude of Nscale revenue recognition; (2) whether project-level capex generates sufficient returns to service the convertible stack; (3) dilution timing and magnitude as conversion probabilities crystallize against negative current margins; (4) liquidity trajectory as capex phases complete, given the tight current ratio.

WhiteFiber's August 2026 capital activity reflects deliberate debt refinancing rather than dilution-driven raises. The company issued $310M in convertible senior notes due 2032 at 5%, per claim 20, generating $298M in net proceeds, per claim 21. On the same day, it executed an exchange transaction, per claim 22, swapping $198.15M of existing convertible debt for $118.5M in cash and 6.3M shares. Simultaneously, total convertible notes outstanding declined by $32M, per claim 23. The compressed timeline—all capital events within a single trading day on August 20—indicates a coordinated refinancing strategy rather than reactive capital-raising.

The financial structure shows design rather than improvisation. By exchanging older debt into new debt plus equity, the company reduced debt principal while buying back distressed or expiring convertibles at favorable terms. The new convertibles priced at $34 conversion, per claim 6, establishing a clear valuation anchor. Insider transactions in early July (two buy transactions) may reflect confidence ahead of the capital events, though transaction volumes are incomplete.

Tracking use of proceeds remains unclear. An earnest money deposit of $2.25M for property acquisition was recorded on August 16, per claim 30, before the capital raise. Post-exchange, the company retains approximately $179.5M in net available cash ($298M proceeds minus $118.5M paid in the exchange), but timing is too compressed to confirm actual deployment. Duplicate reporting of the $310M offering, $298M proceeds, and exchange transaction across multiple claim IDs suggests potential data ingestion issues.

The critical question for investors: whether the retained $179.5M capital is deployed for organic growth, infrastructure capex, or reverts to further financial optimization. The earnest money deposit hints at potential acquisition activity, but the August 16 timing predates the capital raise and requires explanation. Investors should monitor the next quarterly guidance for capex commitment levels and cash deployment cadence. Until deployment is visible, this appears as capital structure optimization rather than growth financing.

WhiteFiber's record contains two named, signed customer contracts. The rest is unnamed tenants, utility undertakings, or language short of a binding agreement. The first signed contract is Cerebras at MTL-3: 5 MW of IT load over five years (24237, 17567), installed in October 2025 (19577) and billing CAD 1.4 million a month since November 1, 2025 (2208). The second is Nscale at NC-1: a master services agreement dated November 22, 2025 (3971) for a 10-year, 40 MW initial service order worth about $865 million, including escalators (18325, 2195, 2169). That deployment is not expected to run until mid-2026 (17566).

The cloud book is real but anonymous. The "Initial Customer" is never identified. Its MSA is dated November 9, 2023 in some filings (4718, 4555) and described as finalized December 12, 2023 in others (2202, 2218). Each version covers 2,048 GPUs, plus a second 2,048-GPU tranche added later (2204). The 14 MTL-1 lessees (4575), the 23 total customers (24239) and the 5.5 MW of executed Blönduós contracts (2174) are also unnamed. Claim 2199 is truncated and cannot be assessed.

The Duke Energy items do not belong in the customer column at all. They are utility supply milestones of 24, 40 and 99 MW gross, given on a "commercially reasonable efforts" basis and repeated across at least six filings (4710-4712, 4583-4585, 4546, 2188-2190). NVIDIA Preferred Partner status (2171) is a channel credential, not revenue.

The weakest language sits at NC-1. In August 2025 the company cited an LOI with an unnamed "AI cloud provider" worth about $230 million over seven years (24238). By November it described a "re-marketing process" and "closing stages of discussions with multiple highly creditworthy counterparties" (19579). That implies the LOI never converted. Nscale signed a month later.

Concentration has shifted from one unnamed GPU-cloud customer toward two named colocation tenants. Meanwhile an unnamed arrangement worth about $21 million of annualized cloud run rate is winding down (4125).

The arithmetic mostly reconciles, with caveats. Cerebras at CAD 1.4 million a month comes to about $11.7 million a year in USD, or roughly $59 million over the term. Nscale averages about $86.5 million a year, but escalators mean the opening rate is lower, and that revenue arrives only after commissioning. On that basis, the $21 million wind-down exceeds Cerebras's entire annual contribution. Power is the larger issue. Nscale's 40 MW is critical IT load, while Duke's April 2026 milestone is 40 MW gross. Delivering the full contract likely requires part of the 99 MW tranche, which carries only a best-efforts commitment.

WhiteFiber's record contains three signed and named customer agreements; most of the remaining language is either efforts-based or anonymous. The Cerebras colocation contract at MTL-3 is named and executed: 5 MW of IT load over five years, with billing beginning November 1, 2025 (claims 4708, 24237, 19577, 2170). The Nscale master services agreement is also named and executed. It was entered November 22, 2025 for the first 40 MW of IT load at NC-1 and carries roughly $865 million of contracted revenue over ten years (claims 3971, 18325, 2195). The third is the 2023 cloud MSA, which was expanded in 2024 by 2,048 GPUs, but that counterparty has only ever been called the 'Initial Customer' (claims 4718, 2204). The Blönduós contracts for 5.5 MW (claim 2174), the 14 MTL-1 lessees (claim 4575), and the 23-customer count (claim 24239) are likewise unnamed. The Duke Energy items are not customer revenue at all. They are a power supplier's commitment to 'commercially reasonable efforts' toward 24, 40 and 99 MW gross (claims 4710, 4711, 4712, 2164). NVIDIA Preferred Partner status is a reseller authorization, not a contract for demand (claim 2171).

The NC-1 path shows why the distinction matters. In August 2025 the company presented a Phase 1 LOI with an unnamed AI cloud provider worth about $230 million over seven years (claim 24238). By November that had become a 're-marketing process' and 'discussions with multiple highly creditworthy counterparties' for an unidentified anchor (claim 19579). The LOI evidently did not convert, and Nscale was signed in its place. In the same quarter, a customer representing about $21 million of annualized cloud run rate began winding down (claim 4125), and the filings do not say who it was. Concentration has therefore moved from an anonymous cloud customer toward Nscale. Nscale's average of about $86.5 million a year dwarfs Cerebras at about $11.7 million.

The Cerebras figures reconcile: CAD 1.4 million (about USD 979 thousand) a month for 60 months gives roughly $58.7 million (claim 2170), though the USD figure carries currency exposure. Nscale is harder to reconcile. Forty MW of critical IT load needs more than 40 MW of gross power; MTL-3 runs 5 MW of IT on 6.6 MW gross (claim 24237). Duke's 40 MW gross milestone for April 2026 is efforts-based, and the 99 MW tranche runs to 2029. That leaves the mid-2026 operational target (claim 17566) dependent on power delivery that no one has firmly committed.

WhiteFiber's record has only two signed customer agreements where the counterparty is named. The first is Cerebras, at 5 MW of IT load at MTL-3 (4708, 4642, 24237). Its status moved from "subject to our receipt of all required permits" (4684, 4580) to installed and revenue-generating in October 2025 (19577), with billing from November 1, 2025 (2170). The second is Nscale, under a master services agreement signed November 22, 2025 (3971). Its initial service order covers 40 MW at NC-1 for about $865 million over ten years (18325, 2169, 17566). Everything else is either unnamed or not a customer at all. The "Initial Customer" cloud MSA is never identified (4718, 2202). The record dates it November 9, 2023 in one place and says it was "finalized" December 12, 2023 in another. The 14 lease customers at MTL-1 (4575, 4538) and the parties to the 5.5 MW of executed contracts at Blönduós (2174) are also unnamed. Duke Energy appears in more claims than any other party, but it is the power supplier, not a buyer. Its commitments are "commercially reasonable efforts" milestones of 24, 40 and 99 MW (4710–4712, 4644–4646, 2211), not firm capacity obligations. NVIDIA Preferred Partner status (2171) is a channel credential, not a revenue contract.

The NC-1 history is the clearest case of LOI language not converting. In August 2025 the company presented a Phase 1 LOI with an unnamed "AI cloud provider" expected to produce about $230 million over seven years (24238). By November that party was gone. The company described a "re-marketing process" and "closing stages of discussions with multiple highly creditworthy counterparties" for an anchor agreement (19579). Nscale signed weeks later on different terms, so the $230 million figure should be treated as superseded, not as a backlog item.

Concentration has shifted sharply. In August 2025 the company reported 23 customers (24239), and the base was spread across MTL-1 colocation and GPU cloud. In November it disclosed the wind-down of an unidentified arrangement worth about $21 million of annualized cloud run rate (4125). Once Nscale is fully live, it averages roughly $86.5 million a year including escalators. That would make a single NC-1 tenant the dominant revenue source. Nscale's mid-2026 start date (17566) also depends on Duke reaching 40 MW by April 2026 on a best-efforts basis (4711, 2189), and the record does not confirm that energization.

The arithmetic holds where it can be tested. Cerebras billing of CAD 1.4 million a month (about $979,000) over 60 months gives the stated $58.74 million (2170). That figure is exposed to CAD/USD movements. The Nscale TCV is stated consistently across filings. The Initial Customer figure (2218) is too truncated to reconcile, and no claim states which customer the $21 million wind-down refers to.

WhiteFiber's record contains three counterparties that are both named and tied to an executed agreement. The first is Cerebras, which holds a 5 MW IT-load, five-year colocation contract at MTL-3 (claims 4708, 24237). Installation was completed in October 2025 (claim 19577), and billing began November 1, 2025 at CAD 1.4 million per month (claim 2170). The second is Nscale, whose master services agreement was signed November 22, 2025 (claim 3971) and carries an initial 40 MW, ten-year service order worth roughly $865 million (claims 18325, 2169). The third, Duke Energy, appears repeatedly (claims 4710-4712, 4644-4646, 2211), but it is the power supplier, not a customer. Its obligation is limited to "commercially reasonable efforts" toward 24, 40 and 99 MW gross, which is a best-efforts commitment rather than a firm delivery. NVIDIA Preferred Partner status (claim 2171) is a channel designation, not revenue.

Much of the remaining customer base is unnamed. The "Initial Customer" behind the 2,048-GPU cloud MSA is never identified, and the filings give it two different dates: November 9, 2023 in claim 4718 and December 12, 2023 in claim 2202. The 14 MTL-1 lessees (claim 4575), the 23-customer count (claim 24239) and the Blönduós tenants under 5.5 MW of executed contracts (claim 2174) are also anonymous. The NC-1 Phase 1 counterparty was only ever described as an "AI cloud provider" under an LOI expected to produce about $230 million over seven years (claim 24238). By November that deal had evidently lapsed, because the company described a "re-marketing process" and talks with "multiple highly creditworthy counterparties" for an anchor agreement (claim 19579). Nscale was named a month later.

Concentration has risen sharply. The cloud business started with one unnamed anchor, and a customer arrangement worth about $21 million of annualized cloud run rate began winding down after September 30, 2025 (claim 4125); the filings do not say which customer it was. Contracted backlog is now dominated by Nscale, whose $865 million is roughly fifteen times Cerebras's total.

The Cerebras figures reconcile: $979 thousand a month over 60 months comes to about $58.7 million, or roughly $2.35 million per MW per year. Nscale's figure implies about $86.5 million a year, or $2.16 million per MW per year including escalators, which is consistent with Cerebras. One gap remains. Nscale contracts 40 MW of critical IT load (claim 18325), while Duke's April 2026 target is 40 MW gross (claim 2189), so the power needed above IT load depends on Duke's best-efforts ramp toward 99 MW. The mid-2026 operational date for Nscale (claim 17566) has no confirmation anywhere in the record.

WhiteFiber's record contains only two counterparties that are both named and tied to signed, priced customer contracts. The first is Nscale, under a November 22, 2025 master services agreement for 40 MW at NC-1 with approximately $865 million of total contracted revenue over ten years (claims 3971, 18325, 2169). The second is Cerebras, under a five-year, 5 MW IT-load colocation contract at MTL-3 that has been billing since November 1, 2025 (claims 19577, 2170). Most of the other entries in the record are restatements of these two deals or of the Duke Energy capacity agreement. Duke Energy is a power supplier, not a customer, and its 24/40/99 MW milestones are 'commercially reasonable efforts' undertakings rather than firm delivery obligations (claims 4710–4712, 2211). The NVIDIA Preferred Partner status (2171) is a channel designation. The Blönduós arrangement (2199, 2174) is leased supply from an unnamed facility partner.

The rest of the customer base has never been named. The 'Initial Customer' behind the 2,048-GPU cloud MSA is unidentified. Its contract is dated November 9, 2023 in earlier filings (4718, 4555) but described as finalized December 12, 2023 in the March 2026 filing (2202), a discrepancy the company does not explain. The 14 MTL-1 tenants (4575) and the 23-customer headcount (24239) are also anonymous. The August 2025 NC-1 Phase 1 LOI with an unnamed 'AI cloud provider' for roughly $230 million over seven years (24238) did not convert. By November the site was being re-marketed to 'multiple highly creditworthy counterparties' (19579), and Nscale was signed a month later.

Concentration has inverted. Through mid-2025 revenue came from a diffuse colocation tenant base plus one unnamed cloud anchor. The $21 million annualized cloud run-rate wind-down disclosed in November 2025 (4125) removed a meaningful piece of that cloud book, and the customer involved was not named. Contracted backlog now rests overwhelmingly on Nscale, whose $865 million TCV is roughly fifteen times the Cerebras contract.

The arithmetic reconciles where the company gives the inputs. CAD 1.4 million per month, or about $979 thousand, over 60 months equals the $58.74 million attributed to Cerebras (2170). Nscale's figure implies about $86.5 million per year on average, or roughly $2.2 million per MW-year including escalators, and the company says it is not yet operational (17566). The Initial Customer's '$50' value (2218) is unitized ambiguously. It also cannot be tied to the $21 million wind-down without disclosure of which arrangement was terminated.

WhiteFiber's record has two signed, named customer contracts, a larger base of unnamed customers, and a substantial layer of non-binding or supply-side language. The two named agreements are Cerebras and Nscale. Cerebras holds a 5 MW IT-load, five-year colocation contract at MTL-3, which began billing on November 1, 2025 (claims 4708, 24237, 19577, 2170). Nscale signed a master services agreement on November 22, 2025 for 40 MW at NC-1, a 10-year deal with about $865 million in total contract value (claims 3971, 18325, 2169). Every other customer goes unnamed. These include the 2023 "first cloud customer," also called the "Initial Customer" (claims 4718, 2202), 14 lessees at MTL-1 (claim 4575), a 23-customer headline count (claim 24239), and contracts for 5.5 MW at Blönduós (claim 2174). The Initial Customer's agreement is dated November 9, 2023 in one filing (claim 4718) and December 12, 2023 in a later one (claim 2202), and that discrepancy should be resolved. The many Duke Energy entries (claims 4710 through 2190) describe a power supplier, not a customer. Duke's obligation is only to use "commercially reasonable efforts," which is not a firm delivery commitment. NVIDIA Preferred Partner status (claim 2171) is a channel credential, not revenue. NC-1's path to a signed deal was not linear. In August 2025 the company cited an LOI with an unnamed AI cloud provider worth about $230 million over seven years (claim 24238). By November it had re-marketed the site and was in "closing stages of discussions with multiple" counterparties (claim 19579), which implies the LOI did not convert. Concentration has shifted from the single 2023 GPU customer to Nscale, whose $865 million contract value dwarfs everything else on record. At the same time, an unnamed customer arrangement worth about $21 million of annualized cloud run rate began winding down after September 30, 2025 (claim 4125). The filings do not confirm whether that is the Initial Customer, whose three-year term would end around late 2026. On the arithmetic, Cerebras reconciles cleanly: about $979 thousand per month over 60 months comes to the $58.74 million carried in claim 2170. Nscale averages about $86.5 million a year, but the figure includes escalators and revenue starts only with a mid-2026 deployment (claim 17566), so it is backlog rather than current revenue. One gap needs scrutiny. Nscale's 40 MW is critical IT load, while Duke's April 2026 milestone is 40 MW gross (claim 4711), and no claim on record confirms that either Duke milestone was actually met.

WhiteFiber's record contains three signed customer agreements and a larger body of language that is either non-binding or refers to suppliers rather than customers. The first signed agreement is the three-year cloud Master Service Agreement with an unnamed 'Initial Customer,' dated November 9, 2023 and finalized December 12, 2023 for 2,048 GPUs (claims 4718, 2202, 2218). The second is the 5 MW, five-year Cerebras colocation at MTL-3, which was completed in October 2025 and has been billing since November 1, 2025 (claims 19577, 2170). The third is the Nscale master services agreement of November 22, 2025, a 10-year, 40 MW contract at NC-1 valued at roughly $865 million TCV (claims 3971, 18325, 2169). Contracts covering the 14 MTL-1 tenants and 5.5 MW at Blönduós are also described as executed, but none of those counterparties is named (claims 4575, 2174). The roughly 23-customer total is not broken out by counterparty (claim 24239).

The softer language centers on NC-1. In August 2025 the company presented an LOI with an unnamed 'AI cloud provider' and cited about $230 million over seven years (claim 24238). By November that had become a 're-marketing process' and 'closing stages of discussions' with unnamed counterparties (claim 19579), which implies the LOI did not convert. Nscale signed weeks later. Duke Energy is a power supplier, not a customer. It is repeatedly cited at 24, 40, and 99 MW, but only on a 'commercially reasonable efforts' basis (claims 4546, 2211, 2189, 2190). NVIDIA Preferred Partner status (claim 2171) and the Atlanta capacity lease (claim 2200) are also supply-side items rather than revenue.

Counterparty concentration has moved toward a single customer. Nscale's average of about $86.5 million a year is roughly seven times Cerebras's run rate. At the same time, a customer arrangement worth about $21 million of annualized cloud run rate began winding down after September 30, 2025 (claim 4125). The record does not identify that customer, so attributing it to the Initial Customer would be inference.

The arithmetic reconciles for Cerebras: CAD 1.4 million (about $979,000) a month over 60 months comes to $58.74 million (claim 2170). Nscale's $865 million includes escalators, so its first-year revenue will sit below the average. The power base is the larger issue. Forty MW of critical IT load requires more than the 40 MW gross that Duke targeted for April 2026 on a best-efforts basis. The mid-2026 initial deployment (claim 17566) has no confirmation of energization or billing on record as of this writing.

WhiteFiber's record contains two named, signed customer contracts. Cerebras holds a 5 MW IT-load colocation agreement at MTL-3 (4708, 24237). Its installation was completed in October 2025 (19577), and billing began November 1, 2025 at CAD 1.4 million a month (2170). Nscale signed a master services agreement on November 22, 2025 for 40 MW at NC-1, with roughly $865 million of contracted revenue over ten years (3971, 18325, 2169). Every other revenue relationship is either unnamed or not a customer. The 'Initial Customer' behind the 2,048-GPU three-year MSA is never identified (2202, 2218). Its dating also shifts, from a November 9, 2023 agreement (4718) to one 'finalized' December 12, 2023 (2202). The 14 MTL-1 tenants (4575), the 23-customer count (24239) and the 5.5 MW contracted at Blönduós (2174) are all anonymous. Duke Energy appears in a dozen duplicative claims (4710-4712, 4583-4585, 2211), but it is a power supplier, and its commitments are 'commercially reasonable efforts', not firm delivery. NVIDIA Preferred Partner status (2171) is a channel credential, not a demand signal.

NC-1 shows the gap between non-binding and signed language most clearly. In August 2025 the company cited an LOI with an unnamed AI cloud provider worth about $230 million over seven years for Phase 1 (24238). By November that had become a 're-marketing process' and 'closing stages of discussions with multiple highly creditworthy counterparties' (19579), which implies the LOI did not convert. Nscale signed weeks later.

Counterparty concentration has inverted. In 2023 the business rested on a single unnamed cloud customer. By mid-2025 it described a diversified tenant base. Now, named contracted value is overwhelmingly Nscale: $865 million against about $59 million for Cerebras, more than 90% of disclosed TCV. In the same period, a customer arrangement worth about $21 million of annualized cloud run rate began winding down after September 30, 2025 (4125). The filings do not say which customer that is.

The disclosed arithmetic mostly reconciles. Cerebras at about $979 thousand a month over 60 months gives about $58.7 million, consistent with claim 2170. That is about $2.35 million per MW-year. Nscale averages about $86.5 million a year, including escalators, or about $2.16 million per MW-year, which is a plausible range. The '$50' attached to the Initial Customer (2218) has no stated units and cannot be checked. Nscale revenue depends on an initial deployment 'by mid-2026' (17566) and on Duke reaching 40 MW by April 1, 2026 on an efforts basis (2189). Nothing on record confirms either milestone was met.

WhiteFiber's record contains only two named, signed customer contracts. Cerebras holds a 5 MW IT-load, five-year colocation contract at MTL-3 (4708, 24237). Installation finished in October 2025 (19577), and billing of CAD 1.4 million (about $979 thousand) a month began on November 1, 2025 (2208). Nscale signed a master services agreement through Enovum NC-1 Bidco on November 22, 2025 (3971, 2214) for a 10-year, 40 MW service order of about $865 million in total contract value. That figure includes annual escalators (18325, 2195, 2169), and deployment is expected only by mid-2026 (17566). Everything else is either unnamed or not a customer relationship. The 'Initial Customer' behind the cloud business has never been identified. Its master agreement is dated November 9, 2023 in some filings (4718, 4555) and December 12, 2023 in another (2218), a discrepancy worth resolving. MTL-1's 14 tenants (4575, 4538), the 23-customer headcount (24239) and the 5.5 MW of executed Blönduós contracts (2174) are likewise anonymous. The Duke Energy items (4710-4712, 4644-4646, 4546, 2164, 2190) are supply-side utility undertakings limited to 'commercially reasonable efforts', not revenue contracts. The SuperMicro, Dell and HPE ties (2172) are vendor authorizations.

The NC-1 anchor story is the clearest case of LOI language sitting next to signed language. In August 2025 the company cited an LOI with an unnamed 'AI cloud provider' expected to produce about $230 million over seven years for Phase 1 (24238). By November 2025 it described a 're-marketing process' and 'closing stages of discussions with multiple highly creditworthy counterparties' (19579). That wording implies the original LOI did not convert. Nscale signed ten days later, on different economics.

Concentration has shifted sharply. In mid-2025 revenue came mainly from the unnamed Initial Customer's GPU cloud, including a 2,048-GPU expansion (2204, 2219), plus fragmented MTL-1 colocation. By late 2025 an unidentified customer accounting for about $21 million of annualized cloud run rate was being wound down (4125). Disclosed backlog is now dominated by Nscale. Its $865 million dwarfs Cerebras's roughly $59 million five-year value, so one counterparty that is not yet operational holds most contracted dollars.

The arithmetic partly reconciles. Cerebras billing implies about $11.7 million a year, or about $2.35 million per MW, which fits a 5 MW load. Nscale's $865 million over ten years averages about $86.5 million a year, or about $2.16 million per MW before escalators. That is plausible, but it is unearned until delivery. The Initial Customer's dollar terms are truncated in the record (2218), so cloud revenue cannot be tied back to a named, verifiable contract.

WhiteFiber's record contains only two signed agreements with named customers whose economics are disclosed. The first is Cerebras at MTL-3, a 5 MW IT-load, five-year colocation contract (claims 4708, 24237). Installation was completed in October 2025 (19577), and billing has been CAD 1.4 million a month, about $979 thousand, since November 1, 2025 (2170, 2208). The second is Nscale, which signed a master services agreement on November 22, 2025 (3971, 2214) whose initial service order is worth about $865 million over ten years for 40 MW at NC-1 (18325, 2195, 17566). Everything else needs a discount. The 'first cloud customer' or 'Initial Customer' is never named, and its MSA is dated November 9, 2023 in one filing (4718) and December 12, 2023 in another (2218). The fourteen MTL-1 tenants (4575), the 23-customer count (24239) and the 5.5 MW contracted at Blönduós (2174) are also unnamed. Duke Energy appears repeatedly (4644-4646, 4583-4585, 2189-2190), but it is the utility, not a customer, and its commitment is only 'commercially reasonable efforts.' The NVIDIA, SuperMicro, Dell and HPE entries (2171, 2172) are vendor authorizations, not revenue relationships. The NC-1 story shows clearly how soft language can harden or fall away. In August 2025 the company cited an LOI with an unnamed AI cloud provider worth about $230 million over seven years for Phase 1 (24238). By November that had become a 're-marketing process' with 'multiple highly creditworthy counterparties' in 'closing stages of discussions' (19579). This implies the LOI lapsed before Nscale signed. Concentration has moved from one unnamed GPU-cloud customer, part of which is now being wound down at about $21 million of annualized run rate (4125), to Nscale. Nscale's contract value is roughly fifteen times the Cerebras contract. On the arithmetic, Cerebras reconciles: $979 thousand a month for 60 months is about $58.7 million, which matches the $58.74 million figure (2170). Nscale's $865 million works out to about $86.5 million a year including escalators, but that revenue depends on power. Duke's target is 40 MW gross by April 2026 (4711), while Nscale's contract specifies 40 MW of critical IT load. That suggests full delivery requires Duke to progress toward its 99 MW, four-year, best-efforts milestone, which sits uneasily with the mid-2026 operational guidance. The Initial Customer's '$50' figure (2218) has no stated units in the record, and the claim list provided is truncated at 2211.

WhiteFiber's record contains two signed, named customer contracts. Most of the remaining repeated claims are either unnamed counterparties or are not customer agreements at all. The clearest is Cerebras, whose 5 MW (IT load) colocation at MTL-3 moved from a permit-contingent 'will support' (4708, 4684) to completed installation and revenue generation in October 2025 (19577). Billing of CAD 1.4 million (about $979 thousand) per month runs over a five-year term (2208, 2183, 17567). The second is Nscale, whose master services agreement was signed on November 22, 2025 for the first 40 MW at NC-1. It runs ten years at roughly $865 million of total contract value (3971, 18325, 2214), with initial operation guided to mid-2026 (17566).

Duke Energy appears more often than any other name, but it is a power supplier, not a customer. Its 24, 40 and 99 MW milestones are 'commercially reasonable efforts' undertakings (4710, 4711, 4712, 2211), and the company itself labels Duke a 'Power Provider' (17577). The Blönduós and Atlanta arrangements are capacity WhiteFiber leases in (2199, 2200), and the SuperMicro, Dell and HPE relationships are vendor authorizations (2172). None of these produce revenue.

The cloud business rests on counterparties that are never named. The 'Initial Customer' contracted 2,048 GPUs, plus a further 2,048 in the second quarter of 2024 (2202, 2204, 2219). Its agreement is dated November 9, 2023 in earlier filings (4718) but described as finalized December 12, 2023 later (2202), a discrepancy worth resolving. MTL-1's 14 lessees are likewise unidentified (4575, 4538). In November 2025 the company was in 'closing stages' with 'multiple highly creditworthy counterparties' for an NC-1 anchor (19579), and Nscale was named ten days later. The same filing disclosed the wind-down of an unnamed arrangement worth about $21 million of annualized cloud run rate (4125). Whether that was the Initial Customer is not stated.

Concentration has therefore shifted from a single unnamed GPU customer toward Nscale. At about $86.5 million per year on a straight-line basis, Nscale's contract would outweigh Cerebras's roughly $11.7 million annualized by more than seven to one once it is live. The arithmetic only partly reconciles. The Cerebras monthly figure ties cleanly to its term, at about $59 million over five years. The '$50' associated with the Initial Customer (2218) cannot be verified from the truncated quote. No total cloud run rate is disclosed against which to size the $21 million loss. Nscale's revenue also depends on Duke reaching 40 MW, which is only a best-efforts commitment (4711).

WhiteFiber's record mixes signed revenue contracts with commitments that only resemble customer deals, and it has to be read in that light. The most frequently repeated items, the Duke Energy milestones of 24 MW by September 1, 2025, 40 MW by April 1, 2026 and 99 MW within four years (claims 4710, 4711, 4712, 4546, 2211), are supply-side utility arrangements, not customers. They bind Duke only to "commercially reasonable efforts," and nothing on the record confirms the 24 MW milestone was met. The same caution applies to the NVIDIA, SuperMicro, Dell and HPE authorizations (2171, 2172), the Blönduós partnership (2199) and the Atlanta lease (2200), where WhiteFiber is the lessee.

Only two counterparties are named and signed. The first is Cerebras, with 5 MW over five years at MTL-3 (4708, 17567). It went live in October 2025 (19577) and has billed CAD 1.4 million, about $979 thousand, a month since November 1, 2025 (2183). The second is Nscale, which signed a master services agreement on November 22, 2025 (3971) for 40 MW over 10 years at NC-1, valued at about $865 million (18325, 2169). Every other revenue source is unidentified. These include the "Initial Customer" for 2,048 GPUs, the 14 MTL-1 tenants on leases averaging about 30 months (4575, 4538), and the holders of the executed 5.5 MW at Blönduós (2174). Even the Initial Customer's start date is inconsistent: November 9, 2023 in claims 4718 and 4555, but December 12, 2023 in claims 2218 and 2202. Nscale itself was preceded by "closing stages of discussions with multiple highly creditworthy counterparties" (19579). That was LOI-grade language that converted within weeks, which is to the company's credit.

Concentration has inverted over the period. The original exposure was one anonymous GPU-cloud customer. The company then disclosed that it was winding down an unnamed arrangement worth about $21 million of annualized cloud run rate (4125). Contracted backlog now sits overwhelmingly with Nscale, at roughly $865 million against about $58.7 million for Cerebras. That makes Nscale's credit and its mid-2026 start (17566) the central diligence question.

The revenue arithmetic largely reconciles. Sixty months at $979 thousand equals the $58.74 million tagged in claim 2170, or about $2.35 million per MW-year. Nscale's $865 million over 10 years averages $86.5 million a year including escalators, about $2.16 million per MW-year, which is a plausible colocation discount for a larger anchor tenant. One risk remains: Nscale's 40 MW matches Duke's April 2026 milestone, so contracted revenue rests on a utility promise made on an efforts basis. Some database tags are unreliable and should not be used. Claim 17577's "$0" is really a $0.05/kWh power cost, claim 2218's "$50" is unexplained, and claim 2182 has no quote.

WhiteFiber's record contains only two signed, named customer contracts. The first is a five-year, 5 MW colocation contract with Cerebras at MTL-3, billing since November 1, 2025 (claims 4708, 19577, 17567, 2183). The second is a 10-year, 40 MW master services agreement with Nscale at NC-1, executed November 22, 2025 (claims 3971, 18325, 17566, 2169). Most of the other entries do not describe customers at all. The repeated Duke Energy items are supply-side capacity commitments in which the utility agreed only to 'commercially reasonable efforts' toward 24, 40 and 99 MW (claims 4710, 4711, 4712, 4546, 2211). They carry no revenue, and the $0 tag on claim 17577 is an extraction artifact of a ~$0.05/kWh power cost. The SuperMicro, Dell, HPE and NVIDIA items are vendor and channel authorizations (claims 2172, 2171). The Iceland and Atlanta items describe WhiteFiber leasing capacity as a tenant (claims 2199, 2200).

Several revenue-bearing counterparties remain unidentified. The cloud 'Initial Customer' is described only by its 2,048-GPU, three-year MSA. That agreement is dated November 9, 2023 in earlier filings (claim 4718) but 'finalized' December 12, 2023 in the March 2026 filing (claims 2218, 2202), and the record does not explain the discrepancy. The 14 MTL-1 tenants (claims 4575, 4538) and the 5.5 MW Blönduós contracts (claim 2174) are also unnamed. So is the customer whose wind-down removed roughly $21 million of annualized cloud run rate (claim 4125). The November 2025 language about 'closing stages of discussions with multiple highly creditworthy counterparties' for an NC-1 anchor (claim 19579) was a pre-signing description, and the Nscale signing a month later superseded it.

Counterparty concentration has shifted sharply. Through mid-2025, disclosed revenue rested on one unnamed GPU customer plus a diffuse MTL-1 tenant base. Of the roughly $924 million in named contracted value, about 94% now sits with Nscale ($865 million) and the remainder with Cerebras (about $58.7 million). The Initial Customer's three-year term also runs out around late 2026.

The Cerebras arithmetic reconciles. CAD 1.4 million, or about $979 thousand monthly, over 60 months yields the $58.74 million tagged in claim 2170, which implies about $2.35 million per MW-year. Nscale's $865 million over 10 years averages about $86.5 million annually, or about $2.16 million per MW-year. That figure includes escalators and so is front-end lighter, but it is broadly consistent with the Cerebras pricing. Two caveats matter. Nscale revenue depends on mid-2026 operation (claim 17566), which in turn depends on Duke's best-efforts 40 MW by April 1, 2026, and the record contains no confirmation of commencement. The $50 tag on claim 2218 is also not tied to any disclosed term.

WhiteFiber's record contains two named, signed customer contracts. The first is the Nscale master services agreement, executed November 22, 2025 by subsidiary Enovum NC-1 Bidco: a 10-year, 40 MW colocation order at NC-1 worth about $865 million in total contract value (claims 3971, 18325, 2195, 2214). The second is the Cerebras colocation contract at MTL-3, a 5 MW, five-year arrangement. Installation was completed in October 2025, and billing began November 1, 2025 at CAD 1.4 million per month (claims 19577, 17567, 2208). Several other contracts are signed but name no counterparty. These include the 'Initial Customer' GPU MSA for 2,048 GPUs and a follow-on 2,048-GPU order (claims 2218, 2202, 2204). The record dates that MSA to both November 9 and December 12, 2023 (claims 4718, 2218). The 14 lessees at MTL-1 are also unnamed (claim 4575), as is the 5.5 MW contracted at Blönduós (claim 2174). Duke Energy should not be read as a customer at all. It is the utility supplying power, and its 24/40/99 MW obligations are only 'commercially reasonable efforts' targets (claims 4712, 4710, 4711, 2211). The NVIDIA, SuperMicro, Dell and HPE items are vendor and channel authorizations, not revenue relationships (claims 2171, 2172). The only true discussions-stage language was NC-1's 'closing stages of discussions with multiple highly creditworthy counterparties' after a 're-marketing process' (claim 19579). Nscale resolved that ten days later. Still, the phrase 're-marketing' implies an earlier anchor arrangement that did not close. Concentration has moved from one unnamed cloud customer toward named colocation tenants. A wind-down of a customer arrangement worth about $21 million of annualized cloud run rate (claim 4125) shows the earlier concentration risk materializing. The record does not identify which customer that was. The disclosed figures broadly reconcile. Cerebras at CAD 1.4 million (about $979 thousand) per month over 60 months implies roughly $58.7 million, or about $2.35 million per MW-year. Nscale's $865 million over 10 years and 40 MW implies about $2.16 million per MW-year. That figure includes escalators, so early-year revenue will run lower (claim 2195). The two rates are consistent. The caveat is timing: Nscale's mid-2026 start (claim 17566) depends on Duke delivering 40 MW by April 1, 2026, which is a best-efforts commitment rather than a guaranteed one (claim 2189). The Initial Customer's dollar value (claim 2218) cannot be verified from the truncated quote.

WhiteFiber's record contains two signed, named customer contracts with disclosed economics, and most of its other commitments are either unnamed or not customer agreements at all. The first is Cerebras. The 5 MW IT-load colocation at MTL-3 was first described as subject to permits (4684, 4642, 4580). Installation was completed in October 2025 (19577), and billing began November 1, 2025 at CAD 1.4 million per month over five years (2183, 17567). The second is Nscale. Its master services agreement was signed November 22, 2025 (3971, 2214) for a 10-year, 40 MW term valued at roughly $865 million in total contracted revenue (18325, 2169, 2195). The recurring Duke Energy entries (4710-4712, 4546, 2164) are not customer agreements. They are utility capacity commitments made on a commercially-reasonable-efforts basis, and they recur across filings in a way that inflates the apparent count of counterparties. The NVIDIA Preferred Partner status (2171) is likewise a channel designation, not revenue.

Several material counterparties are never named. These include the Initial Customer under the three-year cloud MSA and its 2,048-GPU expansion (4718, 2204), the 14 MTL-1 lessees (4575), and the buyers of 5.5 MW at Blönduós (2174). The Initial Customer's agreement is dated November 9, 2023 in the 2025 filings (4718, 4555) but December 12, 2023, as amended, in the 2026 filing (2202). That discrepancy deserves clarification. The NC-1 anchor was disclosed in November 2025 only as closing-stage discussions with multiple unidentified creditworthy parties (19579). It became a named, signed contract with Nscale within weeks.

Concentration has shifted from an unnamed cloud tenant toward Nscale. On contracted value, Nscale's $865 million dwarfs Cerebras's roughly $59 million, giving it about 94 percent of named backlog. At the same time, an unnamed customer arrangement worth about $21 million of annualized cloud run rate began winding down after September 2025 (4125). That removes a revenue stream that exceeds Cerebras's entire annual billing.

The Cerebras arithmetic reconciles. CAD 1.4 million times 60 months gives about $58.7 million at the stated exchange rate (2170), with foreign-exchange exposure left open. Nscale's figure implies roughly $86.5 million a year on average, and that average is flattered by escalators (2195). Initial operation was guided for mid-2026 (17566), and the record contains no confirmation of commencement. More importantly, Nscale's 40 MW is critical IT load, while Duke's April 2026 tranche is 40 MW gross (2189). Full delivery therefore likely depends on the best-efforts 99 MW gross tranche, which runs to May 2029 (2190).

WhiteFiber's record contains two named, signed customer contracts with disclosed economics. Cerebras holds a 5 MW IT-load, five-year colocation agreement at MTL-3. It was first announced subject to permits (4708, 4684) and later confirmed as installed and generating revenue from October 2025 (19577, 17567). Nscale signed a 10-year, 40 MW master services agreement at NC-1 on November 22, 2025 (3971, 2214), with roughly $865 million in contracted revenue (18325, 2169). Duke Energy appears in more claims than any other counterparty, but it is a utility, not a customer. Its commitments are 'commercially reasonable efforts' capacity targets of 24, 40 and 99 MW gross (4644, 4645, 4646, 2164), and they carry no revenue. The NVIDIA Preferred Partner status (2171) is a channel designation, not a contract. Most other counterparties are unnamed. The cloud business rests on an 'Initial Customer' whose MSA is dated November 9, 2023 in one filing (4718) and December 12, 2023 in another (2202). That customer was later expanded by 2,048 GPUs (2204). MTL-1 is described only as 14 customers averaging about 30 months (4575), and Blönduós as 'executed contracts' for 5.5 MW (2174). The NC-1 anchor was pitched as being in closing 'discussions with multiple highly creditworthy counterparties' (19579), which is LOI-grade language. Five weeks later it resolved into the signed Nscale deal, a conversion worth crediting. Concentration has inverted. Revenue was once dominated by a single anonymous cloud customer. A roughly $21 million annualized cloud arrangement, also unnamed, began winding down after September 30, 2025 (4125). Contracted backlog is now overwhelmingly Nscale. Its $865 million dwarfs Cerebras's total, which is about $58.7 million (2170). The Cerebras figures reconcile: CAD 1.4 million, or about $979 thousand, per month over 60 months matches the $58.74 million tagged value (2183). Nscale averages about $86.5 million a year, or about $2.2 million per MW-year. That total includes escalators (2195), and the pass-through terms are cut off in the quote. It depends on 40 MW of critical IT load being live by mid-2026 (17566). Duke's April 2026 milestone is only 40 MW gross on a best-efforts basis (2189), which leaves timing and ramp risk that the headline TCV does not reflect. Cerebras billing replaces only about half of the $21 million run rate being wound down, so near-term reported revenue may dip before Nscale starts billing.

WhiteFiber's record contains two named, signed revenue contracts. Most of the remaining partnership language describes either suppliers or customers the company has not identified. The Nscale master services agreement, entered on November 22, 2025 by subsidiary Enovum NC-1 Bidco, is the most substantial. It covers a 10-year, 40 MW critical-IT-load colocation deal at NC-1 valued at approximately $865 million in total contracted revenue (claims 3971, 18325, 2169, 2214). The Cerebras 5 MW, five-year colocation contract at MTL-3 is the second, and it has been billing since November 1, 2025 (claims 19577, 17567, 2170). The Duke Energy entries are not customer agreements. They are utility capacity commitments, repeated across at least five filings, under which Duke will use only "commercially reasonable efforts" to reach 24, 40 and 99 MW gross (claims 4710-4712, 4546, 2164, 2189, 2190). The NVIDIA, SuperMicro, Dell and HPE references describe vendor and channel authorizations, not demand (claims 2171, 2172).

Much of the base remains anonymous. The "first cloud customer" under a November 2023 three-year MSA and its later 2,048-GPU expansion are identified only as the "Initial Customer" (claims 4718, 2204, 2219). MTL-1's 14 tenants are unnamed (claim 4575), as is the counterparty behind 5.5 MW of executed contracts at Blönduós (claim 2174). Before Nscale was named, NC-1 went through a period in which management described only being in the "closing stages of discussions with multiple highly creditworthy counterparties" after a re-marketing process (claim 19579). That wording suggests an earlier prospect fell away.

Concentration has shifted from one undisclosed GPU-cloud customer to a single named colocation tenant. Nscale's $865 million dwarfs Cerebras. In the same quarter, the company began winding down an unnamed arrangement worth roughly $21 million of annualized cloud run rate (claim 4125), which erodes the legacy cloud book.

The arithmetic reconciles cleanly for Cerebras: approximately $979 thousand a month over 60 months matches the $58.74 million figure (claims 2170, 2183). Nscale's headline averages about $86.5 million a year, but it includes annual escalators, so early-year revenue should run below that average. Initial deployment was guided only to mid-2026 (claim 17566). The power math also needs scrutiny. A 40 MW critical IT load cannot be served from Duke's 40 MW gross milestone once cooling and overhead are counted, so full Nscale delivery appears to depend on progress toward the best-efforts 99 MW tranche. The record provides no disclosed revenue total against which the $21 million wind-down and the unnamed cloud contracts can be reconciled.

WhiteFiber's record contains two named, signed customer contracts, and both carry their own economics. The first is Cerebras, a 5 MW IT-load, five-year colocation at MTL-3 (claims 4708, 17567, 2182). Filings first framed it as "previously announced" and "subject to our receipt of all required permits" (4684, 4642, 4580). Billing began November 1, 2025 at CAD 1.4 million, roughly $979 thousand, per month (2170, 2183). That run-rate multiplied over sixty months gives the $58.74 million tagged on claim 2170, so the Cerebras figures reconcile. The second is Nscale, whose master services agreement was signed November 22, 2025 (3971). Its initial service order is a 10-year, 40 MW critical-IT-load colocation at NC-1 worth about $865 million in total contract value, including escalators (18325, 17566, 2169). It is not yet producing revenue, with deployment expected by mid-2026 (17566). A large share of the remaining disclosures do not identify counterparties. The "Initial Customer" for 2,048 GPUs has never been named (2202, 2218). Its contract date is also inconsistent: November 9, 2023 in earlier filings (4718, 4555) and December 12, 2023 in the latest (2202). The fourteen MTL-1 tenants (4575, 4538) and the 5.5 MW of executed contracts at Blönduós (2174) are likewise disclosed only in aggregate. Several items look like customer relationships but are not. The Duke Energy capacity agreements are utility supply commitments on a "commercially reasonable efforts" basis (4710-4712, 2164, 2211), not revenue. NVIDIA, SuperMicro, Dell and HPE are vendor authorizations (2171, 2172). Iceland and Atlanta describe leased capacity that WhiteFiber takes in, not capacity it sells (2199, 2200). Concentration has moved from a single anonymous anchor cloud customer toward one named colocation tenant. Nscale's $865 million accounts for about 94% of the named contracted value. After September 30, 2025, WhiteFiber began winding down an unidentified arrangement worth about $21 million of annualized cloud run rate (4125). The filings do not say whether this is the Initial Customer. Two reconciliation gaps remain. First, the $50 figure attached to claim 2218 has no stated unit, so it cannot be tested against the 2,048-GPU contract or the $21 million wind-down. Second, Nscale's 40 MW of IT load would normally require more than the 40 MW gross that Duke committed by April 2026 (2189). The 2026 filings repeat the September 2025 24 MW target (2188) without confirming it was met. The record supplied ends at claim 2195, so later disclosures are not reflected here.

WhiteFiber's record is thinner on firm customer commitments than the volume of claims suggests. Most entries repeat the same few disclosures across successive filings. Only two counterparties are both named and tied to signed, priced contracts. Nscale signed a master services agreement on November 22, 2025 for a 10-year, 40 MW colocation deployment at NC-1, valued at roughly $865 million in total contracted revenue including escalators (claims 3971, 18325, 2169). Cerebras holds a 5 MW, five-year colocation contract at MTL-3 (claims 17567, 2182). In mid-2025 that contract was still described as 'subject to our receipt of all required permits' (claims 4684, 4642). By March 2026 it had become a billing relationship starting November 1, 2025 (claim 2183). Everything else is either unnamed or not a customer relationship. The 'Initial Customer' cloud MSA for 2,048 GPUs is never identified. Its contract date also shifts between November 9, 2023 (claims 4718, 4555) and a December 12, 2023 amended finalization (claims 2202, 2218). The 14 MTL-1 tenants (claims 4575, 4538) and the 5.5 MW of executed Blönduós contracts (claim 2174) carry no counterparty names either. Duke Energy is a power supplier, not a customer. Its NC-1 commitments are 'commercially reasonable efforts' targets, not firm obligations (claims 4710-4712, 2211). The NVIDIA, SuperMicro, Dell and HPE items are vendor authorizations that carry no revenue (claims 2171, 2172). Concentration has shifted sharply. Through mid-2025, revenue rested on the unnamed Initial Customer and a fragmented MTL-1 tenant base. After the Nscale signing, Nscale accounts for roughly 94% of named contracted value. Separately, a customer arrangement worth about $21 million of annualized cloud run rate began winding down after September 30, 2025, and that customer was also not named (claim 4125). The Cerebras arithmetic reconciles: roughly $979 thousand a month over 60 months equals the $58.74 million figure (claims 2170, 2183). The USD value will move with the CAD rate. Nscale works out to about $86.5 million a year, or roughly $180 thousand per MW-month. That sits close to the Cerebras rate of about $196 thousand, so the two contracts are internally consistent. The unresolved question is power. Nscale's first 40 MW of critical IT load is expected in mid-2026 (claim 17566), but Duke's April 2026 target is only 40 MW gross (claim 2189). After cooling and overhead, 40 MW gross cannot fully serve 40 MW of IT load. Full delivery therefore likely depends on the 99 MW phase, which carries a four-year, efforts-based horizon (claim 2190).

WhiteFiber's record contains two named, signed customer contracts with disclosed economics. Cerebras holds a five-year, 5 MW colocation contract at MTL-3 (claims 4708, 17567, 2182). It began billing on November 1, 2025 at CAD 1.4 million, roughly US$979 thousand, per month (claim 2208). The mid-2025 filings had qualified that deal as 'subject to our receipt of all required permits' (claims 4684, 4642, 4580). Nscale signed a master services agreement on November 22, 2025 for 40 MW at NC-1, valued at about $865 million in total contracted revenue over ten years, including escalators (claims 3971, 18325, 2195, 2214). Duke Energy appears repeatedly, but it is the power supplier, not a customer. Its capacity agreement commits only 'commercially reasonable efforts' toward 24, 40 and 99 MW (claims 4710, 4711, 4712, 2211, 2189, 2190), and the $0 tag on claim 17577 is a parsing artifact of a roughly $0.05/kWh energy cost. The NVIDIA, SuperMicro, Dell and HPE items are vendor authorizations, not revenue relationships (claims 2171, 2172).

Much of the remaining book is unnamed. The 'Initial Customer' behind the 2,048-GPU master agreement, and the 2,048-GPU add-on finalized in 2024, is never identified (claims 2202, 2204, 2218). The record also dates that agreement inconsistently: November 9, 2023 in earlier filings (claims 4718, 4555) and December 12, 2023 'as amended' in later ones (claim 2218). MTL-1's 14 tenants are anonymous (claims 4575, 4538). So is the counterparty on 5.5 MW of executed Blönduós contracts (claim 2174). The customer being wound down, representing about $21 million of annualized cloud run rate, is not named either (claim 4125). The record does not say whether that customer is the Initial Customer.

Concentration has shifted sharply. Through mid-2025, the business rested on one unnamed GPU cloud customer and a diffuse colocation tenant base. By early 2026, Nscale dominates. Its average of about $86.5 million per year outweighs Cerebras's roughly $11.7 million annualized billing by about seven to one.

The arithmetic reconciles where figures exist. Cerebras's monthly rate implies about $59 million over five years. Nscale's contract implies about $2.2 million per MW per year. However, the Initial Customer's contract value is truncated in the record at '$50' (claim 2218), and total cloud run rate is never disclosed, so the size of the $21 million wind-down relative to the whole cannot be tested. Nscale's mid-2026 start (claim 17566) also depends on Duke reaching 40 MW by April 1, 2026, and that milestone is a best-efforts target rather than a guarantee.

WhiteFiber's record contains two named, signed customer contracts, and both carry specific terms. The first is Cerebras at MTL-3: 5 MW of IT load over five years (claim 17567). It was first disclosed on the condition that permits were received (claims 4684, 4642), and it has since moved to billing. The site began billing on November 1, 2025, at CAD 1.4 million, or about USD 979 thousand, per month (claim 2208). The second is Nscale at NC-1: a 10-year, 40 MW master services agreement signed November 22, 2025, worth roughly $865 million of total contracted revenue (claims 3971, 18325, 2195). Much of what else is on record is not customer demand. The Duke Energy items (claims 4686, 4687, 4688, 2211) are a supply-side Capacity Agreement. It commits the utility only to commercially reasonable efforts toward 24, 40 and 99 MW, not to firm delivery. The NVIDIA, SuperMicro, Dell and HPE items (claims 2171, 2172) are vendor and channel authorizations, and the Blönduós and Atlanta entries (claims 2199, 2200) are capacity leases where WhiteFiber is the tenant.

The unnamed demand is large. The cloud business rests on an Initial Customer that is never identified, contracted for 2,048 GPUs plus another 2,048 (claims 2218, 2204). Its contract date also varies across filings: an MSA dated November 9, 2023 (claims 4718, 4555) versus one finalized December 12, 2023, as amended (claim 2202). MTL-1's 14 tenants (claim 4575) and the 5.5 MW contracted at Blönduós (claim 2174) are not named either. The $21 million annualized cloud run-rate arrangement being wound down (claim 4125) is also unattributed. The record does not say whether it is the Initial Customer, although that contract's three-year term makes it a plausible candidate.

Concentration has changed sharply. Before November 2025, contracted exposure rested on one anonymous GPU customer and a small Cerebras lease. Now Nscale's $865 million is more than ten times Cerebras's roughly $59 million five-year value, which makes WhiteFiber's backlog effectively a single-counterparty bet on NC-1.

The per-megawatt arithmetic is consistent. Cerebras works out to about $11.7 million a year, or $2.35 million per MW-year. Nscale averages $86.5 million a year, or $2.16 million per MW-year, with escalators included, so early years should run below that average. The reconciliation risk is timing, not pricing. Nscale's mid-2026 start (claim 17566) depends on Duke reaching 40 MW by April 1, 2026 on an efforts basis, and the latest filing still describes that level as a target (claim 2189). Meanwhile, the $21 million run-rate loss hits the cloud line before NC-1 revenue is documented. The Initial Customer dollar figure in claim 2218 is truncated in the record and cannot be checked.

WhiteFiber's signed customer record is narrower than its volume of announcements suggests. Much of the repetition in the filings is the Duke Energy NC-1 capacity agreement, restated across at least a dozen claims (4710, 4711, 4712, 2211, 2164). That agreement is a utility supply commitment, not a revenue contract. It is framed only as 'commercially reasonable efforts' toward 24 MW by September 2025, 40 MW by April 2026 and 99 MW by 2029. As of this writing, nothing on record confirms that the April 2026 milestone was met (2189). The NVIDIA, SuperMicro, Dell and HPE relationships (2171, 2172) are channel authorizations. The Iceland 'partnership' and the Atlanta capacity leases (2199, 2200) are arrangements in which WhiteFiber is the tenant, not the vendor.

Three customer contracts are both signed and named. Cerebras has a five-year, 5 MW colocation at MTL-3, billing since November 1, 2025 (2183, 17567). Nscale has a 10-year, 40 MW master services agreement at NC-1 dated November 22, 2025 (3971, 2169). The third is the unnamed 'Initial Customer' GPU agreement. That agreement is dated November 9, 2023 in earlier filings (4718) but described as 'finalized' December 12, 2023, 'as amended', in the latest (2202). This date drift deserves a footnote. The other contracted revenue is unidentified: 14 MTL-1 tenants averaging about 30 months (4575) and 5.5 MW contracted at Blönduós (2174).

Customer concentration has inverted. In 2023–2025 the business depended on a single unnamed cloud customer. After September 30, 2025 the company began winding down an arrangement worth roughly $21 million of annualized cloud run rate (4125), and it did not name that counterparty. Within weeks, Nscale's $865 million of total contracted value (18325, 2195) became the dominant item in the backlog, so concentration has shifted to Nscale.

The revenue arithmetic mostly reconciles. Cerebras at about $979 thousand per month over 60 months gives the $58.74 million recorded in claim 2170, although that figure fixes an exchange rate on a CAD 1.4 million monthly fee. Nscale's $865 million over 10 years and 40 MW works out to about $180 thousand per MW per month, consistent with Cerebras at about $196 thousand. Two things do not tie out. First, Nscale's 40 MW is critical IT load, while Duke's 40 MW milestone is gross, so the power commitment on record does not by itself cover the Nscale deployment once facility overhead is included. Second, no filing confirms that Nscale has started billing, even though operations were guided for mid-2026 (17566). Finally, the Initial Customer contract value appears only as a truncated '$50' (2218), so it cannot be reconciled against the $21 million wind-down.

WhiteFiber's record contains two signed, named customer agreements. The rest is unnamed customers, utility commitments, or duplicate restatements of those two. The first named deal is Cerebras at MTL-3: 5 MW of IT load on a five-year term, billing since November 1, 2025 at CAD 1.4 million (about USD 979 thousand) a month (claims 2208, 2182, 17567). Earlier filings conditioned it on permits (4684, 4642). The second is the Nscale master services agreement signed November 22, 2025 for 40 MW at NC-1, at roughly $865 million over ten years (3971, 18325, 2195, 2214). That figure includes rate escalators, and the quoted language cuts off mid-clause, so its full scope is unverified. Duke Energy appears most often, but it is the power supplier, not a customer. It committed only to 'commercially reasonable efforts' to reach 24, 40 and 99 MW gross (4710, 4711, 4712, 2211). That is best-efforts language, not a binding delivery obligation. The SuperMicro, Dell and HPE items (2172) are vendor authorizations and carry no revenue. The largest pre-2025 relationship has never been named. The 'Initial Customer' took 2,048 GPUs under an MSA, plus another 2,048 GPUs added in the second quarter of 2024 (2218, 2204). Its contract is dated November 9, 2023 in some filings (4718, 4555) and December 12, 2023 in others (2202). The 14 MTL-1 lessees (4575, 4538) and the 5.5 MW contracted at Blönduós (2174) are also unnamed. Counterparty concentration has shifted from one anonymous GPU cloud customer to one named colocation tenant. In November 2025 the company disclosed it was winding down an arrangement worth about $21 million of annualized cloud run rate (4125), without naming the customer. Nscale then became the whole of NC-1's first delivery. The arithmetic only partly reconciles. Cerebras works out to about $11.7 million a year, or roughly $59 million over the term. That does not replace the $21 million lost to the wind-down. Nscale averages about $86.5 million a year before escalator effects, but deployment is not expected until mid-2026 (17566). There is also a power question. Nscale takes 40 MW of IT load, while Duke's best-efforts target is 40 MW gross by April 2026. Once cooling and other overhead are counted, that gross figure likely leaves little or no headroom for 40 MW of IT load, and it is not firm. The record supplied here ends partway through a claim entry, and it includes no reported revenue totals to test these figures against.

Three counterparties are actually named, and only two of them are customers. Cerebras Wafer Scale ULC Systems has been identified since the first July 2025 disclosures (4708, 4684, 4642, 4580, 4543), where it was hedged as "subject to our receipt of all required permits," and is now a signed five-year, 5 MW colocation lease at MTL-3 with billing live since November 1, 2025 (17567, 2170, 2208, 2182). Nscale Services US Inc. and Nscale Global Holdings signed a master services agreement on November 22, 2025, with an initial Service Order of roughly $865 million over ten years for 40 MW at NC-1 (3971, 17566, 2195, 2214, 2169). Duke Energy appears in this register nine times but is the electric utility, not a revenue counterparty, and its language is the softest in the file: "commercially reasonable efforts" toward 24 MW, 40 MW and 99 MW (4710, 4711, 4712, 2188, 2189, 2190, 2211). That distinction matters because Nscale's 40 MW is precisely the Duke milestone dated April 1, 2026, so the largest contracted revenue line rests on an effort-standard utility commitment, and the March 2026 filing still restates that target rather than confirming delivery. NVIDIA appears only as a channel designation, "authorized NVIDIA Preferred Partner" (2171), not a customer.

Everything else is unnamed. The cloud business rests on an "Initial Customer" identified only as such across two and a half years (4718, 4695, 4555, 2202, 2204, 2218, 2219): 2,048 GPUs from December 2023, another 2,048 from Q2 2024. MTL-1's fourteen tenants (4575, 4538) and the 5.5 MW of executed Blonduos contracts (2174) are likewise anonymous.

Concentration has not diluted so much as rotated. It moved from one unnamed cloud customer, to a fragmented colocation tenant base, back to a single dominant name: Nscale is roughly 94% of disclosed contracted TCV against Cerebras' $58.7 million.

The arithmetic mostly reconciles. CAD 1.4 million monthly for sixty months is $58.74 million at the disclosed conversion, matching 2170, though that TCV is CAD-denominated and therefore FX-exposed. Nscale's $865 million implies $86.5 million annually, about $180 per kW-month, rich for wholesale colocation and explicitly escalator-inclusive, so early years run below that average and none of it reaches revenue before mid-2026. Unreconciled: the roughly $21 million annualized run-rate wind-down (4125) cannot be matched to any named customer.

WhiteFiber's record contains three identified commercial counterparties. The rest of its customer base is described but unnamed, and several of the most frequently repeated items are not customer contracts at all.

The firmest named deal is Cerebras. Mid-2025 filings described a "previously announced" 5 MW IT colocation agreement at MTL-3 that was "subject to our receipt of all required permits" (4684, 4642, 4580). By March 2026 the company reported that billing had begun on November 1, 2025, at CAD 1.4 million (about $979 thousand) a month over a five-year term (2208, 2183). The second named deal is the Nscale master services agreement signed November 22, 2025, at NC-1 (3971, 2214). Its initial service order carries about $865 million of total contracted revenue over ten years, including escalators (2195, 2169).

The largest historical cloud relationship is still anonymous. The "Initial Customer" MSA covers 2,048 GPUs over three years (2202, 2218), and an agreement for another 2,048 GPUs followed in the second quarter of 2024 (2204, 2219). Earlier filings date that first MSA to November 9, 2023 (4718, 4555), while the 2026 filing uses December 12, 2023. That discrepancy is unexplained. The 14 MTL-1 tenants (4575, 4538) and the counterparties behind 5.5 MW of executed Blönduós contracts (2174) are also unnamed.

The record contains no MOU or LOI language. The softer wording sits elsewhere. Duke Energy is a utility supplier, not a customer, and it committed only to "commercially reasonable efforts" to deliver 24, 40 and 99 MW to NC-1 (2211, 2188–2190). This language recurs across nearly every filing, and none of them confirms the September 2025 milestone was met. That matters because the Nscale revenue depends on NC-1 power. Blönduós and Atlanta are leases in which WhiteFiber is the tenant (2199, 2200), and the SuperMicro, Dell and HPE items are channel authorizations (2172).

Concentration has flipped. In mid-2025 the business rested on one unnamed GPU customer, 14 small tenants and 5 MW for Cerebras. After November 2025, a single named colocation tenant dominates. On a straight-line basis Nscale's contract averages about $86.5 million a year. Cerebras annualizes to roughly $11.7 million, about $59 million over its term. Meanwhile an unnamed arrangement worth about $21 million of annualized cloud run rate began winding down after September 30, 2025 (4125). The figures reconcile only partly. The Cerebras monthly rate matches its term, but Nscale's MW, start date and ramp are not disclosed. The $50 figure tagged to claim 2218 cannot be checked against its truncated quote.

WhiteFiber's record contains three categories of claims that need separating. Only a few are signed, revenue-bearing customer contracts. Some are supply-side agreements. The rest are vendor or channel relationships that generate no revenue.

The signed customer contracts with named counterparties are few. Cerebras holds a 5 MW (IT load) colocation agreement at MTL-3, billing since November 1, 2025 at CAD 1.4 million monthly over five years (claims 4708, 2182, 2183, 2170). Nscale signed a master services agreement at NC-1 on November 22, 2025, with an initial service order of approximately $865 million over ten years (claims 3971, 2169, 2195, 2214).

Everything else on the demand side is anonymous. The cloud business has always rested on an unnamed 'Initial Customer' (claims 4718, 2202, 2219, 2204). The record gives two dates for its MSA: November 9, 2023 in earlier filings and December 12, 2023, 'as amended', in the March 2026 filing. The difference may reflect signing versus finalization, but the filings do not explain it. MTL-1's 14 tenants (claims 4575, 4538) and the 5.5 MW of executed contracts at Blönduós (claim 2174) are also unidentified. The Duke Energy milestones of 24, 40 and 99 MW (claims 4710-4712, 2211, 2188-2190) are a utility capacity agreement written on a 'commercially reasonable efforts' basis. They are not a customer commitment. The March 2026 filing repeats the September 1, 2025 24 MW target without confirming it was met. The NVIDIA, SuperMicro, Dell and HPE relationships (claims 2171, 2172) and the Atlanta and Iceland capacity leases (claims 2199, 2200) are inputs to the business, not revenue.

Counterparty concentration has changed sharply. Through mid-2025 the company depended on a single unnamed cloud customer plus a fragmented MTL-1 tenant base. The wind-down of an arrangement worth roughly $21 million of annualized cloud run rate (claim 4125) points to a real contraction in that base. The counterparty is not named, and the record does not disclose total run rate, so its share cannot be measured. Nscale now dominates contracted value. Its average of about $86.5 million per year compares with about $11.7 million per year from Cerebras, which makes Nscale the main concentration risk.

The revenue arithmetic partly reconciles. The Cerebras figure is internally consistent: $979 thousand multiplied by 60 months equals $58.74 million (claim 2170), although the USD amount will move with the CAD exchange rate. The Nscale figure cannot be converted to annual revenue from this record. It includes escalators, so early-year revenue will fall below the straight-line average. The billing-commencement language is cut off in the excerpt. Its delivery also depends on NC-1 power arriving on Duke's efforts-based schedule, with no confirmed energization in the record.

WhiteFiber's record has fewer distinct agreements than its claim count suggests, because the same few contracts are restated across successive filings. Only two customer counterparties are both named and priced. The first is Cerebras, with a five-year, 5 MW IT-load colocation contract at MTL-3 (claims 4708, 2182). The second is Nscale, which signed a master services agreement on November 22, 2025 for NC-1 (claims 3971, 2214). The largest historical cloud relationship is never named. It is referred to only as the 'Initial Customer' under a three-year MSA for 2,048 GPUs (claims 2202, 2218), with a second 2,048-GPU order added in Q2 2024 (claim 2219). The execution date is inconsistent: earlier filings cite November 9, 2023 (claims 4718, 4555) and later ones cite December 12, 2023 (claim 2202). The 14 MTL-1 tenants (claims 4575, 4538) and the 5.5 MW of executed Blönduós contracts (claim 2174) are also unidentified.

The recurring Duke Energy items (claims 4710–4712, 4546, 2164, 2188–2190) should not be read as customer wins. They describe one utility Capacity Agreement, and it commits Duke only to 'commercially reasonable efforts' toward 24, 40 and 99 MW. That is best-efforts supply language, not a binding delivery obligation. It also matters for the revenue case, because NC-1's revenue depends on that power actually arriving, and the record does not confirm that the April 1, 2026 40 MW milestone was met. The NVIDIA, SuperMicro, Dell and HPE items (claims 2171, 2172) are vendor authorizations and generate no revenue.

Counterparty concentration has shifted sharply. Through mid-2025, cloud revenue rested on one unnamed GPU customer, alongside a fragmented colocation tenant base. In November 2025 the company disclosed that it was winding down an unnamed arrangement worth about $21 million of annualized cloud run rate (claim 4125). Contracted backlog is now dominated by Nscale. Nscale's $865 million (claim 2169) is roughly 94% of the combined value of the two named contracts.

The arithmetic partly reconciles. Cerebras billing of CAD 1.4 million a month (about $979 thousand) over 60 months gives the $58.74 million total in claim 2170, although the USD figure moves with the exchange rate. Nscale's $865 million over ten years averages about $86.5 million a year. That figure includes escalators, so early-year revenue will fall below the average, and the billing-commencement language is truncated in the record (claim 2169). The Initial Customer's contract value, flagged as $50 in claim 2218, is also truncated. As a result, the $21 million wind-down cannot be tied to that customer or checked against the GPU order book.

WhiteFiber's record contains two signed agreements with named counterparties. The first is Cerebras, a 5 MW IT-load colocation contract at MTL-3 (4708, 2182). The second is Nscale, whose master services agreement for NC-1 was signed November 22, 2025 (3971, 2214). Everything else is either unnamed or not a customer relationship. The largest cloud relationship is still called only the 'Initial Customer'. Its agreement is dated November 9, 2023 in one filing (4718) and December 12, 2023 as an amended Master Services and Lease Agreement in another (2218, 2202), a discrepancy the record does not explain. That agreement covers 2,048 GPUs, and a second 2,048-GPU supply agreement followed in Q2 2024 (2204, 2219). The 14 MTL-1 tenants (4575, 4538) and the Blönduós contracts for 5.5 MW (2174) are likewise unidentified. Duke Energy appears in more claims than anyone else (4710-4712, 4686-4688, 4644-4646, 4583-4585, 4546, 2164, 2188-2190, 2211), but it is the utility supplying NC-1, not a customer. Its 'commercially reasonable efforts' language is a best-efforts target, not a firm commitment. The March 2026 filing still repeats the September 2025 and April 2026 milestones without confirming either was met. SuperMicro, Dell/Advania and HPE are vendor authorizations (2172), and the Blönduós and Atlanta arrangements are capacity WhiteFiber leases from others (2199, 2200). None of these generate revenue. Concentration has shifted from one anonymous GPU customer to one named colocation tenant. In November 2025 the company disclosed it was winding down a customer arrangement worth about $21 million of annualized cloud run rate (4125), without naming the customer. Within weeks, Nscale's roughly $865 million, 10-year service order (2195, 2169) became the dominant contracted figure, dwarfing everything else on the record. That revenue depends on the NC-1 power ramp, which in turn depends on Duke's best-efforts schedule. The Cerebras numbers reconcile. CAD 1.4 million, or about $979 thousand, per month implies an exchange rate near 0.70 and works out to about $11.7 million a year, or $58.7 million over five years (2208, 2183). Nscale's $865 million averages $86.5 million a year, but it includes annual rate escalators and the quote is truncated, so early-year revenue is likely lower. The start date is unstated. The $50 tag on claim 2218 cannot be verified from the truncated quote. The record also does not say whether the $21 million wind-down is the Initial Customer, so cloud revenue after the wind-down cannot be reconciled to the announced GPU agreements.

WhiteFiber's disclosed customer base comprises three categories with materially different contract certainty. Cerebras (claims 2183, 2208, 2170) and Nscale Services (claims 2169, 2214) represent signed, named agreements with disclosed revenues: Cerebras commenced monthly billing at CAD 1.4M (~$979K USD) in November 2025 under a five-year contract totaling $58.74M (claims 2208, 2170); Nscale signed a master services agreement on November 22, 2025 (claim 3971) valued at $865M over ten years (claim 2169). By contrast, Duke Energy's three-phase capacity commitment (claims 4710, 4711, 4712, and repeated at claims 4688, 4646, 4644, 4645) uses the language "agreed to use commercially reasonable efforts," which signals an aspirational framework rather than a binding customer agreement—24 MW by September 1, 2025 remains contingent, not contracted. The "Initial Customer" identified only by that descriptor (claims 4718, 2218, 2202) signed a master services and lease agreement dated December 12, 2023 for 2,048 GPUs with an additional 2,048 GPU order in Q2 2024 (claim 2219), yet a material deterioration emerged: as of September 30, 2025, the company disclosed a wind-down of this arrangement representing approximately $21M of annualized cloud run rate (claim 4125). Fourteen unnamed customers occupy MTL-1's full capacity under lease agreements averaging 30 months (claims 4575, 4538), but no revenue is quantified. The concentration picture has materially shifted from 2023–2024, when the Initial Customer dominated, to 2026, where Nscale ($86.5M annualized over ten years) and Cerebras ($11.75M annualized) together represent disclosed contracted value exceeding the now-deteriorating Initial Customer relationship. Revenue reconciliation is incomplete: Nscale plus Cerebras yield approximately $98M in annual contracted run rate, while the unwinding of the $21M Initial Customer arrangement and the unquantified MTL-1 customer mix remain unresolved against any disclosed total-revenue guidance. The nomenclature also obscures customer identity—an unnamed "Initial Customer" representing the company's foundational relationship and fourteen MTL-1 occupants undermine transparency regarding true counterparty concentration.

WhiteFiber's disclosed customer base comprises three substantively different tiers of commitment. Only two named counterparties represent binding, executable service agreements: Cerebras and Nscale Services. Cerebras commenced billing in November 2025 at approximately CAD 1.4 million monthly (claim 2183, 2208), yielding roughly $58.7 million in contracted five-year revenue (claim 2170). Nscale Services, formalized November 22, 2025 (claim 3971), represents approximately $865 million in total contracted revenue over ten years inclusive of escalators (claims 2169, 2214). Both arrangements evidence active service delivery or binding contractual obligation. By contrast, Duke Energy's three phased capacity commitments—24 MW by September 1, 2025, 40 MW by April 1, 2026, and 99 MW within four years of May 16, 2025 (claims 4711, 4712, 4688, 4684, 4687, 4686)—rely consistently on the qualifier "use commercially reasonable efforts," which constitutes a softer standard than firm contractual obligation. These are utility capacity agreements, not customer offtake contracts. The company's most opaque counterparty is its "Initial Customer," described via Master Services and Lease Agreement signed December 12, 2023, covering 2,048 GPUs with a three-year term (claims 2218, 2202), supplemented by 2,048 additional GPUs in Q2 2024 (claim 2204). This customer remains entirely unnamed across all filings. The customer concentration shifted materially in late 2025: the Initial Customer had represented an annualized $21 million cloud run rate before entering wind-down status on September 30, 2025 (claim 4125). Nscale's November 2025 execution and Cerebras's November 2025 billing commencement partially offset this departure, though the Initial Customer's unnamed status prevents full revenue reconciliation. The company's publicly disclosed revenue arithmetic aligns with contract terms for Cerebras ($979K monthly × 60 months ≈ $58.7M) and Nscale ($865M over ten years), but historical dependence on an unidentified single customer creates significant opacity regarding total addressable revenue and customer concentration risk.

WhiteFiber's customer disclosure exhibits significant concentration among named counterparties coupled with material anonymity in its customer base. Only two counterparties are unambiguously identified with binding, executed service agreements: Cerebras (5 MW colocation, CAD 1.4 million monthly billing since November 2025, ~$58.7 million contracted over five years per claims 2170, 2208, 2183) and Nscale Services (Master Services Agreement signed November 22, 2025, $865 million over ten years per claims 2169, 2214). The Initial Customer, first secured in Q4 2023 through a Master Services Agreement dated December 12, 2023 (claims 4718, 2202), remains unnamed despite representing the earliest material engagement; its revenue contribution is undisclosed despite GPU commitments expanding from 2,048 to 4,096 units by Q2 2024 (claims 2204, 2219).

Duke Energy occupies a distinct category. All three capacity commitment phases—24 MW by September 1, 2025 (claims 4583, 4644, 2188); 40 MW by April 1, 2026 (claims 4584, 4645, 2189); and 99 MW within four years of May 16, 2025 (claims 4585, 4646, 2190)—are structured as "commercially reasonable efforts" obligations rather than binding service delivery commitments. This language, repeated across claims 4710, 4711, 4712, 4686, 4687, 4688, and 4546, signals conditional rather than executed arrangements.

MTL-1's 14 customers (claims 4575, 4538) and the Blönduós Campus partnership (claim 2199) remain entirely anonymous, with no counterparty identification or revenue disclosure despite stated full capacity occupation and ~$5.5 MW contracted IT load respectively. A customer arrangement wound down in September–December 2025 represented approximately $21 million of annualized cloud run rate (claim 4125), indicating material customer attrition concurrent with Nscale's November 2025 signing.

Revenue arithmetic does not fully reconcile disclosed announcements. Cerebras and Nscale together represent approximately $923.7 million in contracted value over their respective terms, yet Initial Customer revenue remains unquantified, the 14 MTL-1 counterparties undisclosed by name or individual value, and the departed customer's total contract value unspecified. Duke Energy's best-efforts framework precludes revenue recognition pending actual service delivery. Counterparty concentration appears extreme: two named entities represent the disclosed pipeline, suggesting either high dependency on Nscale or material reliance on unnamed customers whose stability is evidenced by the $21 million wind-down event.

WhiteFiber discloses three named, signed revenue-generating customers with material contract terms, alongside one undisclosed anchor customer now materially impaired. Nscale Services (master services agreement signed November 22, 2025) represents approximately $865 million over ten years; Cerebras (colocation agreement, commencing billing November 1, 2025) generates CAD 1.4 million monthly, or roughly $58.74 million over its five-year term; claims 2214, 2170, 2208 establish these as binding, performing contracts. The unnamed "Initial Customer," signed December 12, 2023, for GPU services valued at approximately $50 million, remains the only identified cloud customer from the founding period, yet since September 30, 2025, WhiteFiber has wound down this arrangement, eliminating approximately $21 million of annualized run rate (claims 2202, 2218, 4125). This wind-down is disclosed without explanation or counterparty identity.

Duke Energy occupies material disclosure real estate across ten claims spanning July 2025 through March 2026, but the relationship structure differs fundamentally. All Duke Energy commitments employ best-efforts language—"agreed to use commercially reasonable efforts"—for capacity milestones of 24 MW by September 1, 2025, 40 MW by April 1, 2026, and 99 MW within four years of May 16, 2025 (claims 2211, 2189, 2190). These are aspirational capacity targets, not binding customer agreements, and zero revenue is attributed to Duke Energy across the claim set. The repeated filing of identical Duke Energy language across multiple periods suggests neither the September 2025 nor April 2026 milestones have been publicly confirmed as achieved.

WhiteFiber names several technology partners—NVIDIA Preferred Partner (claim 2171), SuperMicro, Dell, and Hewlett Packard Enterprise (claim 2172)—but these are framework authorizations generating no disclosed revenue. Counterparty concentration is acute: Nscale alone represents approximately $86.5 million annualized contracted value; Cerebras adds $11.75 million annually. Disclosed revenue arithmetic cannot be independently reconciled against announced deals, as the Initial Customer identity remains undisclosed and the $21 million wind-down obscures historical revenue baselines. Actual billing performance against contracts is mentioned only for Cerebras (November 1, 2025 commencement) and Nscale (post-November 22, 2025), with no third-party verification of achievement against the disclosed contract values.

WhiteFiber's disclosed customer base reveals a material distinction between named signed contracts and conditional capacity commitments. The company has signed Master Service Agreements with exactly three identifiable counterparties: an 'Initial Customer' (undisclosed entity, claim 2202) for GPU cloud services via an MSA dated December 12, 2023, covering 2,048 GPUs with an additional 2,048-unit order in Q2 2024 (claims 2204, 2219); Nscale Services US Inc. and Nscale Global Holdings Limited for a ten-year colocation services agreement representing approximately $865 million in total contracted revenue (claim 2195); and Cerebras Wafer Scale ULC Systems for 5 MW of IT load capacity, with billing commencing November 1, 2025, at approximately $979,000 USD monthly (claim 2208). In contrast, Duke Energy's three-tranche capacity commitment—24 MW by September 2025, 40 MW by April 2026, and 99 MW within four years of May 2025—is consistently characterized by language obligating Duke to use 'commercially reasonable efforts,' not guaranteed delivery (claims 2164, 2189, 2190). This is materially softer than a binding service agreement. WhiteFiber also discloses that MTL-1's full capacity is occupied by fourteen customers under lease agreements averaging thirty months (claim 4538), yet none of these counterparties are named. Combined with Blönduós Campus partnerships (5.5 MW IT load, claim 2174) and technology partnerships with SuperMicro, Dell, and Hewlett Packard Enterprise (claim 2172), the unnamed customer base constitutes a concentration risk obscured from readers. Revenue arithmetic presents significant gaps. The Initial Customer's exact contractual value remains unspecified beyond term length; claimed GPU volumes do not translate to disclosed revenue. Cerebras generates approximately $11.7 million annually; annualized Nscale represents $86.5 million, making that single counterparty dominant post-November 2025. The fourteen MTL-1 customers contribute entirely undisclosed revenue. A claim from August 2025 notes the company began winding down a customer arrangement representing $21 million of annualized cloud run rate (claim 4125), suggesting material contract erosion not reflected in subsequent forward guidance. Reconciliation of announced capacity commitments with actual revenue remains opaque.

WhiteFiber's disclosed customer base separates cleanly into named, signed revenue contracts and vaguely attributed infrastructure relationships. Three counterparties carry documented, binding agreements with quantified revenue visibility: Cerebras (claim 2170, 2208), with a five-year colocation contract commencing billing November 2025 at CAD 1.4 million monthly, translating to approximately $58.7 million total contracted value; Nscale Services (claims 3971, 2169, 2195), with an $865 million ten-year master services agreement executed November 2025 for NC-1 colocation; and an unnamed "Initial Customer" (claims 2218, 2202), under a Master Services Agreement from December 2023 for GPU provision, with supplemental capacity agreements adding over three-year periods and total value suggested at approximately $50 million. By contrast, Duke Energy's capacity agreements (claims 4646, 4644, 4645) employ "commercially reasonable efforts" language, rendering them framework commitments rather than firm purchase orders. MTL-1's "14 customers under lease agreements with an average duration of approximately 30 months" (claims 4575, 4538) remains entirely anonymous, precluding meaningful concentration analysis. The company's partnership announcements (claim 2172) conflate technology authorizations—SuperMicro partnership, Dell CSP status, HPE official partnership, NVIDIA Preferred Partner status (claim 2171)—with revenue contracts; these represent supply-chain credentials, not customer agreements. Concentration risk appears acute: the three named operational customers represent the disclosed revenue base, while a "customer arrangement" wound down post-September 2025 accounted for $21 million in annualized run rate (claim 4125), unnamed and now dormant. The arithmetic suggests annual revenue in the $100–$115 million range from named sources (Cerebras $11.75M, Nscale $86.5M, Initial Customer $16.7M), but the absence of named customers for MTL-1's 14 lease arrangements and the recent $21 million contraction indicate both hidden concentration and deterioration masked by headline announcements. Critically, neither Duke Energy's phased capacity targets nor the Blönduós Campus GPU contracts (claim 2174) disclose binding revenue commitments, leaving the gap between announced capacity and actual contracted revenue substantial.

WhiteFiber's disclosed partnership portfolio distinguishes sharply between named, signed customer agreements and conditional utility-capacity frameworks. The company has executed three clearly identified customer service agreements: Cerebras (5 MW colocation, billing commenced November 2025 at CAD 1.4 million monthly), Nscale Services (November 2025 master services agreement representing $865 million over ten years at NC-1), and an unnamed "Initial Customer" (Master Service Agreement dated December 12, 2023 for 2,048 GPUs, amended to add 2,048 additional GPUs in Q2 2024). Beyond these, MTL-1's full capacity is occupied by 14 customers under lease agreements of approximately 30 months average duration as of May 30, 2025, but no counterparties are identified. Technology relationships with SuperMicro, Dell, and Hewlett Packard Enterprise are described as partnerships or authorizations rather than service agreements.

Duke Energy represents the most cited counterparty across claims but functions as a utility capacity provider, not a customer purchasing WhiteFiber services. The company entered a Capacity Agreement with Duke Energy containing three phased targets: 24 MW by September 1, 2025; 40 MW by April 1, 2026; and 99 MW within four years of May 16, 2025. Critically, all Duke Energy commitments employ the language "agreed to use commercially reasonable efforts to achieve" capacity, which denotes a conditional best-efforts framework rather than a binding service guarantee—language typically associated with letters of intent or pilot frameworks rather than firm contracts.

Counterparty concentration has shifted materially. Through mid-2025, dependency centered on an undisclosed "Initial Customer" and Duke Energy capacity commitments. The November 2025 Nscale agreement introduced substantial revenue ($865 million contracted over ten years) but represents a recent, single concentration point. Simultaneously, claim 4125 discloses that since September 30, 2025, the company commenced winding down a customer arrangement representing approximately $21 million of annualized cloud run rate—likely the Initial Customer referenced across earlier claims.

Revenue arithmetic presents a reconciliation gap. Cerebras contributes approximately $11.7 million annually (CAD 1.4 million monthly). Nscale's $865 million over ten years implies $86.5 million annually, though the November 2025 signature suggests limited revenue recognition in 2025. Set against the $21 million wind-down, the disclosed annualized revenue trajectory remains opaque without clarification of Initial Customer's actual contracted value and the timing of Nscale revenue recognition.

WhiteFiber's disclosed customer base comprises three materially significant named counterparties and a conspicuously unnamed primary customer, with contract quality varying substantially by counterparty. Cerebras Systems (claims 2183, 2170) represents a signed, billing colocation agreement generating approximately $979,000 USD monthly ($11.7M annualized) across a five-year term valued at $58.74M—the company's first demonstrably active revenue stream, commenced November 2025. Nscale Services (claims 3971, 2169) presents a November 2025 master services agreement representing $865M in contracted revenue over ten years, though as an "initial Service Order" language suggests expansion optionality rather than firm commitment. Duke Energy (claims 2211, 2188-2190) constitutes a capacity commitment agreement, not a customer revenue contract, with milestone language employing "commercially reasonable efforts"—softening actual obligation to three phased targets (24 MW by September 2025, 40 MW by April 2026, 99 MW within four years of May 2025), a construct materially weaker than signed service agreements.

The "Initial Customer" disclosed since Q4 2023 (claims 2202, 2218) remains deliberately unnamed despite representing 2,048 GPUs under a master services and lease agreement, with supplementary 2,048-GPU orders added Q2 2024. This anonymity constrains analyst visibility into the company's largest disclosed customer concentration; the counterparty's strategic importance is underscored by its explicit naming as both the first cloud customer and the basis for capital structure planning (sale-leaseback financing, claim 2219). Separately, MTL-1 capacity remains "fully occupied by 14 customers" (claims 4575, 4538) averaging 30 months' lease tenure, yet no counterparty identification or revenue reconciliation is provided.

Revenue arithmetic presents inconsistencies. Claim 4125 (November 2025) discloses that WhiteFiber commenced winding down a customer arrangement representing approximately $21M annualized cloud run rate—materially undercutting narrative continuity if this represents the unnamed Initial Customer. Against this $21M loss, Nscale ($86.5M annualized) and Cerebras ($11.7M annualized) would theoretically offset the departure, yet the timing suggests a customer transition rather than net growth. The repeated re-filing of Duke Energy capacity commitments, Cerebras colocation terms, and Initial Customer agreements across multiple claim dates (July through August 2025, then March 2026) indicates heavy restatement rather than partnership accumulation. Counterparty concentration has evolved from undisclosed Initial Customer dominance toward visible Nscale dependency, introducing single-counterparty revenue risk that contradicts infrastructure diversification claims.

WhiteFiber discloses three named operating counterparties with signed or active agreements: Duke Energy (capacity agreement with phased targets of 24 MW by September 2025, 40 MW by April 2026, and 99 MW within four years, using 'commercially reasonable efforts' language across claims 2211, 2188, 2189, 2190), Cerebras Systems (5 MW colocation at MTL-3, billing commenced November 2025 at CAD 1.4 million monthly—approximately $979K USD—per claim 2183), and Nscale Services (master services agreement signed November 2025, initial service order valued at $865 million over ten years, claims 2214, 2169). However, WhiteFiber's revenue base exhibits material opacity. The company's foundational customer—a cloud services counterparty secured under a Master Services Agreement dated December 12, 2023, for 2,048 GPUs over three years—remains unnamed across all filings (4718, 4695, 4555, 2218, 2202). This customer received an additional GPU order finalized in Q2 2024. Between September and December 2025, the company wound down an unnamed customer arrangement representing $21 million of annualized cloud run rate (claim 4125), suggesting either attrition from the unnamed Initial Customer or a separate counterparty.

MTL-1's full capacity is described only as occupied by "14 customers under lease agreements with approximately 30-month average duration" (claims 4575, 4538), with no counterparty names or revenue splits disclosed.

Duke Energy's capacity commitments employ conditional language—"agreed to use commercially reasonable efforts"—suggesting aspirational targets rather than firm offtake. By contrast, Cerebras and Nscale represent counterparties with quantified, active revenue. Technology partnerships listed (SuperMicro, Dell, HPE, NVIDIA per claim 2172) appear to be authorization or distribution relationships, not customer contracts.

Counterparty concentration has evolved from reliance on a single unnamed cloud customer (Q4 2023) through phased Duke commitments (2025–2029) to diversification via Cerebras and Nscale (November 2025 onward). The September 2025 customer wind-down signals operational portfolio shifts. Annualized run-rates from named contracts are Nscale (~$86.5M) and Cerebras (~$11.75M); the unnamed Initial Customer's revenue contribution and its relationship to the $21 million wind-down remain undisclosed. No reconciliation bridges announced power-delivery capacity to actual cloud-services revenue, nor is the composition of the "annualized cloud run rate" itemized against named contracts.

WhiteFiber's customer disclosures reveal a material distinction between named, binding agreements and contingent capacity commitments. Duke Energy, identified explicitly across filings (claims 4710, 4712, 4711, 4688, 4686, 4687, 4646, 4645, 4644, 4583, 4585, 4584, 4546, 2188, 2189, 2190, 2211, 2164), operates under a Capacity Agreement anchored in "commercially reasonable efforts" language—a conditional commitment to deliver 24 MW by September 1, 2025, 40 MW by April 1, 2026, and 99 MW within four years of May 16, 2025. Cerebras Wafer Scale ULC Systems (claims 4708, 4684, 4642, 4580, 4543, 2183, 2170, 2182), similarly named, operates an executed colocation contract for 5 MW IT load with billing commenced November 1, 2025, at CAD 1.4 million monthly—contractually $58.74 million over five years.

Conversely, WhiteFiber's "first cloud customer" (claims 4718, 4695, 4555, 2202, 2218, 2219) remains unnamed since Q4 2023, identified only as "Initial Customer." This counterparty secured 2,048 GPUs under a signed Master Service Agreement dated December 12, 2023, with an additional 2,048-GPU commitment in Q2 2024. MTL-1's 14 customer lease agreements (claims 4575, 4538) are entirely unidentified. The November 2025 Nscale Services Agreement (claims 3971, 2214, 2169, 2195)—representing $865 million over ten years—is named but constitutes colocation infrastructure services, not proprietary cloud revenue. Marketed partnerships with SuperMicro, Dell, HPE, and NVIDIA (claims 2172, 2171) are reseller authorizations, not customer contracts.

Counterparty concentration evolved from reliance on the unnamed Initial Customer (2023–2024) and Duke Energy's phased capacity targets (mid-2025) toward material dependence on Nscale's infrastructure outsourcing (late 2025). Disclosed revenue shows Cerebras arithmetic reconciliation: $979k monthly × 60 months equals $58.74 million, matching claim 2170. However, claim 4125 disclosed a September 2025 customer wind-down representing $21 million annualized run rate, indicating churn or capacity reallocation. Duke Energy's contingent "commercially reasonable efforts" language, Nscale's colocation focus rather than direct cloud services, and the unnamed Initial Customer's persistent opacity suggest WhiteFiber's material recurring revenue concentration remains undiversified and materially obscured from named counterparty disclosure.

WhiteFiber's public disclosure record contains minimal customer or partnership naming. The only substantive claim on file (claim 27, dated August 20, 2026) references "potential acquisitions, partnerships and joint ventures related thereto"—language that avoids contractual commitment entirely. The use of "potential" signals an exploratory phase, not signed agreements. The phrasing remains generic and unnamed; no counterparty is identified.

This absence is consequential. WhiteFiber has disclosed no named customer agreements, no identified strategic partners, and no anchor customers by designation. The silence on specific counterparties prevents any assessment of customer concentration, revenue per customer, or revenue attribution to announced relationships. Where transparency would typically require naming material customers or partners in SEC filings or press releases, WhiteFiber's record shows only forward-looking aspiration.

Without a historical series of substantive customer announcements, counterparty concentration trends cannot be evaluated. A single statement about "potential" partnerships provides no baseline to measure concentration risk or customer diversification over time. One would need successive announcements—naming specific partners, describing deal scope, disclosing revenue bookings—to track whether the company's customer base is broadening or narrowing, or whether customer concentration is increasing.

Revenue reconciliation is similarly impossible. WhiteFiber has made no revenue claims attributable to named partnerships or customers. The absence of deal-specific revenue disclosure (such as estimated contract value, annual recurring revenue tied to a named customer, or win announcements that cite revenue impact) means one cannot assess whether the company's total revenue or revenue growth aligns with its stated partnership activity.

The takeaway is measured: WhiteFiber's public disclosure on customers and partnerships consists of a single aspirational statement containing no named counterparties, no signed commitments, and no revenue attribution. For a commercial-diligence view, this represents either an early-stage company still in partner evaluation, or one with significant customer concentration or material partnerships that remain undisclosed. Until WhiteFiber names specific, signed customer agreements and ties revenue to those relationships, no meaningful commercial footprint can be verified from public statements alone.

WhiteFiber's NC-1 story has grown on paper while its near-term deliverable has been quietly re-denominated. Through summer 2025 the campus was pitched as 24 MW by Q1 2026, 40 MW by mid-2026, and 99 MW over four years (4713, 4653, 24230). An 'up to 200 MW' ceiling then appeared (4547, 24233) without new power commitments behind it. By December the deliverable had become 40 MW of IT load in two 20 MW phases against 99 MW of electrical capacity (3972, 18327). Measured as IT load rather than utility power, the first phase is 20 MW, not 24. The long-range target also moved: the 99 MW milestone tied to Duke was dated 2029-05-16 (4712, 4646), while the January 2026 framing carries a 2030-01-25 horizon (17571). That roughly eight-month push was never flagged as a revision. MTL-3 shrank slightly, from 7.0 MW site capacity (4706) to 6.6 MW at COD (24235). That looks like a gross-to-critical conversion rather than a cut.

On dates, the record is mixed but legible. MTL-2 is the clear miss. Its completion was given as end of 2025 in four separate disclosures (4647, 4577, 4541, 24234), and all are now marked delayed. An earlier July filing had already said January 2026 (4705), so the company was carrying two dates at once. NC-1's first 20 MW ready-for-service by April 30 is verified (18328, 3975), but the second 20 MW tranche due May 30 is delayed (18329, 3976). That leaves the January assurance of full Phase 1 delivery in April and May (17569) only half-honored.

Independent support is strongest where a utility is involved. Duke Energy's 24 MW and 40 MW power milestones are verified (4710, 4711, 24231, 24232), and the Nscale master services agreement (3971) supplies a named tenant. Nothing in the record covers permits, interconnection queue positions or on-the-ground reporting for NC-2/NC-3, Atlanta, Paris or Blonduos. Those remain project names without capacity or dates. The supplied claim history also stops in late January 2026, so any acknowledgement of the second-phase slip after that point is not visible here.

Over the next 90 days, the key question is whether the second 20 MW NC-1 phase is confirmed as billing, and on what date. The next is whether the year-end 2026 target of 76 MW gross online (24228) is reaffirmed, trimmed or silently dropped. Arithmetic on the disclosed sites makes that figure demanding. Also watch for a firm MTL-2 COD, any Duke filing that pins Phase 3 timing, and the first capacity or permit disclosure for any announced-only site.

WhiteFiber's capacity story has expanded in ambition while its near-term deliverables have contracted. In July 2025, NC-1 was framed as a 24 MW initial build growing to 40 MW by mid-2026 (4689, 4691, 4653), backed by a 99 MW Duke Energy commitment (4712). By August, the ceiling had become a potential 200 MW (4547, 24233), and the company guided to 76 MW gross online by year-end 2026 across a 115 MW four-facility portfolio (24228, 24229). The tranches that actually carry dates have moved the other way. When the Nscale master services agreement arrived in December 2025 (3971), the 24 MW first phase due by March 31, 2026 (4713, 4652, 24236) became a 20 MW first phase ready for service by April 30 (18328), with a second 20 MW block by May 30 (18329). Total phase-one capacity was preserved at 40 MW (3972), but the first block shrank by four megawatts and moved back a month. Only a month earlier, the November update still described 24 MW as on schedule for early 2026 (19578). We could not find any disclosure that acknowledged this re-cut. Revenue start had already drifted from May 1 (4714, 4690) to May 31 (4651, 4549) between filings dated two weeks apart. MTL-2 is the clearer slip. It was guided to Q4 2025 COD repeatedly (4647, 4577, 24234), yet the project is still recorded as announced, not operational, more than nine months later. MTL-3's 7 MW site likewise resolved to a 6.6 MW COD figure (24235). The record contains no independent corroboration: no building permits, no Duke interconnection queue entries, and no site reporting. Every dated claim from 2025 and early 2026 remains marked not_yet_verifiable, even though all of those dates have now passed. Before any read of pace is credible, the ledger needs its past-due claims resolved. NC-1's partially_operational status suggests at least one Nscale phase is energized, but that status is not tied to a dated claim. Over the next 90 days, watch for four things. First, whether Q3 2026 results confirm billing on both 20 MW NC-1 phases (3975, 3976). Second, whether management reaffirms or quietly drops the 76 MW year-end target (24228). Third, any Duke filing that shows progress toward the 99 MW tranche. Fourth, whether NC-2/NC-3, Atlanta, Paris or Blonduos acquire dated milestones or remain undated pipeline. Note that the supplied record is truncated at claim 3976.

WhiteFiber's NC-1 story is one of expanding ambition paired with shrinking near-term delivery. In July 2025 the site was framed as a 24 MW first phase growing to 40 MW and then 99 MW under Duke Energy commitments (4710, 4711, 4712). By August the ceiling had risen to a potential 200 MW (4547, 24233). In December 2025 the first deliverable changed shape: the 24 MW initial phase that had been promised for Q1 2026 (4713, 4652, 24236) became a 20 MW ready-for-service tranche due April 30, 2026, followed by a second 20 MW tranche on May 30 (18328, 18329, 3975, 3976). That was a capacity trim and a one-month slip, and the record contains no explicit acknowledgement of either. As late as November the company still called the 24 MW phase on schedule for early 2026 (19578). Revenue start also drifted within weeks, from May 1 (4714) to May 31, 2026 (4651), and the original September 2025 Duke power milestone (4710, 24231) passed without confirmed delivery in the record. Montreal shows the same pattern at smaller scale. MTL-2 and MTL-3 were first guided to January 2026 (4705, 4707) and then pulled forward to year-end 2025 (4647, 4648, 24234, 24235). MTL-3 was quietly restated from 7.0 MW (4706) to 6.6 MW (24235). The 16 MW year-end 2025 fleet target (4649, 24227) is flagged delayed. Paris, Blonduos, Atlanta and NC-2/NC-3 carry no capacity, cost or date claims at all, so they should be read as optionality rather than pipeline. On independent evidence, nothing in this record corroborates the announced pace. It includes no Duke interconnection queue filings, no Guilford or Yadkin County permits, no Quebec occupancy records and no site reporting, so every delivery here rests on company statements. The claim list is also truncated after 3971, so later restatements may exist that are not reflected here. Over the next 90 days, the key test is the 76 MW gross-online target for year-end 2026 (24228), which looks very difficult to hit unless both NC-1 tranches are already billing and MTL-2/MTL-3 are confirmed operating. Watch the Q3 2026 results for explicit confirmation of NC-1 billing commencement and a restated year-end figure. Watch Duke filings for the 99 MW path, and county records for any NC-2/NC-3 permitting that would turn the 115 MW four-facility framing (24229) into something verifiable.

WhiteFiber's stated ambition grew steadily through 2025 while the near-term deliverable got smaller. NC-1 was first framed as a 24 MW initial phase inside a Duke Energy ramp to 40 MW and then 99 MW (4710, 4711, 4712). By August it carried a 200 MW potential ceiling (4547, 24233) and sat inside a four-facility, 115 MW portfolio (24229) with 76 MW gross targeted online by year-end 2026 (24228). In the December 2025 disclosures, however, the first deliverable became a 20 MW ready-for-service block and the facility was described as 40 MW of IT load in two phases (3972, 18328, 18329). The 24-to-20 change may be a shift from utility megawatts to IT megawatts rather than a cut, but it was not reconciled publicly. MTL-3 also moved quietly from 7.0 MW (4706) to 6.6 MW (24235). Paris, Blonduos, NC-2/NC-3 and Atlanta are listed as projects with no capacity or date claims at all.

Almost every date has passed, and most slipped without being acknowledged. The Montreal completions were first given as January 2026 (4705, 4707), then restated as Q4 2025 (4647, 4648, 24234, 24235), and none of the material on record confirms either date. For NC-1, the 24 MW power date of September 2025 (4710, 24231) lapsed. The November assurance of early-2026 delivery (19578) gave way within five weeks to an April 30 ready-for-service date (18328), a month after the earlier Q1 2026 completion promise (4713, 24236). Revenue start drifted from May 1 to May 31 (4714, 4651). The record is also cut off partway through a 2025-12-17 entry, so roughly ten months of statements, including any 2026 restatements, are absent from what was supplied.

The record contains no independent corroboration. The dated items are transactions: the Greensboro-area property purchase (4643, 4685), the MTL-2 and MTL-3 acquisitions (4640, 4682), and the Duke earnout (24240). The Duke capacity schedule is a commercial commitment, not evidence of interconnection progress. To support the announced pace, the next things to check are Rockingham County building and certificate-of-occupancy records for 805 Island Drive, Duke large-load and substation filings, and Hydro-Québec allocations for the Saint-Jérôme and Montreal sites.

Over the next 90 days, the Q3 2026 results should show whether billing has actually started on both NC-1 20 MW phases (3975, 3976), whether the 76 MW year-end 2026 target (24228) is reaffirmed, quietly dropped, or redefined, whether MTL-2 and MTL-3 are reported as revenue-generating, and whether any of the claim-less projects acquire dated commitments before NC-1 is shown to be complete.

WhiteFiber's announced footprint has expanded much faster than its delivered one. NC-1 was first framed as a 24 MW initial build with a 40 MW phased expansion (4689, 4691). It was then anchored to a 99 MW Duke Energy commitment (4712, 4688), and by August 2025 it was described as having potential up to 200 MW (4547, 24233). Portfolio language followed the same pattern, with four facilities totaling 115 MW (24229) and 76 MW gross online by year-end 2026 (24228). The project roster has since added Paris, Blonduos, NC-2/NC-3 and Atlanta, none with a stated capacity. The only figure that moved down was MTL-3, which slipped from 7.0 MW (4706) to 6.6 MW (24235) without comment.

The dates have not held. MTL-2 and MTL-3 were first guided to January 2026 (4705, 4707). Weeks later they were pulled forward to year-end 2025 (4647, 4648, 24234, 24235), and both targets are now flagged delayed, as is the 16 MW year-end 2025 portfolio guide (4649, 24227). NC-1 shows quieter drift. The Q1 2026 energization of 24 MW (4713, 4652, 24236) was reaffirmed as on schedule in November (19578). In December it became two 20 MW tranches with ready-for-service dates of April 30 and May 30 (18328, 18329, 3975, 3976), which pushed the first block back by a month and cut it by 4 MW. The record shows no acknowledgement that this was a revision. Revenue start also moved from May 1 (4714) to May 31 (4651). The first Duke milestone of 24 MW by September 2025 (4710, 24231) is marked delayed, which matters because every later NC-1 date depends on it.

The record contains no independent corroboration of any of this: no building permits, no Duke interconnection queue filings, and no site reporting. The Duke figures are company-reported, and the claim history provided ends in December 2025, so 2026 statements could not be assessed. The pace should therefore be treated as unverified rather than as on track.

Over the next 90 days, the key test is whether the 76 MW year-end 2026 target (24228) is reaffirmed, trimmed, or dropped in the Q3 report. Specifically, look for an explicit statement on whether both 20 MW NC-1 tranches are energized and billing, confirmation that MTL-2 and MTL-3 actually reached commercial operation, and any Duke-side or county-level documentation of the 40 MW energization (4687, 24232). Also watch whether the new sites gain capacity figures before NC-1 demonstrates delivery.

WhiteFiber's capacity narrative has grown much faster than its delivered megawatts. NC-1 was introduced as a 24 MW initial build tied to a Duke Energy commitment (4710, 4689). It then became a 40 MW cumulative target (4653, 4691), a 99 MW four-year commitment (4712, 24230), and by August 2025 a site with up to 200 MW of potential (4547, 24233). Portfolio framing grew the same way, from 16 MW gross online by year-end 2025 (4649, 24227) to 76 MW by year-end 2026 (24228) and four facilities totaling 115 MW (24229). The one move in the other direction is small but notable: MTL-3 was carried at 7.0 MW through mid-2025 (4706, 4641, 4578) and then appeared at 6.6 MW in the August COD claim (24235), with no stated reason.

The dates have not held. The first Duke tranche of 24 MW by September 2025 (4710, 24231) is marked delayed, and so is every NC-1 completion and revenue date since then. The December 2025 restructuring matters most here. The 24 MW Q1 2026 initial phase (4713, 4652, 24236) was recast as two 20 MW ready-for-service tranches due April 30 and May 30, 2026 (18328, 18329, 3975, 3976). That one change cut the first-phase figure and pushed the dates back. It came a month after management called the 24 MW phase on schedule for early 2026 (19578), and the record shows no acknowledgement of a slip. Montreal followed an unusual path. MTL-2 and MTL-3 completion moved from January 31, 2026 (4705, 4707) to December 31, 2025 (4647, 4648, 24234, 24235), an earlier date that was then missed. The 16 MW year-end 2025 guidance depended on those two sites and is also delayed.

The record contains no independent corroboration: no permits, no interconnection queue positions for the 99 MW Duke tranche, no construction reporting. Paris, Blönduós, NC-2/NC-3, and Atlanta appear as announced projects but have no capacity or timeline claims at all. The claim list provided here also cuts off at 3971, so later revisions may be missing.

The next 90 days end on the 76 MW year-end 2026 target (24228). That claim is still marked stated, but meeting it would require NC-1's full 40 MW plus both Montreal expansions to be energized and billing. Watch whether the next quarterly report confirms billing on either 20 MW NC-1 tranche, whether the 76 MW figure is quietly restated, whether MTL-3's 6.6 MW is explained, and whether any filing names a Duke interconnection milestone beyond 40 MW.

WhiteFiber's capacity story has expanded on paper while the deliverable tranches have shrunk. In July 2025, NC-1 was described as a 24 MW initial build (4689) stepping to 40 MW (4691) under a 99 MW Duke Energy commitment running to 2029 (4688, 4712). Within weeks, management added a potential ceiling of 200 MW (4547, 24233) and a portfolio framing of 115 MW across four facilities (24229). Near-term commitments moved the other way. The December 2025 Nscale agreement recast the first NC-1 phase as 20 MW (18328) rather than 24 MW (4713). In Montreal, MTL-3 drifted from 7.0 MW (4706, 4578) to 6.6 MW (24235) without comment, while MTL-1 has stayed at 4 MW throughout (4702, 4537).

On timing, the record shows repeated quiet resets rather than acknowledged slips. MTL-2 and MTL-3 were first dated to January 31, 2026 (4705, 4707), then pulled forward to year-end 2025 (4647, 4648, 24234). Both are now flagged delayed, so the earlier date was the more honest one. The NC-1 revenue start moved from May 1 (4714) to May 31, 2026 (4651) between two July filings. The Duke 24 MW energization targeted for September 2025 (4710, 24231) became 'early 2026' by November (19578). It was then superseded by an April 30, 2026 ready-for-service date for a smaller 20 MW block (18328), and the second 20 MW moved to May 30 (18329). The year-end 2025 target of 16 MW (4649, 24227) is also flagged delayed. Every dated NC-1 and Montreal milestone in this record is past due as of October 2026 with no confirmed delivery logged. The record is truncated after 18327, however, so later confirmations may exist and should be checked before publication.

Independent corroboration is thin. The Duke figures are company-reported (4644, 4645, 4646), and the record holds no interconnection queue entries, building permits, or site reporting. Paris, Blonduos, Atlanta and NC-2/NC-3 have no capacity, cost or dated claims, so they are pipeline names, not projects. Acquisition claims flagged delayed against dates earlier than the filings themselves (4640, 4685) look like retrospective closing dates, not slippage.

Over the next 90 days, the key test is the Q3 2026 report. It should state whether Nscale billing has begun on both 20 MW phases (3975, 3976), whether MTL-2 and MTL-3 are generating revenue, and whether the 76 MW year-end 2026 guidance (24228) is reaffirmed. Watch also for Duke or NCUC documentation of actual NC-1 energization and Yadkin County permitting for NC-2/NC-3.

WhiteFiber's headline capacity has grown far faster than its delivered capacity. NC-1 was first framed as a 24 MW initial build (4689) with a 40 MW phased expansion (4691). It then took on a 99 MW Duke Energy commitment running to 2029 (4712, 18327) and a potential ceiling of 200 MW (4547, 24233). At the portfolio level the company described four facilities totaling 115 MW (24229) and guided to 16 MW gross online by year-end 2025 and 76 MW by year-end 2026 (24227, 24228). The smaller numbers have quietly shrunk. MTL-3 went from 7.0 MW (4706, 4641) to 6.6 MW at commercial operation (24235). NC-1's 24 MW first phase was recut into two 20 MW tranches tied to the Nscale agreement (3971, 18328, 18329), and the company did not explain the change.

Dates have moved in both directions without much acknowledgement. MTL-2 and MTL-3 were promised for January 31, 2026 in one July filing (4705, 4707). Weeks later they were pulled forward to year-end 2025 (4647, 4648, 24234, 24235), and both now carry delayed status. The first Duke tranche of 24 MW by September 2025 (4710, 24231) and the 16 MW year-end 2025 target (4649) are also marked delayed. On NC-1, the company said in November 2025 that it was on schedule for early 2026 (19578). One month later, the ready-for-service date became April 30, 2026 for a smaller 20 MW block (18328), and the revenue start had already drifted from May 1 to May 31 (4714, 4651). Every NC-1 delivery milestone is now marked delayed, and the record contains no restated dates to replace them.

There is no independent corroboration in the record: no building permits, no interconnection queue filings, no inspection reports, and no third-party site reporting. The Duke figures are commitments as described by WhiteFiber, not utility-published service dates. Paris, Blonduos, Atlanta and NC-2/NC-3 have no capacity or timeline claims at all. That gap matters most for Yadkin County, which would carry much of any scaling beyond NC-1.

Over the next 90 days, the year-end 2026 target of 76 MW (24228) is the clearest test. Watch for disclosure of whether Nscale billing actually began on either 20 MW tranche (3975, 3976), for restated commercial operation dates for MTL-2 and MTL-3, and for any Duke or North Carolina filings confirming energized load at NC-1. If the next quarterly report reaffirms the 99 MW and 200 MW figures without dating the delayed tranches, that would point to a widening gap between ambition and delivery.

WhiteFiber's capacity narrative has expanded on paper while the deliverable increments have quietly narrowed. NC-1 was introduced as a 24 MW initial build (4689, 4713) rising to 40 MW (4691, 4653) and ultimately 99 MW under the Duke Energy schedule (4712, 4646). By August 2025 a 200 MW site potential had been added (4547, 24233), and the portfolio was framed as four facilities totaling 115 MW (24229). The executable phases went the other way. The December 2025 Nscale agreement (3971) restated NC-1's first tranche as 20 MW ready-for-service by April 30, 2026 (18328), not 24 MW energized in Q1 2026 (4652, 24236). Billing on the second 20 MW was set for May 30, 2026 (3976). In Montreal, MTL-3 slipped from 7.0 MW (4706, 4641) to 6.6 MW (24235). Neither the NC-1 rescoping nor the MTL-3 trim was flagged as a change.

Dates have drifted in the same quiet way. NC-1 was described as on schedule for early 2026 in November 2025 (19578), a month before the April and May targets appeared. MTL-2 and MTL-3 completion moved between January 31, 2026 (4705, 4707) and December 31, 2025 (4647, 4648) across filings weeks apart, with no explanation. The year-end 2025 target of 16 MW gross online (4649, 24227) depended on both Montreal sites reaching commercial operation on time. Nothing in the record confirms that they did. The Duke 24 MW tranche due by September 2025 (4710, 24231) has no subsequent confirmation either.

Independent corroboration is absent. The record contains no permits, no Duke interconnection or load-study entries, no Quebec utility filings, and no on-the-ground reporting. The announced pace therefore rests entirely on company statements. Paris, Blönduós, Atlanta and NC-2/NC-3 carry no capacity or cost figures at all. The company's claims stop in December 2025, so every NC-1 milestone has since passed without a confirming statement in this dataset.

Over the next 90 days, the Q3 2026 results are the key test. They should show whether Nscale billing began on both 20 MW phases and whether MTL-2 and MTL-3 are generating revenue. They should also show whether the year-end 2026 target of 76 MW gross (24228) is reiterated, trimmed, or dropped. Beyond the results, watch for Rockingham and Yadkin County building permits, Duke load-interconnection milestones for the 40 MW step, and any move to attach megawatts or dates to the unquantified sites.

WhiteFiber's headline capacity has grown steadily on paper. NC-1 began as a 24 MW initial build within a 40 MW phased plan (4689, 4691). It was then framed against a 99 MW Duke Energy commitment (4712) and a 200 MW potential ceiling (4547, 24233). Fleet guidance rose from 16 MW by year-end 2025 (4649, 24227) to 76 MW by year-end 2026 (24228), against a 115 MW four-facility total (24229). The one direction of shrinkage is modest: MTL-3 moved from 7.0 MW (4706, 4578) to 6.6 MW at COD (24235) without comment.

The date record is weaker. NC-1 revenue start moved from 1 May 2026 (4714, 4690) to 31 May 2026 (4651, 4549) within eighteen days in July 2025, with no acknowledgment. In November the company described the initial 24 MW as on schedule for early 2026 (19578). The December Nscale agreement (3971) then quietly re-tranched the build into two 20 MW phases with ready-for-service dates of 30 April and 30 May 2026 (18328, 18329, 3975, 3976). That pushed the old Q1 2026 energization target (4652, 24236) into the second quarter and replaced the 24 MW figure with 20 MW. The Duke 24 MW tranche promised by September 2025 (4710, 24231) carries a delayed flag across every restatement. That matters because utility power gates everything downstream. The Montreal dates are internally inconsistent: MTL-2 and MTL-3 appear as 31 January 2026 in early July (4705, 4707), then as Q4 2025 weeks later (4647, 4648, 24234, 24235). Both are flagged delayed.

The record contains no independent corroboration. There are no permits, no interconnection-queue entries and no site reporting for NC-1, Montreal, or the newer Paris, Blönduós, Atlanta and NC-2/NC-3 entries, which have no capacity claims at all. The delivery pace therefore rests entirely on company statements. The claim list also ends mid-record at 3973, so later restatements may be missing.

Over the next 90 days, the third-quarter report should show whether Nscale billing actually began on either 20 MW phase. Both dates have now passed, and that is the clearest test of whether NC-1 exists as revenue-generating capacity. Watch for any reset of the 76 MW year-end 2026 figure, which already requires capacity beyond NC-1's 40 MW and the 15.6 MW Montreal base. Look for confirmation that MTL-2 and MTL-3 are energized and billing. Finally, look for any Duke energization disclosure or North Carolina permitting record that would supply the first independent evidence for this program.

WhiteFiber's delivery story is essentially an NC-1 story, and its capacity framing has moved in both directions. The July 2025 disclosures anchored on a 24 MW initial phase (4713, 4689), a 40 MW phased build (4691), and a 99 MW Duke Energy ceiling by May 2029 (4712, 4646). Within weeks the site was described as having 200 MW of potential (4547). The December 2025 Nscale agreement (3971) then reset the near-term unit of account: 40 MW of IT load in two 20 MW tranches (3972, 18328, 18329) inside a 99 MW electrical envelope (18327). By January 2026 the portfolio-level figure was 115 MW across four facilities over four years (17568), which is smaller than the 200 MW rhetoric applied to a single site. The quiet substitution of a 20 MW first tranche for the 24 MW first phase is worth noting, because the original number was never formally withdrawn.

The dates have drifted more than the company has acknowledged. NC-1 revenue start moved from May 1, 2026 (4714, 4690) to May 31 (4651, 4549) in under a week in July 2025, with no stated reason. The Q1 2026 energization target for 24 MW (4652, 4650) became April 30 ready-for-service for 20 MW (18328), and then April-May 2026 for the full 40 MW (17569). The Duke 24 MW milestone dated September 2025 (4710, 4583) appears to have been folded into a 44 MW April 2026 figure (17570) without being reconciled. The Montreal expansions show the opposite inconsistency. MTL-2 and MTL-3 targets were given as January 31, 2026 (4705, 4707), then as December 31, 2025 (4647, 4648), and the 16 MW year-end 2025 guidance (4649) is flagged delayed. In November 2025 the company said NC-1 was on schedule for early 2026 (19578), and every dated NC-1 milestone on record is now marked delayed.

The record contains no independent corroboration. There are no permits, Duke interconnection filings, or site reporting to test the announced pace against, so the trajectory rests entirely on company statements. The most recent claims date to late January 2026, which leaves an eight-month gap. Paris, Blönduós, Atlanta, and NC-2/NC-3 carry no capacity, cost, or timing claims at all.

Over the next 90 days, the third-quarter report is the main test. It should show whether Nscale billing actually began on either 20 MW tranche (3975, 3976), whether MTL-2 and MTL-3 are generating revenue, and whether the 115 MW target is restated. Any Duke filing confirming energized load at Madison would be the first external check on this timeline.

WhiteFiber's capacity story has grown much faster than its delivered megawatts. NC-1 entered the record as a 24 MW initial build (4713) inside a 40 MW phased plan (4691) and a 99 MW Duke Energy envelope stretching to 2029 (4712). It was then floated as a 200 MW potential site (4547), before settling back to a 99 MW total target by 2030 (17571) and a portfolio-wide 115 MW across four facilities (17568). The Montreal assets have stayed flat at 4, 5 and 7 MW (4702, 4704, 4706). Growth in the headline numbers has come almost entirely from North Carolina entitlements, not from new Quebec construction.

The dates have moved, and mostly without being acknowledged. NC-1's 24 MW Q1 2026 energization (4652, 4713) was still described as on schedule in November (19578). In December it was quietly restated as two 20 MW ready-for-service tranches on April 30 and May 30 (18328, 18329), and in January as Phase 1 across April and May (17569). That is a one-to-two-month slip, plus a change in unit from a 24 MW build to 40 MW of IT load billed in halves (3972). None of the restatements framed it as a delay. The first Duke tranche, due September 2025 (4710), was never confirmed. Guidance for MTL-2 and MTL-3 swung between end-2025 (4647, 4648) and January 2026 (4705, 4707) in filings weeks apart. The end-2025 total capacity target of 16 MW (4649) has no completion claim behind it. Several acquisition items flagged as delayed (4680, 4685) are backward-looking closings, not real slippage, and Enovum's completion is confirmed (4333).

The record contains no independent corroboration: no permits, no NCUC or Duke large-load filings, no Hydro-Québec allocations, and no site reporting. The Duke figures (17570, 17577) and the Nscale agreement (3971) are company-sourced. The project address in Madison, Rockingham County, is also inconsistent with the Greensboro description in the claims (4581, 4643). That should be checked against county records.

The supplied claims end in late January 2026, so as of October nothing here confirms that either 20 MW tranche reached billing. Over the next 90 days, the third-quarter report should state plainly whether Nscale billing started for both tranches (3975, 3976) and whether MTL-2 and MTL-3 are generating revenue. Also watch for permits or interconnection filings for NC-2 and NC-3 in Yadkin County, and for any further restatement of the 99 MW and 115 MW targets.

WhiteFiber's record is dominated by NC-1, and its capacity story has grown in framing rather than in delivered megawatts. The July 2025 disclosures set a 24 MW first tranche, a 40 MW cumulative step, and a 99 MW Duke Energy ceiling by May 2029 (4713, 4653, 4712). An August filing floated a 200 MW potential maximum (4547), but that figure does not recur. By January 2026 the ceiling was back to 99 MW, now dated to 2030 (17571), and a portfolio-wide target of 115 MW across four facilities within four years was added (17568). The Nscale master services agreement in December 2025 (3971) reshaped the first tranche from 24 MW into two 20 MW phases totaling 40 MW of IT load (3972, 18328, 18329). That is a larger initial commitment, and it ties delivery to a single counterparty's billing schedule.

The dates have drifted quietly. NC-1 revenue was first pegged to May 1, 2026 (4714) and moved to May 31 (4651) without comment. In November 2025 the company described the 24 MW phase as on schedule for early 2026 (19578). One month later the Q1 2026 energization (4652) had become April 30 and May 30 ready-for-service dates (18328, 18329). In Montreal, MTL-2 and MTL-3 were promised by January 31, 2026 (4705, 4707), then pulled forward to year-end 2025 within three weeks (4647, 4648). The 16 MW year-end 2025 total (4649) was not met. The Duke 24 MW milestone for September 2025 (4710, 4583) simply stops appearing after July 2025. It was superseded by a 44 MW April 2026 figure (17570) rather than acknowledged as missed. Note also that several historical acquisitions carry delayed flags (4638, 4640, 4643). This looks like a date-classification artifact rather than evidence of slippage, and those flags should be read cautiously.

The record contains no independent corroboration: no permits, no Duke interconnection filings, and no site reporting for 805 Island Drive. Filings also alternate between Madison and Greensboro as the NC-1 location (4709, 4643), which should be reconciled. The visible record ends in late January 2026 (17584). Every NC-1 phase date has since passed without a confirming claim, and Paris, Blönduós, Atlanta, and NC-2/NC-3 carry no dated claims at all.

Over the next 90 days, Q3 2026 results should be checked for explicit confirmation that Nscale billing began on both 20 MW phases. Look for Duke energization evidence at the 40 to 44 MW level, any commissioning disclosure for MTL-2 and MTL-3, and whether the 115 MW and 99 MW targets are restated, re-dated, or dropped. Yadkin County permitting records for NC-2 and NC-3 would be the first independent signal that the pipeline extends beyond NC-1.

WhiteFiber's capacity story has expanded in headline terms while getting less specific near term. In August 2025, NC-1 was framed with a 200 MW potential ceiling (4547). By January 2026 the company was instead describing a 99 MW total target for the site (17571) and 115 MW across four facilities within four years (17568). The 200 MW figure has not been repeated, and nothing in the record says why. Meanwhile, the first NC-1 phase grew from 24 MW (4689, 4713) to 40 MW split into two 20 MW tranches under the Nscale master services agreement (3971, 3972). That is a larger first phase, but a later one.

The dates show a pattern of quiet re-sequencing. In July 2025 the 24 MW initial phase was set for completion by end of Q1 2026 (4713, 4650, 4652), and as late as November it was described as on schedule for early 2026 (19578). The December Nscale terms then set ready-for-service and billing for April 30 and May 30, 2026 (18328, 18329, 3975, 3976). The January filings restated this as Phase 1 delivering 40 MW across April and May (17569), backed by 44 MW of Duke power by April (17570). The Q1 24 MW milestone was never formally retired; it was absorbed. NC-1's revenue start likewise moved from May 1 (4714, 4690) to May 31 (4651, 4549) within three weeks of July 2025 disclosures. In Montreal, MTL-2 and MTL-3 slid from a January 2026 revenue target (4705, 4707) to a December 2025 completion framing (4647, 4648, 4577, 4579), which supported a 16 MW year-end total (4649). The record contains no later claim confirming that either site was finished.

The record also contains no independent corroboration of the announced pace. There are no building permits for the Madison site, no Duke interconnection filings, and no on-the-ground reporting. The Duke milestones (4710, 4711, 4712) are company-relayed, and the 24 MW September 2025 tranche was never separately confirmed as energized. The claim history ends in late January 2026 (17585), so the April and May delivery dates have passed with nothing on file either confirming or revising them.

The Q3 2026 results, likely due in November, are the key test over the next 90 days. Watch for whether Nscale billing actually began on both 20 MW tranches, and whether MTL-2, MTL-3 and the Cerebras 5 MW deployment (4708) are reported as revenue-generating. Watch also for any timeline beyond 40 MW at NC-1. NC-2 and NC-3 in Yadkin County and the Atlanta, Paris and Blonduos sites have no stated capacity or dates at all. County permit dockets there would be the first real evidence of whether the 115 MW plan is moving.

WhiteFiber's capacity story is concentrated almost entirely in NC-1, and it has expanded upward while near-term delivery has been renegotiated. In July 2025 the site was framed as a 24 MW initial build (4689) growing to 40 MW (4691), with a 99 MW Duke Energy commitment through 2029 (4712). By August 2025 a 200 MW potential ceiling had appeared (4547). In January 2026 the portfolio was restated as 115 MW across four facilities within four years (17568). The large headline numbers have grown; the first tranche has shrunk and moved. The 24 MW Q1 2026 energization (4652, 4713) was replaced in December 2025 by the Nscale agreement's two 20 MW phases, with ready-for-service dates of April 30 and May 30, 2026 (18328, 18329). That restructuring was never presented as a slip, even though it followed a November assurance that the 24 MW phase was on schedule for early 2026 (19578). Revenue start drifted from May 1 (4714) to May 31 (4651) within three weeks of the same July disclosure cycle, also without comment. The 99 MW horizon moved from May 2029 (4712) to roughly January 2030 (17571). Montreal is harder to read. MTL-2 and MTL-3 were promised for end of 2025 (4647, 4648), which together with MTL-1 supports the 16 MW year-end guidance (4649). None of those deliveries is confirmed anywhere in the record. Several acquisition claims flagged as delayed, such as Enovum (4638) and MTL-2 (4640), are past-dated closings rather than failures, and Enovum's completion is stated in 4333. No independent evidence is on file: no permits, no Duke interconnection filings, and no site reporting. The first Duke milestone of 24 MW by September 2025 (4710) is still unverified, and the later claim of 44 MW by April 2026 (17570) is also unconfirmed. The record also ends at a truncated January 26, 2026 claim on NC-1 project-level debt (17584). Both Nscale ready-for-service dates have now passed with no delivery claim captured. Over the next 90 days, the third-quarter report should state whether Nscale billing began on one or both 20 MW phases, and whether NC-1 debt financing closed. Watch for Duke energization or interconnection documentation that would independently confirm power at the Madison site. MTL-2 and MTL-3 need an explicit in-service statement or an acknowledged delay. Paris, Blonduos, Atlanta and NC-2/NC-3 carry no capacity figures at all; any number attached to them would show whether the 115 MW plan has real sites behind it.

WhiteFiber's NC-1 story has been reframed more than it has grown. In mid-2025 the company described a 24 MW initial phase (4713, 4689) rising to 40 MW (4691, 4653) under a 99 MW Duke Energy commitment (4712). An August 2025 filing floated a 200 MW theoretical ceiling (4547) that never reappears. By December 2025 the Nscale master services agreement (3971) defined NC-1 as 40 MW of IT load in two 20 MW tranches (3972) against 99 MW of electrical capacity (18327). The underlying 40 MW first build was stable. The framing shifted from power milestones to billing milestones. January 2026 added a 44 MW Duke figure (17570) and a portfolio target of 115 MW across four facilities within four years (17568). That portfolio figure implies little beyond NC-1 and Montreal, even though Paris, Blonduos, Atlanta and NC-2/NC-3 carry no capacity claims at all.

The dates have moved quietly. The first Duke tranche of 24 MW by September 2025 (4710) passed without a restated date. NC-1 revenue start moved from 2026-05-01 (4714) to 2026-05-31 (4651) within eighteen days and without comment. Full 99 MW moved from 2029-05-16 (4646) to roughly January 2030 (17571). In November 2025 the company called the first phase on schedule for early 2026 (19578). One month later it was re-expressed as ready-for-service in April and May 2026 (18328, 18329). In Montreal, MTL-2 and MTL-3 completion was stated as January 2026 (4705, 4707) and as December 2025 (4647, 4648) in filings eighteen days apart. The 16 MW year-end 2025 guidance (4649) is flagged as missed. Some delayed flags, such as 4680 and 4682, mark historical closing dates rather than slips. The MTL-1 4 MW figure has been consistent throughout (4702, 4537).

The record contains no independent corroboration: no Guilford or Rockingham County permits, no Duke interconnection documentation beyond company statements (17577), and no site reporting. Every NC-1 deadline through May 2026 has now passed. None is matched by a delivery or billing confirmation, and the record breaks off mid-entry at 17584, a project-level debt item. Over the next 90 days, watch the Q3 2026 report for whether Nscale billing has actually begun on either 20 MW tranche. Also watch for whether Duke's 44 MW is energized, what terms the NC-1 debt financing carries, and any Yadkin County filings that would give NC-2/NC-3 substance beyond a name.

WhiteFiber's North Carolina story has been reshaped more than it has grown. In July 2025 the company framed NC-1 as a 24 MW first phase stepping to 40 MW, backed by Duke Energy commitments ending at 99 MW (4713, 4711, 4712), and by August it was citing a 200 MW potential ceiling (4547). The December 2025 Nscale agreement anchored something narrower: 40 MW of IT load delivered as two 20 MW tranches (3972, 18328, 18329), with the 99 MW number recast as electrical capacity (18327). The 200 MW figure does not reappear. The four-facility, 115 MW target (17568) moves the portfolio ambition out to a four-year horizon. In Montreal, MTL-1, MTL-2 and MTL-3 have stayed fixed at 4, 5 and 7 MW, and the 5 MW Cerebras deal (4708) accounts for most of MTL-3. The later sites (NC-2/NC-3, Atlanta, Paris, Blonduos) have no stated capacity at all.

The dates have moved in small steps that were never acknowledged. The first Duke milestone of 24 MW by September 2025 (4710) is the one hard deadline already past in the record, and no later claim confirms it was met. NC-1's first phase went from 24 MW by end of Q1 2026 (4713, 4652), to 'early 2026' in November (19578), to 20 MW ready-for-service by April 30 (18328). The tranche got smaller and the date moved later, and no disclosure tied the two changes together. Revenue start moved from May 1 to May 31, 2026 within three weeks of July filings (4714, 4651). The 99 MW endpoint moved from May 2029 (4712) to January 2030 (17571), though the later date may simply be a 'within four years' phrasing anchored to the filing date. The Montreal completion dates are internally inconsistent, cited as both January 31, 2026 and December 31, 2025 in July filings (4705, 4647).

Nothing in this record independently supports the announced pace. There are no permits, no interconnection queue entries and no site reporting, and the Duke figures come from the company's own disclosures. The record ends in late January 2026, with a truncated financing claim (17584), so it cannot show whether the April and May 2026 milestones were met.

Over the next 90 days, check Q3 2026 disclosures for explicit confirmation that both NC-1 tranches reached ready-for-service and that Nscale billing began (3975, 3976), rather than restated targets. Also look for evidence that Duke's 44 MW (17570) is energized, and for the operating status of MTL-2 and MTL-3. The first stated capacity for any of the later-stage sites would be a meaningful signal, as would Duke or state utility filings that corroborate the load ramp.

WhiteFiber's announced capacity has grown on paper while the dates underneath it have moved. At NC-1, the company first framed a 24 MW initial tranche (4689, 4713) inside a 40 MW phased build (4691). The December 2025 Nscale master services agreement recast this as two 20 MW phases totalling 40 MW (3972, 18328, 18329), so the first deliverable block shrank from 24 to 20 MW even as the headline stayed at 40. The ceiling has moved the other way. It went from a 99 MW Duke commitment (4712) to a 200 MW potential maximum (4547), and then to a 115 MW four-facility target (17568). That 115 MW figure appears to be NC-1's 99 MW plus the 16 MW Montreal trio (4649). NC-2/NC-3, Atlanta, Paris and Blonduos carry no capacity claims at all.

Dates have not held, and the revisions are mostly unacknowledged. The 99 MW horizon moved from 2029-05-16 (4712, 4646) to 2030-01-25 (17571) without comment. The Duke 24 MW milestone due 2025-09-01 (4710) is not mentioned in later filings; it was replaced by 44 MW by April 2026 (17570). NC-1 revenue drifted from 2026-05-01 (4714) to 2026-05-31 (4651), then to staggered April and May billing (3975, 3976, 17569). Montreal is internally inconsistent. MTL-2 and MTL-3 carried a January 2026 completion in one July disclosure (4705, 4707) and a December 2025 date weeks later (4647, 4648). The 16 MW end-2025 guidance (4649) is flagged delayed. The NC-1 project debt close targeted for March 2026 (17584) is also past due. As of today, every near-term dated claim in the record has lapsed without a logged confirmation.

The record contains no independent corroboration. There are no Rockingham or Yadkin County permits, no Duke interconnection filings, no Quebec municipal approvals, and no on-the-ground reporting. The record as supplied also ends in late January 2026, which leaves roughly eight months without a traceable company update. Given that, the announced pace should be treated as unverified. That is not the same as evidence of failure. The Enovum close (4333) is the only milestone with clear completion language.

Over the next 90 days, the key test is the Q3 2026 report. Watch for three things. The first is explicit confirmation that one or both NC-1 20 MW phases reached ready-for-service and that Nscale billing began. The second is whether the NC-1 debt facility closed, and on what terms. The third is energized MW in Montreal against the 16 MW guidance. Beyond the filings, it is worth checking for Duke interconnection or substation filings that support 44 MW, and for any first capacity or timeline figures for NC-2 and NC-3. If the 115 MW target is restated without a new dated interim milestone, treat that as a further soft slip.

WhiteFiber's headline capacity has grown mainly by relabeling what counts as capacity, not by adding delivered megawatts. In July 2025 NC-1 was an initial 24 MW site (4689, 4713) with a 40 MW phased build (4691). The same month it was framed as a 99 MW Duke Energy commitment (4712) and then as a 200 MW potential maximum (4547). By December the company described 40 MW of IT load in two 20 MW phases against 99 MW of electrical capacity (3972, 18327). In January 2026 it added a portfolio target of 115 MW across four facilities within four years (17568). The Montreal footprint has barely changed. MTL-1 has been stated at 4 MW throughout (4702, 4574, 4537), and MTL-2 and MTL-3 at 5 MW and 7 MW, which together make up the 16 MW end-2025 guidance (4649). The operating base remains small relative to the 99 MW, 115 MW and 200 MW figures.

The dates have mostly drifted without acknowledgement. NC-1 revenue start moved from May 1 to May 31, 2026 within eighteen days (4714, 4651). The first-phase target changed from 24 MW energized in Q1 2026 (4652) to 20 MW ready for service by April 30 (18328), which is both later and smaller, and no reconciliation appears in the record. Duke power commitments moved from 24 MW by September 2025 (4710) to 44 MW by April 2026 (17570). MTL-2 and MTL-3 completion was pulled forward from January 2026 (4705, 4707) to year-end 2025 (4647, 4648). After August 2025 neither site has another dated claim, which is itself a signal. Every dated delivery milestone in the record has now passed, and the record contains no confirmation that any was met.

None of this can be checked independently from what is on file. There are no permits, no Duke interconnection filings or queue entries, and no on-the-ground reporting. The Duke figures, the Nscale agreement (3971) and the Cerebras colocation (4708) are all sourced from the company. The record also ends in late January 2026, eight months before this analysis.

Over the next 90 days, the Q3 results should show whether Nscale billing on both 20 MW phases (3975, 3976) has actually begun. The debt facility flagged in 4717584 is the wrong id here; the relevant claim is 17584, the NC-1 project-level debt facility, and whether it closes is the key financing test. MTL-2 and MTL-3 need explicit energization disclosures. Finally, NC-2, NC-3, Atlanta, Paris and Blonduos still need their first capacity and date claims before the 115 MW target can be judged at all.

WhiteFiber's capacity story has grown in headline terms while the near-term deliverable has quietly narrowed. At NC-1 the July 2025 framing was an initial 24 MW (4689, 4713) building to 40 MW (4691, 4653), with a 200 MW ceiling floated in August (4547). By December the same site was described as 40 MW of IT load in two 20 MW tranches (3972, 18328, 18329) against 99 MW of electrical capacity (18327). That makes the first deliverable 20 MW rather than 24, and it mixes IT and electrical figures, which flatters the larger number. January 2026 widened the frame again to 115 MW across four facilities within four years (17568). Yet NC-2/NC-3, Atlanta, Paris and Blonduos appear on the project list with no capacity, cost or date claims attached, so the 115 MW has no project-level backing in the record.

The dates show drift without acknowledgement. NC-1 revenue start moved from 2026-05-01 (4714) to 2026-05-31 (4651) within eighteen days in July 2025, with no stated reason. The March 2026 energization target (4652, 4582, 4548) was later restated as ready-for-service in April and May (18328, 18329, 17569) rather than acknowledged as a slip. Montreal ran the other way: MTL-2 and MTL-3 moved from 2026-01-31 (4705, 4707) to 2025-12-31 (4647, 4648), alongside a 16 MW end-2025 total (4649). Then Montreal simply drops out of the record after August 2025. Every one of those dates has now passed, and there is no delivery confirmation on file for any of them, nor for the 2025-09-01 Duke tranche (4710).

No permits, interconnection-queue filings or on-the-ground reporting appear in this record. Every data point is company-sourced. The Nscale master services agreement (3971) is the most concrete external signal, because it ties billing dates (3975, 3976) to a named counterparty. The Duke Energy commitments (4712, 17570) are relayed by WhiteFiber, not by the utility.

Over the next 90 days, the Q3 2026 report is the key test. It should state whether both NC-1 20 MW phases are billing to Nscale and whether the project-level debt facility (17584) closed. It should also give any accounting of MTL-2 and MTL-3, which have been silent for over a year. Beyond the report, look for county permits or Duke queue entries for the Yadkin sites. Those would give the 115 MW target something other than the company's word to rest on.

WhiteFiber's capacity narrative for NC-1 has expanded, then narrowed, and the units have changed along the way. In July 2025 the company framed the site as a 24 MW initial phase building to 40 MW, backed by Duke Energy commitments of 24, 40 and 99 MW (4710, 4711, 4712, 4691). By August it cited a 200 MW potential ceiling (4547). That figure does not appear again. The December Nscale contract instead describes 40 MW of IT load in two 20 MW tranches against 99 MW of electrical capacity (3972, 18327, 3971). Moving from utility megawatts to IT megawatts is a legitimate refinement, but the company never reconciled the two, and the Duke figure itself drifts from 40 to 44 MW (4711, 17570). The January 2026 target of 115 MW across four facilities within four years (17568) is smaller than the single-site ceiling floated five months earlier.

The dates have moved quietly. Revenue start for NC-1 went from May 1 to May 31, 2026 within three weeks (4714, 4651), with no stated reason. The Q1 2026 energization of 24 MW (4652) became two ready-for-service dates of April 30 and May 30 (18328, 18329), then a looser "April and May" (17569). The Duke 24 MW milestone for September 2025 (4710) was never marked as met or revised. In Montreal, MTL-2 and MTL-3 were due by year-end 2025 (4647, 4648), feeding a 16 MW total-capacity target (4649). After August 2025 neither site gets a timeline update in the record.

The record contains no independent corroboration. There are no Rockingham County building permits, no Duke interconnection or NCUC filings, and no site reporting. The Duke commitments and the Nscale contract are company-described, and the contract is evidence of demand, not of construction. The record also goes silent after late January 2026, so every dated NC-1 milestone has now passed with no delivery claim logged. That silence is the central fact. It may reflect a gap in ingestion rather than a miss, but it cannot be read as on-time delivery.

Over the next 90 days, the Q3 2026 results should confirm or deny that both 20 MW tranches are billing to Nscale, and should state what MW is actually energized. Watch for disclosure of the project-level debt facility (17584), since lenders' engineers impose their own milestone discipline. The results should also say whether MTL-2, MTL-3 and the Cerebras 5 MW deployment (4708) are live. Finally, watch for any first substantive claims on NC-2/NC-3, Atlanta, Paris or Blönduós, which so far exist only as names.

WhiteFiber's capacity story for NC-1 has been redrawn several times while its headline delivery dates stayed nearly fixed. In July 2025 the site was a 24 MW initial build (4689, 4713) with a 40 MW phased path (4691, 4653) and a 99 MW Duke Energy ceiling by 2029 (4688, 4646). By August the ceiling had become a 200 MW potential maximum (4547). In December it settled at 99 MW of electrical capacity (18327) supporting 40 MW of IT load in two 20 MW tranches (3972, 18328, 18329). The first tranche is therefore smaller than the original 24 MW figure, even as the January 2026 framing raised the portfolio to 115 MW across four facilities within four years (17568), from a 16 MW end-2025 target (4649).

The dates have moved quietly rather than dramatically. NC-1 revenue went from May 1 (4714, 4690) to May 31 (4651, 4549), then became staged billing on April 30 and May 30 (3975, 3976). The January restatement set Phase 1 for April and May 2026 (17569). The original Duke 24 MW milestone of September 2025 (4710, 4686) was never re-dated. It was simply superseded by a 44 MW-by-April-2026 figure (17570). The Montreal expansions are less clean. MTL-2 and MTL-3 carried a January 31, 2026 completion (4705, 4707), but later July and August filings cite December 31, 2025 (4647, 4648, 4541, 4544). No refreshed date appears after August. The NC-1 project debt close targeted for Q1 2026 (17584) is also flagged delayed. Several acquisition entries tagged delayed (4638, 4640, 4682) are backward-looking completion dates rather than missed promises.

The independent evidence in this record is thin. The Duke commitments and the Nscale master services agreement (3971) are company-reported, and no permits, interconnection-queue filings or site reporting are on file for NC-1 or Quebec. The site is also described as both Greensboro (4581) and Madison. NC-2/NC-3, Atlanta, Paris and Blonduos carry no capacity or cost at all.

Over the next 90 days, the key question is whether the Q3 report confirms that both NC-1 tranches are actually billing to Nscale, and at what megawatt count. The other items to watch are whether the project debt facility has closed, and whether MTL-2 and MTL-3 receive an explicit in-service disclosure. Any Duke or county filing that independently dates energization would be the first outside check on this timeline.

The capacity story has moved in two directions at once. The NC-1 envelope has only ever widened in the telling — 24 MW initial (4713, 4689), a phased 40 MW (4691, 4653), a Duke commitment ladder to 99 MW (4712, 4688, 4646, 2164), and at one point a 200 MW theoretical ceiling (4547) — while the portfolio has sprawled into Paris, Blonduos, Yadkin County, Atlanta and additional Montreal sites, most of them announced with no megawatt or cost figure attached at all. Against that, the January 2026 framing of four facilities totalling 115 MW inside four years (17568) is materially more modest than the sum of the earlier individual asks, and NC-1's 99 MW quietly acquired a 2030 horizon (17571) where the Duke agreement had said May 2029. That re-dating was not presented as a revision.

The date record is where the strain shows. Duke Phase 1 at 24 MW was due 2025-09-01 (4710, 4686, 4644, 4583) and has not been confirmed. MTL-2 and MTL-3 were first put at 2026-01-31 (4705, 4707), then restated to 2025-12-31 (4647, 4648, 4577, 4579, 4541, 4544) — a tightening, not a slip, and now roughly nine months past due. NC-1 revenue slid 2026-05-01 to 2026-05-31 (4714, 4690, 4651, 4549), was recast in December as billing on two 20 MW phases in April and May (3975, 3976), then appeared as 44 MW by 2026-04-30 (17570). Today those dates are four to five months behind. Every acquisition on the list — Enovum (4638, 4536), MTL-2 (4640), MTL-3 (4542), the NC-1 land (4685, 4643) — closed late relative to its own stated date.

Independent corroboration is thin in this record, and the site labelling itself is loose: the property is described as Greensboro (4709, 4581) but sits at 805 Island Drive in Madison, a different county for permitting purposes. Nothing here is sourced to a Duke large-load queue position, a Rockingham County permit, or Quebec interconnection filings.

Over the next ninety days, watch for a disclosed energized-versus-contracted MW split, confirmation that Nscale (3971) is actually being billed, evidence the project debt facility due 2026-03-31 (17584) closed, in-service proof at MTL-3 against the Cerebras 5 MW anchor (4543), and whether Atlanta (2175) or Yadkin County ever acquire numbers. The record supplied here truncates at claim 2180, so later disclosures may qualify this.

WhiteFiber's record points to a portfolio whose headline capacity has grown mainly through new sites, partner capacity and optionality, while the core self-build milestones have repeatedly carried delayed flags. At NC-1 the utility ladder of 24, 40 and 99 MW was restated almost word for word across the July and August 2025 filings (4710-4712, 4644-4646, 4583-4585). The upside was then widened to a 200 MW potential ceiling (4547). The December 2025 Nscale agreement narrowed the near-term picture to 40 MW of IT load in two 20 MW phases (3972). That is a reasonable way to translate grid capacity into billable load, but it quietly replaced the earlier 24 MW-by-Q1-2026 framing (4713, 4650, 4582). Revenue start also drifted without comment. It was 1 May 2026 in 4714 and 4690, then 31 May in 4651 and 4549, before splitting into billing dates of 30 April and 30 May (3975, 3976). The 24 MW Duke tranche dated 1 September 2025 was still being repeated weeks before that deadline (4583). By the March 2026 filing, only the 2029 99 MW commitment was referenced (2164, 2211). There was no statement on whether the 24 and 40 MW steps arrived. The $8 million price increment tied to power capacity by May 2028 (2163) suggests that the full power envelope was not yet locked.

In Quebec the capacity figures held steady at 4, 5 and 7 MW. The dates did not. MTL-2 and MTL-3 completion was given as 31 January 2026 in one July 2025 document (4705, 4707) and as year-end 2025 in others that same month (4647, 4648). The 16 MW year-end guidance depended on both sites and is flagged delayed (4649). MTL-3 also moved from being described as a secured site or acquisition (4641, 4578) to a lease with a purchase option due to close by 30 June 2026 (2161). The Atlanta expansion date of 1 February 2026 had already passed when it was disclosed (2175). Blönduós added 5.5 MW through a partnership (2199). NC-2, NC-3 and Paris carry no dated claims in the data shown.

This record contains no independent evidence, such as permits, interconnection queue positions or site reporting, so the announced pace cannot be verified from it. Note also that the input was cut off after claim 2177.

Over the next 90 days, watch for Q3 results that confirm or deny Nscale billing on both NC-1 phases. Look for Duke or NCUC filings that show energized load, and for Rockingham County permits that reconcile the Madison address with the Greensboro references. Other open questions are whether the MTL-3 option actually closed, whether MTL-2 is generating revenue, and whether Atlanta, NC-2 or NC-3 receive dates that can be tested.

WhiteFiber's capacity story splits into a small, stable Quebec base and a North Carolina campus whose framing has widened faster than its delivery record. The Montreal figures have not moved: MTL-1 at 4 MW (4702, 4574, 4537), MTL-2 at 5 MW (4704, 2177) and MTL-3 at 7 MW (4706, 2180) add up to the 16 MW year-end 2025 guidance (4649). The one change in Quebec is structural. MTL-3 was described as an acquisition (4578) and later as a lease with a purchase option due by 2026-06-30 (2161). NC-1 runs the other way. It went from a 24 MW initial phase (4689) to 40 MW phased (4691), then to a 99 MW Duke Energy commitment (4712, 2164) and a 200 MW potential ceiling (4547). The December 2025 Nscale contract recast the near-term deliverable as 40 MW of IT load in two 20 MW tranches (3972, 3975, 3976).

Dates have drifted in small steps without being flagged. NC-1 revenue start moved from 2026-05-01 (4714, 4690) to 2026-05-31 (4651, 4549) within three weeks of July 2025 disclosures. Nscale billing was split across 2026-04-30 and 2026-05-30 (3975, 3976), then consolidated into a single 40 MW start on 2026-06-01 (2196). That pushes the first tranche back a month, and the visible record never calls it a revision. The Montreal completion language wavered between 2026-01-31 (4705, 4707) and 2025-12-31 (4647, 4648) in filings weeks apart. The Duke 24 MW milestone of 2025-09-01 (4710, 4583) does not reappear in the March 2026 set, and the 40 MW 2026-04-01 target was restated only days before it came due (2189). Atlanta first appears with a date already past (2175). Every dated milestone in the record is now marked delayed. None has a confirmation of completion attached, and the latest claims date to March 2026 despite six further months elapsing.

The record contains no independent corroboration. There are no Rockingham County or Madison permits, no Duke interconnection or large-load filings, no Hydro-Québec service confirmations and no site reporting. The pace is supported by company disclosure alone. The only hard transactional anchors are the closed Enovum deal (4333) and purchase prices (2158, 2162).

Over the next 90 days, the Q3 2026 report should show whether Nscale billing actually began on both 20 MW tranches and whether the MTL-3 option was exercised or quietly extended. It should also show whether Duke energized 40 MW at NC-1 and whether Atlanta receives a real date. Any restatement of the 99 MW or 200 MW figures, or silence on them, would be the clearest signal of how firm the long-range numbers are.

WhiteFiber's capacity story has grown in North Carolina and stayed flat in Quebec. NC-1 was first described as a 24 MW initial build (4689) with a phased path to 40 MW (4691, 4653). It was then tied to a 99 MW Duke Energy commitment by May 2029 (4688, 4646), and by August 2025 to a 200 MW potential maximum (4547). The December 2025 Nscale agreement (3971) gave the site its most concrete form: 40 MW of IT load in two 20 MW phases (3972) at 150 kW density (3973). The 99 MW and 200 MW figures are option value, not contracted demand. Montreal has not moved. MTL-1 at 4 MW (4537), MTL-2 at 5 MW (2177) and MTL-3 at 7 MW (2180) add up to the 16 MW year-end 2025 guidance (4649), with no additions since. The one Quebec change is structural. MTL-3 was first called a secured site acquisition (4682, 4641) and later a lease with a purchase option to close by June 30, 2026 (2180, 2161).

The timeline record is weaker. Filings in the same July 2025 window put MTL-2 and MTL-3 completion at both December 31, 2025 (4647, 4648) and January 31, 2026 (4705, 4707), and nothing reconciles the two. The Duke 24 MW milestone of September 2025 (4583) passed with no confirmation on record. The 40 MW April 1, 2026 target was repeated unchanged in late March (2189), days before it fell due. NC-1 revenue start moved from May 1 (4714) to May 31, 2026 (4651). Nscale billing was then set for April 30 and May 30 in two tranches (3975, 3976), and by March 2026 it had become a single 40 MW commencement on June 1 (2196). That merge pushed the first phase back a month without saying so. The Atlanta expansion date of February 1, 2026 (2175) had already passed when it was reported.

No independent corroboration appears in this record. There are no Rockingham County building or occupancy permits, no Duke interconnection or NCUC filings, and no Hydro-Québec allocations or local reporting. The Duke figures are the company's own description of its utility commitments. Paris, Blonduos and NC-2/NC-3 have no capacity or date claims at all.

Over the next 90 days, watch Q3 results for three things. The first is confirmation that Nscale billing actually began, and at what megawatts. The second is whether the MTL-3 purchase option closed after its June deadline. The third is whether the MTL-2 and MTL-3 megawatts are now producing revenue. Also watch whether the 99 MW and 200 MW figures are repeated or quietly dropped, and whether Duke energization evidence surfaces to support the conditional power-linked purchase price increment (2163).

WhiteFiber's stated capacity has mostly held its size, but the timelines attached to it have been quietly renegotiated. At NC-1 the near-term figure has barely moved. The 24 MW initial phase and 40 MW cumulative build (4713, 4653) became a 40 MW Nscale contract billed in two 20 MW tranches (3972, 3975, 3976). Meanwhile the long-range framing has grown, from a 99 MW Duke Energy commitment (4712) to a 200 MW potential ceiling (4547). The Montreal cluster has stayed fixed at 4, 5 and 7 MW (4702, 4704, 4706). That cluster was the basis for the 16 MW year-end 2025 guidance (4649), which is flagged as delayed. Iceland has since been added at 5.5 MW (2220).

The dates are the weaker part of the record. NC-1 revenue start moved from May 1 to May 31, 2026 within three weeks of July 2025 (4714, 4651), with no stated reason. The Nscale billing schedule first called for April 30 and May 30 (3975, 3976). By March 2026 it had become a single June 1 date for the full 40 MW (2196), which collapses the first tranche into the second without acknowledging a slip. The Duke 24 MW milestone of September 2025 was still being repeated in March 2026 (2188), long after it had passed. MTL-2 and MTL-3 completion dates moved in the wrong direction for credibility: they read January 2026 (4705, 4707), then year-end 2025 (4647, 4648). By March 2026, MTL-3 was being described through a purchase-option exercise dated June 30, 2026 (2207), and MTL-2 no longer carried a date at all (2206). The Cerebras 5 MW agreement at MTL-3 (4708, 4543) is not visibly reaffirmed in the March 2026 filing. The same filing gives the NC-1 contingent price step two different dates, 2027 and 2028 (2210, 2163).

No independent corroboration appears in this record: no permits, no Duke interconnection or energization filings, and no site reporting from Madison or Saint-Jérôme. The Duke figures are company-reported restatements of a utility agreement, not utility disclosures. Paris, Atlanta and NC-2/NC-3 carry no claims of any kind. The record as supplied also ends mid-entry, so later claims may exist.

Over the next 90 days, watch whether the Q3 report confirms Nscale billing on the full 40 MW or restates it. Watch for any Duke confirmation that 40 MW is actually energized, and whether MTL-3's purchase closed and Cerebras is live. Finally, watch whether the 200 MW ceiling is tied to a dated power commitment or remains aspirational.

WhiteFiber's capacity guidance has held flat rather than scaled upward. The company maintains 24 MW for NC-1's initial phase (claims 4713, 4652), 40 MW potential by mid-2026 (claim 4653), and a Duke Energy phased partnership extending to 99 MW by 2029 (claim 4712). Montreal-region facilities remain at 4 MW for MTL-1 (claims 4702, 4679), 5 MW for MTL-2 (claims 4704, 2177), and 7 MW for MTL-3 (claims 4706, 2160), plus a 5 MW Cerebras colocation agreement (claims 4708, 4580). No announced capacity has been withdrawn. However, only MTL-1 operates at the stated 4 MW with 5,000 GPU units deployed (claim 4539), meaning the announced portfolio remains constrained to a single facility while growth projects stall in announcement phase. Timeline slippage has been systematic and unacknowledged. NC-1's initial 24 MW was promised for March 2026 (claims 4713, 4582), with revenue generation by May 2026 (claims 4714, 4690). By December 2025, WhiteFiber revised to phased billing—first 20 MW by April 2026, second 20 MW by May 2026—without acknowledging the implicit delay (claims 3975, 3976, both marked delayed). By March 2026, the company re-promised Nscale's 40 MW billing commencement for June 1, 2026 (claim 2196, marked delayed). MTL-2 and MTL-3 both committed to December 2025 completion (claims 4647, 4648); both remain marked delayed. Prerequisite acquisitions slipped across all facilities: MTL-2 from December 2024, MTL-3 from April 2025, NC-1 from May 2025 (claims 4680, 4682, 4685). Enovum's acquisition drifted from October 2024 through August 2025 without resolution (claims 4638, 4536). Rather than consolidating timelines or acknowledging cumulative delays, the company re-filed similar claims with updated dates, treating each disclosure cycle as new rather than admitting prior misses. Independent validation is entirely absent. No permits, grid interconnection queue positions, or construction permitting documents appear on record. No third-party site reporting or utility correspondence corroborates facility progress. All claims originate from company filings alone, lacking infrastructure-level evidence that promised capacity is being constructed. Over ninety days, track whether NC-1 and MTL facilities report actual operational capacity and revenue for promises made in 2026. Specifically: Nscale's 40 MW billing start (promised June 1, 2026); MTL-2 and MTL-3 completion dates and customer-facing launch; Duke Energy Phase 2 (40 MW) and Phase 3 (99 MW) advancement. Any new facility announcements warrant skepticism until the stalled projects show revenue.

WhiteFiber's announced capacity has remained largely flat while delivery timelines have systematically slipped across all major projects. Montreal facilities MTL-1, MTL-2, and MTL-3 are positioned at 4 MW, 5 MW, and 7 MW respectively, with these figures repeated identically in claims spanning July through August 2025 and March 2026, suggesting no incremental progress. NC-1 in North Carolina shows apparent growth from 24 MW initial (claim 4713) to 40 MW cumulative (claim 4691) to 200 MW maximum potential (claim 4547), but this represents theoretical scaling rather than announced capacity delivery. The company provided 16 MW total capacity guidance for end of 2025 (claim 4649), which has already slipped.

Timeline performance reveals systematic miss patterns. MTL-2 completion was promised 2025-12-31 (claims 4647, 4577, 4541)—now nine months overdue—following a property acquisition deadline of 2024-12-27 (claims 4680, 4640, 4540) already missed by nine months. MTL-3 faced identical end-2025 completion targets (claims 4648, 4544, 4579) and site-secured deadlines of 2025-04-11 (claims 4641, 4542) all marked delayed. NC-1 shows the most severe pattern: initial 24 MW energization promised 2026-03-31 (claims 4713, 4582, 4548) has passed without reported activation. Revenue commencement deadlines shifted from 2026-05-01 (claim 4714) to 2026-05-31 (claim 4651), both now missed. Duke Energy phasing shows acute delays—the Phase 1 24 MW commitment due 2025-09-01 (claims 4710, 4583) is eleven months overdue with no acknowledgment. Phases 2 (40 MW by 2026-04-01, claims 4711, 4645) and billing milestones (claims 3975, 3976, 2196) all trail by several months.

Repeated identical claim announcements across July-August 2025, December 2025, and March 2026 suggest re-tabling of unfulfilled commitments rather than announced achievements. The Enovum acquisition promised by 2024-10-01 or 2024-10-11 (claims 4638, 4536) appears still incomplete despite claim 4333's September 2025 completion statement. No permits, interconnection queue positions, or third-party verification appears in the record to validate announced pace. Watch over the next 90 days whether NC-1's reported 24 MW activates with customer billing, whether MTL-2 and MTL-3 complete, and whether Duke Energy Phase 1 capacity agreement materializes or receives formal extension without disclosure.

WhiteFiber's delivery narrative has contracted sharply while timelines have accumulated severe delays across all major projects. The company initially promised 24 MW at NC-1 (Madison, North Carolina) by Q1 2026, with revenue generation by May 2026 and Duke Energy's 24 MW Phase 1 by September 2025 (claims 4713, 4714, 4710). As of September 2026, all three milestones remain unmet. The property acquisition itself, promised for May 2025 (claim 4685), has not closed. MTL-2 and MTL-3 in the Montreal region were both promised operational status by end of 2025 (claims 4705, 4707), a deadline now nine months past. Critically, these are acquisitions or leases, not construction milestones—their persistent delay suggests foundational obstacles rather than execution friction. The repeated announcement of identical timelines across multiple claim batches (July 10, July 22, July 28, July 31, and August 7) without date adjustments suggests companies are re-releasing prior guidance rather than providing updated trajectory. Duke Energy's three-phase capacity commitment shows scaling intent (24 MW by September 2025, 40 MW by April 2026, 99 MW by May 2029, claims 4710–4712), but only the longest-dated Phase 3 remains on track. The claimed maximum capacity of 200 MW at NC-1 (claim 4547) appears aspirational. Independent verification is absent from the record—no permits, interconnection queue positions, or third-party confirmations anchor these claims. The Montreal facilities cite Cerebras colocation (claim 4708) and 5,000 GPU units at MTL-1 (claim 4539), yet no customer deployment announcements appear. Over 14 months of missed deadlines across property, infrastructure, and utility coordination suggests systemic delay rather than minor slippage. In the next 90 days, watch whether NC-1 achieves any load energization or Duke Energy formalizes a revised timeline. Property closures—Enovum (originally promised October 2024), MTL-2 (December 2024), and MTL-3 (April 2025)—would establish baseline credibility. Any timeline revision for 2026 completion targets would signal realistic re-baselining.

WhiteFiber's announced capacity—approximately 20 to 40 megawatts across MTL-1, MTL-2, MTL-3, and NC-1's initial phase, with a stated potential ceiling of 200 megawatts (4547)—has remained stable across successive announcements. The company has not reduced any stated facility capacity figure. MTL-1 consistently claims 4.0 megawatts; MTL-2 5.0 megawatts; MTL-3 7.0 megawatts; NC-1 initial phase 24 megawatts. This near-term portfolio, however, sits in sharp tension with Duke Energy partnership commitments of 163 megawatts across three phases due by May 2029 (4646, 4688), implying significant undisclosed buildout ahead.

Every facility and timeline claim with a promised date has slipped. NC-1's 24-megawatt initial phase, due March 31, 2026 (4713, 4650, 4582, 4548), is now five months overdue. NC-1 revenue generation, promised May through May 31, 2026 (4714, 4690, 4651, 4549), remains incomplete. MTL-3 was targeted for December 2025 (4648, 4579); MTL-2 for December 2025 (4647, 4577). Duke Energy Phase 1—24 megawatts due September 1, 2025 (4710, 4686, 4644, 4583)—passed twelve months ago. Phase 2, 40 megawatts, was due April 1, 2026 (4711, 4687, 4645, 4584, 2189) and is five months overdue. Nscale service commencement was rescheduled at least three times across April 30, May 30, and June 1, 2026 (3975, 3976, 2196).

Delays compound sequentially: real estate acquisitions slipped first, deferring facility buildout, which then deferred power delivery and revenue generation. Multiple identical restatements—the Duke Energy 24-megawatt phase appears across six separate claims—suggest rolling rescheduling rather than discrete revisions. The company has not publicly acknowledged root causes or provided a coherent explanation for why every single facility missed its announced date.

No independent evidence (permits, queue positions, on-the-ground construction status, third-party reporting) has been provided to validate revised timelines. In the next 90 days, watch whether NC-1 achieves 24-megawatt operational status, whether MTL-2 and MTL-3 move from acquisition phase to revenue generation, whether Nscale billing actually commences, and critically, whether revised Duke Energy milestones or capacity guidance emerges without acknowledgment of previous misses.

WhiteFiber's announced capacity has remained remarkably flat: MTL-1 holds steady at 4.0 MW (4702, 4679, 4639, 4574), MTL-2 at 5.0 MW (4704, 4680, 4576), MTL-3 at 7.0 MW (4706, 4578, 2180), and NC-1 phased in at 24 MW then 40 MW (4713, 4689, 4691, 4653). The company guided 16 MW operational by end-2025 (4649); that target was breached. No evidence of facility rightsizing or capacity retreats exists across the record—only deferrals.

Timeline discipline has collapsed across all facilities. NC-1's property acquisition slipped from May 2025 to still-pending (4685, 4643, 4545). Enovum acquisition, promised October 2024 (4638), remains open even at the September 2025 restatement (4333), a 9-month miss with no acknowledged revision. MTL-2 and MTL-3 completions shifted from January 2026 to December 2025 (4705, 4707 to 4647, 4577, 4579, 4544)—only moved forward by 4 weeks in restatement, not addressing underlying delays.

Most striking: Duke Energy Phase 1, promised September 2025 (4710, 4686, 4644, 4583), is now 12 months overdue as of today (2026-09-05). Phase 2, promised April 2026 (4711, 4645, 4584, 2189), has also breached. NC-1 revenue start drifted from May to May 31, 2026 (4714 to 4651, 4549), yet both dates have passed without reported achievement. Nscale phase billing dates (April 30, May 30, June 1: 3975, 3976, 2196) suggest revenue recognition was promised but not delivered.

No independent verification data—permits, interconnection queue positions, or regional utility filings—exists in the record to validate claimed construction pace or Duke Energy commitments. Absence of ground-truth corroboration compounds confidence concerns.

WhiteFiber's pattern shows capacity stable but timelines systematically eroding. The gap between guidance and delivery widens: a 4-month miss on NC-1 property, 9 months on Enovum, 12 months on Duke Phase 1. Over the next 90 days, watch for: (1) whether MTL-3 purchase option closes by end-June deadline or slips further (2161); (2) any new timeline guidance on NC-1 or Enovum; (3) whether Duke Energy Phase 1 finally energizes or is formally abandoned; (4) whether capacity guidance is revised downward for any facility. Any new announcement without acknowledgment of prior misses would signal communication pattern, not project health.

WhiteFiber's announced capacity trajectory has remained relatively stable at incremental phases, but the delivery timeline shows a pattern of consistent slippage across nearly all major milestones. The company committed to three phased Duke Energy partnerships for North Carolina (24 MW by September 2025, 40 MW by April 2026, and 99 MW by May 2029), with the first phase now a full year overdue and the second phase five months behind schedule as of the analysis date. Simultaneously, Montreal expansion projects (MTL-2 and MTL-3) each missed end-of-2025 completion targets by nine months, with MTL-3's purchase option exercise deadline of June 2026 also passed without public confirmation of closure. The Enovum data center acquisition—originally promised for October 2024—remains unremarked upon after nearly two years, creating ambiguity around MTL-1's role in the portfolio. NC-1 in Greensboro presents the most critical watch point, as initial 24 MW energization was promised for March 2026 with revenue start by May 2026 (claim 4714, 4713); both dates have now expired without revision statements or progress milestones. This contrasts with the stated 200 MW potential at the site, suggesting phasing rather than fundamental constraint. Independent verification is sparse in the available record—permit activity, interconnection queue positions, and on-the-ground reporting from Madison, North Carolina would clarify whether delays reflect permitting friction, supply chain lag, or demand-side customer commitments. Over the next 90 days, monitor for: (1) any Duke Energy activation announcements or revised timelines for NC-1 phases, indicating whether the September 2025 and April 2026 dates will be formally rebooked or abandoned; (2) MTL-3 purchase option closure or extension announcement; (3) Enovum integration status or standalone facility plans; and (4) any revised guidance on Atlanta expansion (claim 2175, originally due February 2026). The absence of revised statements alongside missed dates suggests either ongoing confidence in future completion or limited disclosure discipline.

WhiteFiber's stated delivery trajectory reveals widening gaps between capacity announcements and timeline adherence. Capacity projections have inflated substantially without corresponding progress on foundational deliverables. NC-1, initially announced for 24 megawatts operational by March 31, 2026 (claims 4713, 4582), is now five months past that date. The company has simultaneously expanded the project's theoretical scope to 99 megawatts under Duke Energy commitment (claim 4712) and 200 megawatts speculative maximum (claim 4547), layering larger claims without resolving smaller ones.

The pattern of delays shows systematic compression without acknowledgment. Enovum acquisition was promised October 1, 2024 (claim 4638), then October 11, 2024 (claim 4536). By September 2025 (claim 4333), the company stopped providing a completion date, simply marking it stated. This represents nine months of overrun with no public revised timeline. MTL-2 real estate acquisition promised December 27, 2024 (claim 4640) remains undated and undelivered eight months later. MTL-3 site acquisition promised April 11, 2025 (claim 4641) is five months overdue. NC-1 property purchase promised May 20, 2025 (claim 4643) is now three months past due.

Timeline compression appears deliberate. NC-1 revenue start shifted between May 1 and May 31, 2026 across claims (4714, 4690 versus 4651, 4549), then both dates passed without announcement. Duke Energy's 24-megawatt Phase 1 was promised September 1, 2025 (claim 4583). That deadline passed one year ago. The 40-megawatt Phase 2 promised April 1, 2026 (claim 4584) is five months late. Atlanta expansion promised February 1, 2026 (claim 2175) lacks any updated status.

Independent verification is absent. No grid interconnection queue entries cited, no building permits filed, no third-party construction reporting. MTL-1 remains the sole facility with apparent operational status at 4 megawatts (claim 4574), inherited from prior Enovum assets. Only this pre-acquisition footprint appears verified.

Over 90 days, watch for actual NC-1 revenue commencement now several months overdue. Monitor whether Enovum closes with published facility roster. Observe Duke Energy's response if the 99-megawatt commitment remains unmet as Phase 2 deadline passes. Track whether further capacity announcements occur before foundational projects stabilize, suggesting continued projection creep.

WhiteFiber's facility delivery shows systematic delays across all major projects without public revision of previously announced timelines. The company promised NC-1 in Madison, North Carolina would reach 24 MW operational capacity by March 31, 2026 (claims 4713, 4650, 4582), with revenue generation by May 31, 2026 (claim 4651). Both dates have passed with no facility online and no revised announcement on record. MTL-2 and MTL-3 in the Montreal region each promised completion by December 31, 2025 (claims 4705, 4681, 4647, 4577 for MTL-2; 4683, 4707, 4648, 4579 for MTL-3)—both now nine months overdue without revision.

The Duke Energy partnership timeline is the most severe miss. Phase 1, comprising 24 MW of capacity commitment, was promised by September 1, 2025 (claims 4710, 4686, 4644, 4583). That date is now 12 months past. No revised timeline has been announced, yet subsequent phases (40 MW by April 2026 and 99 MW by May 2029) remain on record. The disconnect between Phase 1's lapsed commitment and the later phases raises questions about the partnership's actual status.

Capacity announcements have also contracted without explicit revision. WhiteFiber stated 16 MW would be operational by end of 2025 (claim 4649). The only documented operational capacity is MTL-1 at 4 MW from the Enovum acquisition, which itself slipped from October 2024 to completion in September 2025. The trajectory announced for NC-1—24 MW initial, scaling to 40 MW, then 99 MW under Duke Energy phases—remains articulated, but execution delays of over a year undermine the phased capacity roadmap.

No independent corroboration is cited in the record. No permits, interconnection queue entries, or third-party site reporting validate construction pace or power infrastructure readiness. The most recent detailed claims date to August 7, 2025; the absence of updates through September 2026 is notable in itself.

Watch for any revised NC-1 completion or revenue dates, construction or equipment deployment evidence at NC-1, MTL-2, or MTL-3, public status on the Duke Energy Phase 1 commitment and whether it has been modified or terminated, and any 2026 capacity or revenue guidance from management.

WhiteFiber's North American datacenter trajectory shows an ambitious ramp that has consistently missed its delivery dates. Initial capacity guidance centered on modest facilities—4 MW at MTL-1 (claim 4702), 5 MW at MTL-2 (claim 4704), 7 MW at MTL-3 (claim 4706)—before expanding dramatically through the Duke Energy partnership, which promised 24 MW, then 40 MW, then 99 MW for NC-1 (claims 4711, 4712). Capacity has scaled upward repeatedly in claims, but execution lags significantly behind every announced date.

Promised dates have slipped across the portfolio without exception. The Duke Energy 24 MW milestone, promised for September 2025 (claim 4710), is now a full year overdue as of this writing. MTL-2's acquisition, promised December 2024 (claim 4680), remains incomplete 20+ months later. MTL-3's site acquisition, promised April 2025 (claim 4682), has not occurred. NC-1's initial 24 MW energization, promised March 2026 (claim 4652), is now five months past deadline. Revenue start dates—first promised May 2026 (claim 4714), then revised to May 31 (claim 4651)—have already elapsed. The Nscale billing phases for NC-1, promised April and May 2026 (claims 3975, 3976), show similar delays.

Revisions without acknowledgement are apparent in the Enovum acquisition timeline, which shifted from October 2024 (claim 4638) to October 2024 again (claim 4536) despite being claimed complete in September 2025 (claim 4333). Contingent purchase price timing for NC-1 also shows inconsistency: claim 2163 targets 2028-05-20, while claim 2210 states 2027-05-20. No external evidence is provided—no NERC queue entries, county permits, power interconnection agreements, or third-party construction timelines appear in available claims. Duke Energy's capacity commitments lack confirmation through independent regulatory filings.

The next 90 days will test whether NC-1 finally achieves revenue operations or slips further. Watch for any Nscale billing updates on the long-delayed 40 MW phases, confirmation of MTL-2 and MTL-3 completion status, and whether the company issues fresh capacity or timeline guidance without revisiting prior misses. The absence of independent verification in the record, combined with pervasive delay patterns, suggests claimed delivery dates should be treated with substantial skepticism.

WhiteFiber's announced capacity roadmap has expanded aggressively without acknowledging prior execution gaps. The company claimed MTL-1 at 4 MW, MTL-2 at 5 MW, and MTL-3 at 7 MW throughout 2025 (claims 4702, 4704, 4706), then progressively escalated NC-1 from an initial 24 MW (claims 4713, 4689) to 40 MW phased (claim 4691) and eventually positioned it as a potential 200 MW site (claim 4547, August 2025). This progression occurred while simultaneously missing every promised delivery milestone.

All major completion dates have slipped uniformly. MTL-2 and MTL-3 were both promised operational by end-2025 (claims 4705, 4707, 4647, 4648), now acknowledged as delayed. NC-1's initial 24 MW was targeted for March 2026 (claim 4650, 4713), with revenue starting by May 2026 (claims 4714, 4651)—both now past and marked delayed. The Duke Energy partnership tells the starkest story: Phase 1 (24 MW) was due September 2025 (claim 4644), Phase 2 (40 MW) April 2026 (claim 4645), yet these remain unfulfilled. Phase 3's 99 MW extends to May 2029, suggesting management is purchasing time with long-dated commitments (claims 4712, 4646).

Notably, July-August 2025 claims repeat identical timelines without revision or acknowledgement of prior misses (compare claims 4714 and 4690, both dated 2025-07-10 and 2025-07-22, both citing May 2026 for NC-1 revenue). The Enovum acquisition, originally due October 2024 (claim 4638), remained marked delayed through August 2025 but appears to have closed by September 2025 (claim 4333). MTL-2 acquisition was overdue by nine months (target December 2024, claim 4680). NC-1 property acquisition, promised May 2025 (claim 4685), is marked delayed. December 2025 claims (3975, 3976) refined the NC-1 phasing to 20 MW billing tranches rather than 24 MW, suggesting scope recalibration but no public timeline reset.

The trajectory shows operational execution lag widening relative to public commitments. Watch for: actual NC-1 power-up within 90 days (required if billing phases are to meet any revised 2026 schedule), Duke Energy Phase 2 announcement or delay acknowledgement, and whether MTL-2 or MTL-3 achieve operational status before year-end 2026. Permit data or interconnection queue entries would indicate whether delays stem from regulatory pace or internal capacity constraints.

WhiteFiber's portfolio expansion shows a striking pattern of announced ambition followed by consistent execution delays. The company has staked its near-term strategy on North Carolina's NC-1 facility (claims 4713, 4714, 4650, 4651), with promised initial capacity of 24 MW by March 31, 2026 and revenue generation by May 2026 (claims 4714, 4690). As of today—nearly six months past those dates—no public evidence confirms these milestones achieved. Duke Energy power commitments, announced in phases (24 MW by September 2025, 40 MW by April 2026, 99 MW by May 2029), show similar slippage (claims 4583, 4584, 4585). The first tranche missed its September 2025 deadline by months.

Montreal-area facilities tell a grimmer story. MTL-2's property acquisition, promised for December 27, 2024 (claim 4640), remains undone as of September 2026—nine months late (claim 2158 notes a $23.3M purchase). MTL-3's facility completion and revenue generation, promised by December 31, 2025 (claims 4648, 4579), similarly has not materialized. Only MTL-1, already operational, claims 4 MW deployed with 5,000 GPUs (claims 4574, 4539).

Announced total capacity has not scaled as promised. Claim 4649 targeted 16 MW operational capacity by December 31, 2025; that guidance missed. The company now appears to be tracking toward a phased NC-1 rollout (claim 3972 describes 40 MW in two phases), with billing milestones in April and June 2026 (claims 3975, 3976) already passed without confirmation. The Enovum acquisition, promised October 2024 and repeatedly re-announced, shows no public completion signal.

Without independent evidence—interconnection queue filings, building permits, utility records—only WhiteFiber's own milestones are available. The pattern of delays without revised guidance suggests either internal projections that don't reflect execution, or external constraints (permitting, supply chain, utility coordination) that aren't being communicated.

In the next 90 days: watch for any North Carolina utility commission filings, Greensboro permitting records, or Duke Energy interconnection updates confirming NC-1 readiness. Monitor whether MTL-2 and MTL-3 announce completion. Any further date revisions without capacity commencement would deepen credibility concerns.

WhiteFiber has entered its infrastructure expansion phase with a densely clustered set of claims spanning August 16–25, 2026, all focused on North Carolina data center capacity and real estate acquisition. The stated roadmap reveals internal tension around near-term and medium-term targets that will test execution credibility almost immediately. Claim 31 commits to a minimum of 60 MW electrical capacity at acquired properties by October 30, 2026—just over two months away—while simultaneously claim 9 targets 100+ MW capacity for 2027, and claim 32 projects 198 MW of potential expansion within seven years by October 2033. The three-tier capacity ladder suggests either phased deployment or misalignment in public messaging. The 60 MW near-term commitment is anchored to a specific acquisition: claim 29 promises a $60 million real estate purchase in Yadkin County, North Carolina by the same October 30 deadline. This creates a binary verification point: either the acquisition closes and power infrastructure is in place by late October, or both timelines slip. Claim 19, referencing a purchase agreement with Unifi Manufacturing (dated August 25), appears strategically sequenced after the Yadkin County claims, suggesting potential property or operational synergy, though no delivery date is specified. Claims 24 and 25 further commit to data center expansion, property acquisition and construction, and energy service agreements, but provide no temporal anchors. Claim 27, on potential acquisitions and partnerships, is deliberately open-ended. The vagueness of these last four claims—filed four to nine days after the concrete Yadkin County commitments—signals either exploratory activity or hedging language to preserve flexibility if near-term targets slip. All claims remain unverified by permit filings, interconnection queue entries, or third-party reporting; all are stated as company assertions. The market should scrutinize October permit activity and SEC filings for the acquisition closing, then monitor Q4 2026 for permitting of electrical infrastructure and construction starts. Any delay past October 30 in the Yadkin County acquisition or absence of power capacity declarations will indicate whether the 2027 100+ MW target is achievable or subject to revision.

The company disclosed a three-year cloud service agreement with its first cloud customer on 2025-07-10 and again on 2025-07-22, cited as claims #4718 and #4695; what has not been said is who that customer is, the total contract value, the GPU or MW footprint committed, and whether the agreement carries take-or-pay or early-termination protections. Without those, a reader cannot tell whether this is an anchor tenant or a small pilot.

The company disclosed a 5 MW Cerebras colocation agreement on 2025-07-10, cited as claims #4708 and #4684; what has not been said is which facility hosts it, the term and pricing, and whether it depends on MTL-2. MTL-2 is also 5 MW (claim #4704), and its completion and revenue start, promised by 2026-01-31, is now marked delayed (claim #4705). Has the Cerebras revenue start slipped with it, and what is the revised MTL-2 date?

The company guided to 16 MW of total capacity by end of 2025 on 2025-07-10 and 2025-07-22, cited as claims #4715 and #4692; what has not been said is whether that target was met. The 16 MW appears to be exactly MTL-1 (4 MW, #4702), MTL-2 (5 MW, #4704) and MTL-3 (7 MW, #4706). With MTL-2 delayed, how much capacity was actually energized and billing at 2025-12-31? The filings also leave unclear whether these MW figures are critical IT load or gross utility capacity.

The company guided to 76 MW by end of 2026 (claims #4717 and #4654) and cited a 99 MW Duke Energy commitment by 2029-05-16 (claims #4712 and #4688); what has not been said is the bridge to 76 MW. The 16 MW from Montreal plus 40 MW at NC-1 (claims #4716 and #4653) totals 56 MW, leaving roughly 20 MW with no named site, interconnection agreement or customer. Separately, is the Duke 99 MW a signed electric service agreement with deposits posted, or a capacity reservation or letter of intent?

The company disclosed that Bit Digital would own 80 percent after the IPO (claim #4719) and contributed assets (claim #4694), and named CEO Sam Tabar (claim #4701) and director nominees David Andre and Pruitt Hall (claims #28291 and #28292); what has not been said is the terms of the intercompany agreements with Bit Digital. These include shared-services fees, transfer pricing on contributed GPUs and facilities, and how the CEO's time and compensation are split between parent and subsidiary. It is also not stated whether either nominee has prior ties to Bit Digital that bear on independence under a controlled-company structure.

The company disclosed the Enovum acquisition and the NC-1 Greensboro property purchase (claims #4703, #4678, #4709 and #4643), along with a 1,171,875-share underwriter over-allotment option (claim #4656); what has not been said is the purchase price of each asset, NC-1's capex per MW to reach 40 MW, and whether the over-allotment was exercised and for what net proceeds. Without those figures, a reader cannot judge whether the 2026 build-out is fully funded.

The company disclosed a three-year service agreement with its first cloud customer on 2025-07-10 and again on 2025-07-22, cited as claims #4718 and #4695; what has not been said is who that customer is, what the total contract value or annual revenue run-rate is, and how much GPU capacity or MW the agreement covers. Given how concentrated early revenue is likely to be, would management name the counterparty or at least state its share of revenue?

The company disclosed an MTL-2 completion target of 2025-12-31 on 2025-07-28, cited as claim #4647, and a completion and revenue start of 2026-01-31 in the 2025-07-10 materials, cited as claim #4705; both are now marked delayed, and the MTL-2 operational timeline in claim #4681 never carried a date. What has not been said is the revised in-service date for the 5 MW expansion (claims #4704, #4680, #4640), what caused the slip, and whether the delay changed the budget or any customer commitments tied to the site.

The company disclosed 76 MW of total capacity by Q4 2026 on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4717, #4693 and #4654. The named sites are MTL-1 at 4 MW (#4702), MTL-2 at 5 MW (#4704), MTL-3 at 7 MW (#4706) and NC-1 at 40 MW (#4716). Those add up to 56 MW, so what has not been said is which site or sites supply the remaining 20 MW. Separately, are these figures critical IT load or gross utility power? The Duke Energy commitments of 24, 40 and 99 MW (#4710, #4711, #4712) look like utility-side numbers.

The company disclosed a 5 MW Cerebras colocation agreement at the 7 MW MTL-3 Saint-Jérôme site on 2025-07-28, cited as claims #4642 and #4641, with earlier mentions in #4708 and #4684. What has not been said is the contract term, the pricing per kW, when MTL-3 will be energized, whether the site needs a utility upgrade from Hydro-Québec, and who will take the remaining roughly 2 MW.

The company disclosed that NC-1 in Greensboro reached its 40 MW and revenue-start milestones, cited as claims #4716, #4714 and #4690, along with a Duke Energy commitment ramping to 99 MW by 2029-05-16, cited as claims #4712, #4688 and #4646. What has not been said is who the paying tenant at NC-1 is, and what capex and financing the build from 40 MW to 99 MW will require. The record contains no debt facility or construction budget beyond the IPO.

The company disclosed an IPO of 7,812,500 shares plus a 1,171,875-share over-allotment option on 2025-07-28, cited as claims #4655 and #4656, and said Bit Digital would own about 80 percent after the offering, cited as claim #4719, following the contribution of assets described in claim #4694. What has not been said is whether the over-allotment was exercised, how the net proceeds were split across the sites, and what intercompany, services or tax-sharing agreements are in place with Bit Digital. Given that Sam Tabar is named CEO in claim #4701, is that role shared with the parent, and if so, how is compensation allocated between the two companies?

The company disclosed on 2025-07-10 a three-year service agreement with a cloud customer (claim #4718), repeated on 2025-07-22 as its "first cloud customer" (claim #4695); what has not been said is who that customer is, the contract value or annual revenue run-rate, and how many GPUs or megawatts the agreement covers. How concentrated is WhiteFiber's cloud revenue in this single unnamed counterparty?

The company disclosed a 16 MW total capacity target for year-end 2025 on 2025-07-10 and 2025-07-22 (claims #4715, #4692), which appears to equal MTL-1 at 4 MW, MTL-2 at 5 MW and MTL-3 at 7 MW (claims #4702, #4704, #4706). What has not been said is how that target was met when MTL-2 missed both its 2025-12-31 and 2026-01-31 completion dates (claims #4647, #4705, both now delayed), and whether the 16 MW figure was ever restated. What is MTL-2's revised in-service date, and what caused the slip?

The company disclosed a 76 MW capacity target for Q4 2026 on 2025-07-10, 2025-07-22 and 2025-07-28 (claims #4717, #4693, #4654). The identified sites (16 MW in Montreal plus 40 MW at NC-1, claims #4691, #4716) account for only about 56 MW, so what has not been said is which site supplies the remaining roughly 20 MW. Is the target stated in critical IT load or in gross utility capacity?

The company disclosed Duke Energy commitments of 24 MW, 40 MW and 99 MW at NC-1 Greensboro (claims #4710, #4711, #4712, repeated as #4645 and #4646), with the 99 MW tranche promised by 2029-05-16. What has not been said is whether these tranches are cumulative or incremental, whether the 99 MW is a signed electric service agreement or only a capacity reservation, and what capital expenditure and financing the NC-1 build-out beyond 40 MW requires. Beyond naming the city, has the company identified the NC-1 site's parcel, substation or interconnection point (claims #4709, #4643)?

The company disclosed on 2025-07-28 a 5 MW Cerebras colocation agreement at MTL-3, a 7 MW Saint-Jérôme site (claims #4642, #4641, also #4708, #4684); what has not been said is the contract term, pricing per kW, ramp schedule, or when MTL-3 is expected to energize. MTL-3 has a stated capacity but no completion date on record, so when does Cerebras revenue begin?

The company disclosed on 2025-07-28 an IPO of 7,812,500 shares with a 1,171,875-share over-allotment option (claims #4655, #4656), and said Bit Digital would retain 80 percent ownership after the IPO (claim #4719). What has not been said is whether the over-allotment was exercised, what the net proceeds were and how they were allocated, the terms of intercompany agreements arising from the Bit Digital contribution (claim #4694), and the Enovum purchase price (claims #4703, #4678, #4638). The company also named Sam Tabar as CEO (claims #4701, #4696), who also leads Bit Digital; how is executive time and compensation split between the controlling parent and the subsidiary?

The company disclosed a three-year service agreement with its first cloud customer on 2025-07-10 and 2025-07-22, cited as claims #4718 and #4695; what has not been said is who the customer is, the total contract value, the GPU count or capacity committed, and whether the agreement can be terminated or repriced before the three years end. Without these, revenue concentration cannot be assessed.

The company set an MTL-2 completion target of 2025-12-31 on 2025-07-28 (claim #4647) and a completion-and-revenue-start target of 2026-01-31 on 2025-07-10 (claim #4705), and both are now marked delayed, while the 2025-07-22 operational timeline (claim #4681) gives no date at all. What has not been said is the current expected energization date for the 5 MW site (claims #4704, #4680, #4640), the cause of the slip, and how much revenue was deferred. Has the 16 MW end-2025 target (claims #4715, #4692) been formally withdrawn, given that it appears to depend on MTL-2?

The company disclosed Duke Energy capacity commitments of 24 MW, 40 MW and 99 MW on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4710, #4711, #4712, #4686, #4687, #4688, #4645 and #4646; what has not been said is whether these figures are utility interconnection capacity or critical IT load. The interconnection agreement, substation details and any take-or-pay or deposit obligations behind the 99 MW tranche dated 2029-05-16 are also undisclosed.

The company guided to 76 MW of total HPC capacity by Q4 2026 in claims #4717, #4693 and #4654. The site figures it has disclosed are MTL-1 at 4 MW (#4702), MTL-2 at 5 MW (#4704), MTL-3 at 7 MW (#4706) and NC-1 at 40 MW (#4716, #4691), which add up to 56 MW. What has not been said is which site or phase supplies the remaining 20 MW, and whether that capacity is owned, leased or still unsecured.

The company disclosed a 5 MW Cerebras colocation agreement at the 7 MW MTL-3 Saint-Jérôme site on 2025-07-28, cited as claims #4642, #4708 and #4684; what has not been said is the contract term, pricing per kW, the start date of billing, and who will use the remaining 2 MW. The physical address and grid interconnection status of MTL-3 (#4641, #4682) also remain undisclosed, as do the purchase prices paid for the NC-1 Greensboro property (#4709, #4685, #4643) and for Enovum (#4703, #4678, #4638).

The company disclosed that Bit Digital would own 80 percent after the IPO and contributed the business, cited as claims #4719 and #4694. It also disclosed an IPO of 7,812,500 shares with a 1,171,875-share over-allotment, cited as claims #4655 and #4656. What has not been said is the terms of the intercompany, transition-services and cost-allocation agreements with Bit Digital, and whether the over-allotment was exercised. Also undisclosed is how CEO Sam Tabar's compensation (#4701, #4696) is split between WhiteFiber and its controlling parent.

WhiteFiber disclosed on 2025-07-10 a three-year service agreement with a cloud customer (claim #4718), repeated on 2025-07-22 as its "first cloud customer" (claim #4695). What has not been said is who that customer is, what the contract is worth, and how many megawatts or GPUs it covers. Is this the same counterparty as any of the MTL-1 customers referenced in claim #4639, and how concentrated is revenue in this one relationship?

The company disclosed Duke Energy capacity commitments of 24 MW by 2025-09-01, 40 MW by 2026-04-01 and 99 MW by 2029-05-16 (claims #4710, #4711, #4712, restated as #4686–#4688 and #4644/#4646). What has not been said is whether these tranches are cumulative or incremental, whether they are firm interconnection agreements or non-binding allocations, and what substation or service voltage feeds NC-1. Given that the 24 MW power date passed months before NC-1's own 24 MW build completion target of 2026-03-31 (claim #4713), was that power energized, and is the company paying for capacity it cannot yet use?

The company guided to 16 MW by end-2025 (claims #4715, #4692, #4649), which matches MTL-1 at 4 MW, MTL-2 at 5 MW and MTL-3 at 7 MW (claims #4702, #4704, #4706). It also guided to 40 MW at NC-1 by end Q2 2026 (claim #4716) and 76 MW total by end Q4 2026 (claims #4717, #4693, #4654). What has not been said is where the remaining roughly 20 MW comes from, since 16 plus 40 equals 56. Is there an undisclosed site, a further NC-1 phase, or an expansion at an existing Montreal facility? Separately, the MTL-2 completion target of 2026-01-31 (claim #4705) and the end-2025 16 MW target have both passed with no confirming disclosure on record. Were they met?

The company disclosed the NC-1 property acquisition in Greensboro, North Carolina, dated 2025-05-20 (claims #4709, #4685, #4643). What has not been said is the street address or parcel, the purchase price, or whether the building is a conversion or a new build. The NC-1 revenue start date is also stated inconsistently, as 2026-05-01 in claims #4714 and #4690 but 2026-05-31 in claim #4651. Which date is the company's actual commitment, and is any NC-1 capacity pre-leased?

The company disclosed a 5 MW Cerebras colocation agreement at the 7 MW MTL-3 site (claims #4708, #4684, #4642, #4706). What has not been said is the contract term, the pricing per kW, any ramp schedule, or who, if anyone, has contracted the remaining 2 MW. Are there termination or step-down rights that would let Cerebras exit before the build-out is recovered?

The company disclosed that Bit Digital would own about 80 percent of WhiteFiber after the IPO (claims #4719, #4694). The IPO was sized at 7,812,500 shares with a 1,171,875-share over-allotment option (claims #4655, #4656), and Sam Tabar is named as CEO (claim #4701). What has not been said is the terms of the intercompany services, asset-contribution and tax agreements with the controlling parent. The record also does not say whether the CEO holds a concurrent role at Bit Digital, how that individual's time and compensation are split between the two companies, or whether the over-allotment was exercised and what net proceeds were raised.

The company disclosed a three-year service agreement with its first cloud customer on 2025-07-10 and again on 2025-07-22, cited as claims #4718 and #4695; what has not been said is who that customer is, what the contract is worth, and how many GPUs or megawatts it covers. Without these, a reader cannot judge how concentrated revenue is or how durable the business is.

The company disclosed Duke Energy capacity commitments of 24 MW by 2025-09-01, 40 MW by 2026-04-01 and 99 MW by 2029-05-16, cited as claims #4710, #4711, #4712, #4686, #4687, #4644, #4645 and #4646; what has not been said is whether these are firm or interruptible service, whether the figures are utility-side or critical IT megawatts, whether the 99 MW is all at the Greensboro NC-1 site, and what the revised energization dates are now that the first two phases are marked delayed.

The company disclosed a target of 76 MW total capacity by Q4 2026 on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4717, #4693 and #4654, and that target is still on record as unrevised; what has not been said is how it squares with the named sites. MTL-1, MTL-2 and MTL-3 (claims #4702, #4704, #4706) plus NC-1 at 40 MW (claim #4691) add up to only 56 MW, so roughly 20 MW is unattributed. It is also unclear how the 76 MW figure survives when the interim milestones (16 MW by end-2025 in #4715, 24 MW at NC-1 by Q1 2026 in #4713, 40 MW at NC-1 by Q2 2026 in #4716) are all delayed.

The company gave conflicting completion dates for MTL-2 and MTL-3: 2025-12-31 on 2025-07-28 (claims #4647 and #4648) but 2026-01-31 on 2025-07-10 (claims #4705 and #4707). It also gave conflicting NC-1 revenue start dates of 2026-05-01 (claims #4714 and #4690) and 2026-05-31 (claim #4651). What has not been said is which dates are authoritative, what the current revised dates are now that all of them are marked delayed, and who the contracted tenant at NC-1 is, since no NC-1 customer appears on record.

The company disclosed a 5 MW Cerebras colocation agreement at the 7 MW MTL-3 site, cited as claims #4708, #4684 and #4642; what has not been said is the term, the pricing per kW, whether there are penalty or termination rights tied to the MTL-3 delay (claim #4707), and whether the remaining 2 MW at MTL-3 is contracted.

The company disclosed that Bit Digital will hold about 80 percent of the company after the IPO and contributed assets to it, cited as claims #4719 and #4694, alongside an IPO of 7,812,500 shares with a 1,171,875-share over-allotment option, cited as claims #4655 and #4656. What has not been said is the terms of any intercompany services, leases or transition agreements with the parent, whether executives including CEO Sam Tabar (claim #4701) are shared with or paid by Bit Digital, whether the over-allotment option was exercised, and how the net proceeds were split between the delayed NC-1 build-out and the Montreal builds.

The company disclosed a three-year service agreement with its first cloud customer on 2025-07-10 and again on 2025-07-22, cited as claims #4718 and #4695. What has not been said is who the customer is, what the total contract value or annual run-rate is, how many GPUs or megawatts it covers, and which facility serves it. Without these, can a reader tell whether this is a single concentrated counterparty underpinning the cloud segment?

The company disclosed Duke Energy capacity commitments of 24 MW by 2025-09-01, 40 MW by 2026-04-01 and 99 MW by 2029-05-16, cited as claims #4710, #4711 and #4712 (restated as #4644, #4645 and #4646). Separately, claim #4653 calls the NC-1 40 MW figure 'cumulative.' What has not been said is whether the Duke tranches are cumulative or additive, so is NC-1's eventual utility-backed load 99 MW or 163 MW? With the first two tranches now marked delayed, the company also has not said whether the bottleneck is Duke's interconnection or its own construction.

The company guided on 2025-07-10 to 76 MW of total capacity by end of 2026, cited as claim #4717 and reiterated as #4693 and #4654. The named sites do not add up to that figure: MTL-1 is 4 MW (#4702), MTL-2 is 5 MW (#4704), MTL-3 is 7 MW (#4706) and NC-1 is 40 MW (#4716), for a total of 56 MW. What has not been said is where the remaining roughly 20 MW comes from. Also unexplained is why the 76 MW target still stands when the 2025 16 MW guide (#4715), the NC-1 24 MW and 40 MW milestones (#4713, #4716) and NC-1 revenue start (#4714, #4651) have all slipped.

The company disclosed MTL-2 and MTL-3 completion and revenue start by 2026-01-31 on 2025-07-10, cited as claims #4705 and #4707. Eighteen days later it gave 2025-12-31 for the same milestones, cited as claims #4647 and #4648, and both are now delayed. What has not been said is a revised in-service date for either site. The Cerebras 5 MW colocation agreement at MTL-3 (#4708, #4642) has also gone without its term, pricing, or a date when billing begins. Should investors assume Cerebras revenue has not started?

The company disclosed the NC-1 Greensboro property acquisition, cited as claims #4709, #4685 and #4643, with a promised date of 2025-05-20 that is now marked delayed. What has not been said is whether the purchase actually closed and at what price. The company also has not given a street address or parcel identifier, or named the substation and interconnection point that the Duke commitments attach to.

The company disclosed that Bit Digital would own 80 percent of the company after the IPO and contributed assets to it, cited as claims #4719 and #4694. It also disclosed an IPO of 7,812,500 shares with a 1,171,875-share over-allotment, cited as claims #4655 and #4656, and named Sam Tabar as CEO, cited as claim #4701. What has not been said is the terms of the intercompany services, lease and capital agreements with the controlling parent, and whether the over-allotment was exercised. Nor has the company explained how the CEO's time and compensation are split if he also serves the parent.

The company disclosed on 2025-07-10 a three-year service agreement with a cloud customer, cited as claim #4718 and repeated as 'first cloud customer' on 2025-07-22 (claim #4695); what has not been said is who that customer is, the contract value, the GPU or MW footprint it covers, and whether it is concentrated enough to be a material single-customer exposure.

The company disclosed a 5 MW Cerebras colocation agreement on 2025-07-10 and 2025-07-28, cited as claims #4708, #4684 and #4642, sited at MTL-3. MTL-3's completion and revenue start, promised by 2026-01-31 (claim #4707) and earlier by 2025-12-31 (claim #4648), are now flagged as delayed. What has not been said is the term, pricing, and ramp schedule of the Cerebras contract, and whether the MTL-3 slippage has triggered any delay penalties, renegotiation, or termination rights for Cerebras.

The company disclosed Duke Energy capacity commitments of 24 MW by 2025-09-01, 40 MW by 2026-04-01, and 99 MW by 2029-05-16, cited as claims #4710, #4711 and #4712 (and again as #4644, #4645 and #4646). What has not been said is whether these tranches are cumulative or incremental, whether the figures are utility load or critical IT capacity, and whether the 24 MW and 40 MW tranches have actually been energized now that both dates have passed and are flagged as delayed. The company also has not disclosed the substation or interconnection detail for the Greensboro NC-1 site acquired under claims #4709 and #4643.

The company disclosed 76 MW of total capacity by end of 2026, cited as claims #4717, #4693 and #4654. The named sites do not add up to that figure: MTL-1 at 4 MW (claim #4702), MTL-2 at 5 MW (claim #4704), MTL-3 at 7 MW (claim #4706), and NC-1 at 40 MW by Q2 2026 (claim #4716) total 56 MW. What has not been said is which site or tranche supplies the remaining roughly 20 MW, or whether the 76 MW guidance still stands given that the 16 MW year-end 2025 target (claims #4715, #4692, #4649) is flagged as delayed.

The company disclosed inconsistent milestone dates within the same month. MTL-2 and MTL-3 completion was 2026-01-31 on 2025-07-10 (claims #4705, #4707) but 2025-12-31 on 2025-07-28 (claims #4647, #4648). NC-1 revenue start was 2026-05-01 (claims #4714, #4690) versus 2026-05-31 (claim #4651). What has not been said is which date set is authoritative, and what revised dates have replaced every one of these now-missed milestones.

The company disclosed that Bit Digital would own 80 percent post-IPO and contributed the business, cited as claims #4719 and #4694, with Sam Tabar named CEO (claims #4701, #4696, #4657). What has not been said is the terms of any intercompany services, lease, or GPU-transfer agreements with the parent, whether Sam Tabar holds a concurrent role at Bit Digital, and how Sam Tabar's compensation is split between the two entities. The company also has not disclosed the purchase consideration paid for Enovum and the NC-1 property (claims #4703, #4678, #4709).

The company disclosed on 2025-07-10 a three-year service agreement with a cloud customer, cited as claim #4718 and repeated on 2025-07-22 as its "first cloud customer" in claim #4695; what has not been said is who that customer is, the total contract value, the GPU count or MW footprint it covers, and whether it carries early-termination or take-or-pay protections. How concentrated is cloud revenue in this single unnamed counterparty?

The company disclosed a three-phase Duke Energy capacity arrangement of 24 MW, 40 MW and 99 MW on 2025-07-10 and 2025-07-28, cited as claims #4710, #4711, #4712, #4644, #4645 and #4646. What has not been said is whether these are executed electric service agreements or non-binding load letters, and whether the figures are utility-delivered MW or critical IT MW. The Phase 1 date of 2025-09-01 (claim #4710) and the Phase 2 date of 2026-04-01 (claim #4711) have both slipped. What is the revised energization schedule, and which party is responsible for the delay?

The company disclosed the NC-1 property purchase in Greensboro, North Carolina, cited as claims #4709, #4685 and #4643, along with a 24 MW initial and 40 MW phased build, cited as claims #4689 and #4691. What has not been said is the street address or parcel, the purchase price, the substation or interconnection point serving the site, and the capex budget per MW. The 24 MW completion (claim #4713), the 40 MW Q2 2026 target (claim #4716) and the May 2026 revenue start (claims #4714 and #4651) are all now past due. Has any MW at NC-1 actually been energized?

The company disclosed 2026 total capacity guidance of 76 MW on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4717, #4693 and #4654, and that guidance is still standing. Adding up the named sites gives about 56 MW: MTL-1 at 4 MW (claim #4702), MTL-2 at 5 MW (claim #4704), MTL-3 at 7 MW (claim #4706) and NC-1 at 40 MW (claim #4691). What has not been said is which site or sites supply the remaining roughly 20 MW. Given that the 16 MW end-2025 target was missed (claims #4715, #4692 and #4649), is the 76 MW figure being reaffirmed or quietly abandoned?

The company disclosed a 5 MW Cerebras colocation agreement at MTL-3, cited as claims #4708, #4684 and #4642. MTL-3's completion and revenue start have slipped past January 2026 (claims #4707 and #4648). What has not been said is the contract term, the pricing per kW, the committed start date, and whether the delay triggers credits, penalties or termination rights for Cerebras.

The company disclosed that Bit Digital would own 80 percent after the IPO, cited as claim #4719, that Bit Digital contributed the business (claim #4694), and an offering of 7,812,500 shares plus a 1,171,875-share over-allotment, cited as claims #4655 and #4656. What has not been said is the IPO price, the net proceeds, whether the over-allotment was exercised, and the terms of the intercompany services, tax-sharing and transition agreements with the parent. With Sam Tabar named as CEO (claim #4701), is any of the CEO's compensation borne by Bit Digital, and how is it allocated between the two companies?

The company disclosed a three-year service agreement with its first cloud customer on 2025-07-10 and 2025-07-22, cited as claims #4718 and #4695; what has not been said is who that customer is, the total contract value, the GPU count or MW allocated to it, and whether revenue under the agreement is take-or-pay or usage-based.

The company disclosed a Cerebras colocation agreement for 5 MW at MTL-3 on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4708, #4684 and #4642; what has not been said is the contract term, the pricing per kW, and whether Cerebras holds termination or penalty rights given that MTL-3 missed its completion and revenue-start dates of 2025-12-31 and 2026-01-31 (claims #4648 and #4707). The company also has not said how the remaining 2 MW of MTL-3's 7 MW (claim #4706) is contracted.

The company disclosed Duke Energy power commitments of 24 MW by 2025-09-01, 40 MW by 2026-04-01 and 99 MW by 2029-05-16 for NC-1 Greensboro, cited as claims #4710, #4711 and #4712 (repeated in #4644, #4645 and #4646); what has not been said is whether these are executed electric service agreements or only capacity reservations, which substation or interconnection point serves the site, and why the first two tranches slipped. The company also has not said whether the MW figures are utility feed or critical IT load. If the 24 MW Duke tranche is utility power, the 24 MW of NC-1 capacity (claims #4689 and #4652) cannot all be IT load once cooling overhead is counted.

The company guided to 76 MW total capacity by end of 2026 on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4717, #4693 and #4654; what has not been said is which sites make up that number. MTL-1, MTL-2, MTL-3 and NC-1 at 40 MW (claims #4702, #4704, #4706 and #4716) total 56 MW, leaving roughly 20 MW unattributed. The company also has not updated the 76 MW figure even though every interim milestone that leads to it, including 16 MW by end of 2025 (claims #4715, #4692 and #4649) and 40 MW at NC-1 by Q2 2026 (claims #4716 and #4653), is marked delayed.

The company gave conflicting completion dates for MTL-2 and MTL-3: 2025-12-31 on 2025-07-28 (claims #4647 and #4648) and 2026-01-31 on 2025-07-10 (claims #4705 and #4707). Separately, NC-1's revenue start moved from 2026-05-01 (claims #4714 and #4690) to 2026-05-31 (claim #4651). What has not been said is the actual energized capacity and the revenue-generating MW at each site as of the latest quarter end.

The company disclosed an IPO of 7,812,500 shares plus a 1,171,875-share over-allotment option on 2025-07-28, cited as claims #4655 and #4656, and said Bit Digital would retain 80% ownership after the IPO, cited as claims #4719 and #4694; what has not been said is whether the over-allotment option was exercised, the net proceeds, and how much of those proceeds is committed to NC-1 capex. The company also has not described the shared-services, intercompany pricing and executive-compensation arrangements with its controlling parent, given that CEO Sam Tabar (claims #4701, #4696 and #4657) also leads Bit Digital.

The company disclosed a three-year service agreement with its first cloud customer on 2025-07-10 and again on 2025-07-22, cited as claims #4718 and #4695. What has not been said is who that customer is, what the total contract value or annual run-rate is, and how much GPU or MW capacity it covers. Without that, a reader cannot judge customer concentration in a business that is still mostly pre-revenue.

The company disclosed a staged Duke Energy power arrangement of 24 MW by 2025-09-01, 40 MW by 2026-04-01 and 99 MW by 2029-05-16, cited as claims #4710, #4711, #4712 and #4644 through #4646. The first two milestones are now marked delayed. What has not been said is whether these are firm service commitments or only interconnection studies, whether the figures are utility MW or critical IT MW, and what Duke has actually energized to date.

The company disclosed the NC-1 property purchase in Greensboro, cited as claims #4709 and #4643. It also disclosed 24 MW energization by Q1 2026 (#4652, #4713), 40 MW by Q2 2026 (#4716, #4653) and revenue start by 2026-05-01 (#4714). All of these are now delayed. What has not been said is the site's street address and purchase price, the substation or interconnection point it relies on, a revised energization date, and whether any tenant has signed for NC-1 capacity.

The company guided to 16 MW by end-2025 (#4715, #4692, #4649) and to 76 MW by end-Q4 2026 (#4717, #4654, #4693). The 2026 target is still carried as stated even though every NC-1 phase is delayed. MTL-1, MTL-2 and MTL-3 sum to 16 MW (#4702, #4704, #4706), and NC-1's disclosed phase reaches 40 MW, so roughly 20 MW of the 76 MW target is unattributed to any named site. What has not been said is which facility supplies that 20 MW, whether the 76 MW is gross or IT load, and whether management still stands behind the figure.

The company disclosed a 5 MW Cerebras colocation agreement at MTL-3, cited as claims #4708, #4684 and #4642. MTL-3 is a 7 MW site whose completion, promised for end-2025 or 2026-01-31, is now delayed (#4707, #4648). What has not been said is the contract term and pricing, whether the delay triggers credits, penalties or termination rights, and how the remaining 2 MW is allocated.

The company disclosed that Bit Digital would hold about 80 percent after the IPO (#4719, #4694). It also disclosed an offering of 7,812,500 shares plus a 1,171,875-share over-allotment (#4655, #4656) and named Sam Tabar as CEO (#4701, #4696). What has not been said is the terms of the intercompany services, lease or asset-contribution agreements with the parent, and how executive time and compensation are split if leadership overlaps with Bit Digital. Also missing are the offering price, net proceeds and how those proceeds map to the delayed NC-1 and Montreal buildouts.

The company disclosed a three-year service agreement with its first cloud customer on 2025-07-10 and 2025-07-22, cited as claims #4718 and #4695; what has not been said is who that customer is, the total contract value, how many GPUs or MW it covers, and what share of revenue it represents. Given the IPO-era concentration, why is the counterparty still unnamed?

The company disclosed a 5 MW Cerebras colocation agreement at MTL-3 on 2025-07-10 and 2025-07-28, cited as claims #4708, #4684 and #4642, while MTL-3's completion and revenue start, promised by 2026-01-31 (claim #4707) and earlier by 2025-12-31 (claim #4648), is now marked delayed. What has not been said is the contract's term, pricing, and whether the slip triggers delivery-date penalties, termination rights, or deferred revenue recognition with Cerebras.

The company guided to 76 MW by end of Q4 2026 on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4717, #4693 and #4654, and that guidance remains stated. The identified pieces are MTL-1 (4 MW, claim #4702), MTL-2 (5 MW, claim #4704), MTL-3 (7 MW, claim #4706) and NC-1 at 40 MW (claim #4716), which sum to 56 MW. Where are the other 20 MW, and are these figures critical IT load or gross utility capacity? The company also has not reconciled how the 76 MW target survives when the 16 MW end-2025 target (claims #4715, #4692, #4649) and the 40 MW NC-1 Q2 2026 target (claim #4716) are both delayed. What capacity was actually energized at 2025 year-end?

The company disclosed Duke Energy power commitments of 24 MW by 2025-09-01, 40 MW by 2026-04-01 and 99 MW by 2029-05-16, cited as claims #4710, #4711 and #4712 (repeated as #4686–#4688 and #4644–#4646). The first two are now delayed. What has not been said is whether these are executed electric service agreements or non-binding load letters, what deposits or collateral were posted, and whether the delays sit with Duke or with WhiteFiber. NC-1 itself was described only as a Greensboro property (claims #4709 and #4643), with no street address, acreage, substation, or interconnection point. Which one is it?

The company disclosed that Bit Digital would own 80.0 percent after the IPO and contributed the business, cited as claims #4719 and #4694, and named Sam Tabar as CEO, cited as claim #4701. What has not been said is the terms of the intercompany, shared-services and tax agreements with the parent, and whether Tabar's compensation and time are split with Bit Digital. On what basis does the board approve related-party dealings with an 80 percent holder?

The company disclosed an IPO of 7,812,500 shares with a 1,171,875-share over-allotment option on 2025-07-28, cited as claims #4655 and #4656, along with acquisitions of Enovum, MTL-2, MTL-3 and NC-1 (claims #4703, #4680, #4682 and #4685). The record contains no offering price, net proceeds, exercise status for the over-allotment, or purchase price for any site. How much capital was raised, and how much of it is earmarked for the delayed NC-1 build versus repaying the parent?

The company disclosed a three-year cloud service agreement on 2025-07-10, cited as claim #4718, and referred to its first cloud customer on 2025-07-22, cited as claim #4695. What has not been said is who that customer is, what the contract value or annual revenue run-rate is, and how much GPU or MW capacity it consumes. By contrast, Cerebras was named as a colocation customer (claims #4708, #4684, #4642). Is the cloud counterparty withheld because it is concentrated, related, or immaterial?

The company disclosed a 76 MW total capacity target by end of Q4 2026 on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4717, #4693 and #4654, and these remain in stated status. What has not been said is how that target can be reconciled with the delays in nearly every building block: the 16 MW end-2025 guidance (#4715, #4692, #4649), the MTL-2 and MTL-3 completions (#4705, #4707, #4647), and NC-1's 24 MW and 40 MW milestones (#4713, #4716, #4653). The company has also not said whether its MW figures are gross utility capacity or critical IT load. What MW was actually energized and billing at the end of Q2 and Q3 2026, and is the 76 MW figure still reaffirmed?

The company disclosed a phased Duke Energy power arrangement of 24 MW by 2025-09-01, 40 MW by 2026-04-01 and 99 MW by 2029-05-16, cited as claims #4710, #4711, #4712, #4686, #4687, #4688, #4644 and #4646. The first two dates are now delayed. What has not been said is whether these are firm interconnection or service agreements or only capacity reservations, what substation or interconnection queue position serves NC-1, and whether the delay sits with Duke or with WhiteFiber's own build. Which party caused the slippage, and does it put the 2029 99 MW tranche at risk?

The company disclosed the NC-1 property acquisition in Greensboro, North Carolina, cited as claims #4709, #4685 and #4643, with revenue start promised by May 2026 (#4714, #4690, #4651). Both the acquisition date and the revenue start are now delayed. What has not been said is the site's physical address or parcel, the purchase price, the capex budget per MW for the conversion, and whether a tenant has been signed for NC-1 capacity. Is NC-1 being built speculatively, and how is it being funded?

The company disclosed that Bit Digital would own 80 percent after the IPO, cited as claim #4719, along with a Bit Digital asset contribution (#4694) and Sam Tabar as CEO (#4701). What has not been said is the terms of the intercompany services, transition, and asset-contribution agreements with the controlling parent. The company also has not said how the CEO's time and compensation are split if they hold roles at both entities. What related-party payments flow between WhiteFiber and Bit Digital, and what protections exist for minority holders under the controlled-company exemption?

The company disclosed an IPO of 7,812,500 shares plus a 1,171,875-share over-allotment, cited as claims #4655 and #4656. Separately, the Enovum acquisition (#4703, #4678, #4638) and the 7 MW MTL-3 site in Saint-Jérôme (#4706, #4682, #4641) carry delayed dates. What has not been said is the offering price, net proceeds, how much of those proceeds is earmarked for NC-1 versus the Montreal sites, and whether additional debt or equity is needed to reach 76 MW. What is the remaining funding gap?

WhiteFiber disclosed a three-year service agreement with its first cloud customer on 2025-07-10 and 2025-07-22, cited as claims #4718 and #4695. What has not been said is who that customer is, what the total contract value or annualized revenue is, and how many GPUs or how much capacity the agreement covers. Without these, a reader cannot judge concentration risk in what appears to be the company's anchor cloud revenue. The 2025-07-28 disclosure that MTL-1 has operating customers, cited as claim #4639, also names none of them.

The company disclosed a 5 MW Cerebras colocation agreement at MTL-3 on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4708, #4684 and #4642, against a 7 MW site, cited as claim #4706. What has not been said is the term, the pricing per kW, and whether the 5 MW is critical IT load or gross power. The company also has not said who, if anyone, is contracted for the remaining 2 MW. MTL-3 completion was guided for 2025-12-31 or 2026-01-31, cited as claims #4648 and #4707, and both are now delayed. Has the Cerebras revenue start moved with it, and does the contract carry delivery penalties or termination rights?

The company disclosed a target of 76 MW of capacity by Q4 2026, cited as claims #4717, #4693 and #4654. It also disclosed Duke Energy commitments of 24, 40 and 99 MW, cited as claims #4710, #4711, #4712, #4686, #4687, #4688, #4644, #4645 and #4646. What has not been said is whether these figures are utility interconnection capacity or billable critical IT load. That distinction can move revenue capacity by 30% or more after PUE. The 16 MW year-end 2025 target, cited as claims #4715, #4692 and #4649, is marked delayed. The company has not stated what capacity was actually energized and revenue-generating at 2025-12-31, or how the 76 MW target reconciles with that shortfall.

The company disclosed that Duke Energy would deliver 24 MW to NC-1 in Greensboro by 2025-09-01 and 40 MW by 2026-04-01, cited as claims #4710 and #4711. Both are now delayed. What has not been said is whether the first 24 MW has been energized at all, and what substation and service agreement type back the interconnection. It is also unclear whether the 99 MW commitment by 2029-05-16, cited as claim #4712, is firm or subject to Duke upgrades the company must fund. Separately, the company gave the NC-1 revenue start as 2026-05-01 in claims #4714 and #4690 but as 2026-05-31 in claim #4651. Which date is correct, and what is the current one?

The company disclosed the NC-1 Greensboro property purchase, the Enovum acquisition and the Bit Digital contribution of assets, cited as claims #4709, #4685, #4643, #4703, #4678 and #4694. What has not been said is the purchase price or consideration for any of these transactions. The company also has not disclosed the capex budget for building NC-1 out to 40 MW and 99 MW. Relatedly, it disclosed an IPO of 7,812,500 shares with a 1,171,875-share over-allotment option, cited as claims #4655 and #4656. The offering price, the net proceeds, and whether the over-allotment was exercised are absent from the record. So is how the proceeds were allocated against this buildout.

The company disclosed on 2025-07-10 that Bit Digital would own 80% of WhiteFiber after the IPO, and named Sam Tabar as CEO, cited as claims #4719 and #4701. What has not been said is the terms of any intercompany services, lease, GPU supply or cost-sharing agreements with the parent. Nor has the company said whether the CEO holds a concurrent role at Bit Digital, or how his compensation is split and set given the parent's controlling stake.

The company disclosed a three-year cloud service agreement on 2025-07-10 (claim #4718) and again described a "first cloud customer" on 2025-07-22 (claim #4695). What has not been said is who that customer is, the contract's total value or annual revenue run-rate, and how many GPUs or megawatts it occupies. Is that agreement the bulk of current cloud revenue?

The company disclosed a Cerebras colocation agreement for 5.0 MW at MTL-3 (claims #4708, #4684, #4642). MTL-3's completion and revenue start has been marked delayed, from a 2025-12-31 target (claim #4648) and a 2026-01-31 target (claim #4707). What has not been said is the new energization date, whether the Cerebras contract carries delivery-date penalties or termination rights, and the contract's term and pricing. The same question applies to MTL-2, whose completion targets (claims #4647, #4705) have also slipped without a replacement date.

The company disclosed Duke Energy power commitments of 24 MW by 2025-09-01, 40 MW by 2026-04-01 and 99 MW by 2029-05-16 (claims #4710, #4711, #4712, repeated as #4686–#4688 and #4645–#4646). The first two dates have passed and are marked delayed. What has not been said is whether these figures are cumulative or incremental, whether the 99 MW tranche is a signed interconnection agreement or a study-stage request, and the revised dates for the 24 MW and 40 MW tranches. NC-1 in Greensboro (claims #4709, #4643) has also been described without a street address, substation or interconnection-queue identifier.

The company guided to 16 MW of total capacity by end of 2025 (claims #4715, #4692) and 40 MW at NC-1 by end of Q2 2026 (claims #4716, #4653). Both are marked delayed. Yet the 76 MW target for Q4 2026 (claims #4717, #4693, #4654) is still carried as stated. What has not been said is whether 76 MW has been formally reaffirmed or withdrawn, and whether the figure means critical IT load or gross utility power. How many megawatts were actually energized and billing at each quarter-end?

The company disclosed the IPO share count of 7,812,500 shares plus a 1,171,875-share over-allotment option (claims #4655, #4656). It has not said, alongside those claims, the net proceeds, whether the option was exercised, or how proceeds are split between the NC-1 buildout and other uses. It also hasn't said what additional debt or equity it needs to take NC-1 from 24 MW to 99 MW.

The company disclosed that Bit Digital retains 80.0 percent ownership after the IPO (claims #4719, #4694) and named Sam Tabar as CEO (claims #4701, #4696, #4657). What has not been said is the terms of the intercompany services, the transitional agreements and the asset-contribution agreement with the parent. It also hasn't said whether the CEO holds a concurrent role at Bit Digital, and if so, how compensation and time are allocated between the two companies.

The company disclosed a three-year service agreement with its first cloud customer on 2025-07-10 and 2025-07-22, cited as claims #4718 and #4695; what has not been said is who that customer is, the contract value or minimum commitment, and what share of current revenue it represents. Is WhiteFiber dependent on a single unnamed counterparty?

The company disclosed Duke Energy capacity commitments of 24 MW, 40 MW and 99 MW on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4710, #4711, #4712, #4686, #4687, #4688, #4644, #4645 and #4646; what has not been said is whether these figures are utility interconnection capacity or critical IT load, whether the phases are cumulative or additive, and why the 24 MW phase due 2025-09-01 and the 40 MW phase due 2026-04-01 slipped. Has Duke actually energized any capacity at NC-1, and has a revised energization date been agreed in writing?

The company disclosed guidance of 16 MW by end-2025 and 76 MW by end-Q4 2026 on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4715, #4692, #4649, #4717, #4693 and #4654; what has not been said is a site-by-site bridge to 76 MW now that the 2025 target, the NC-1 24 MW Q1 2026 milestone (#4713, #4650, #4652) and the NC-1 40 MW Q2 2026 milestone (#4716, #4653) have all been missed. The disclosed Montreal sites total only 16 MW across MTL-1, MTL-2 and MTL-3 (#4702, #4704, #4706). Is the 76 MW target still intact, and is it stated on a gross or IT-load basis?

The company disclosed a 5 MW Cerebras colocation agreement at MTL-3 on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4708, #4684 and #4642. MTL-3 completion and revenue start, promised by 2026-01-31 in claim #4707, is delayed. What has not been said is the contract term, pricing per kW, and whether Cerebras holds delay penalties, termination rights or service credits tied to the missed delivery date.

The company disclosed that Bit Digital would own 80 percent post-IPO and contributed assets including Enovum, on 2025-07-10 and 2025-07-22, cited as claims #4719, #4694, #4703 and #4678, and named Sam Tabar as CEO in claim #4701. What has not been said is the terms of the intercompany services, transition and tax agreements with the parent. Does the CEO also hold a role at Bit Digital? If so, how is their time and compensation allocated between the two companies?

The company disclosed the NC-1 property purchase in Greensboro and an IPO of 7,812,500 shares plus a 1,171,875-share over-allotment, on 2025-07-10 and 2025-07-28, cited as claims #4709, #4643, #4685, #4655 and #4656. What has not been said is NC-1's street address, its interconnection substation, the IPO's net proceeds, how much of those proceeds is earmarked for NC-1 build-out, and what additional financing is required to reach 99 MW by 2029.

The company disclosed on 2025-07-10 a three-year service agreement with a cloud customer, cited as claim #4718, and on 2025-07-22 referenced its first cloud customer, cited as claim #4695. What has not been said is who that customer is, the total contract value, the number of GPUs or MW it covers, and whether there are termination-for-convenience or take-or-pay provisions. How concentrated is cloud revenue in this single unnamed counterparty?

The company disclosed Duke Energy capacity commitments of 24 MW by 2025-09-01 (claims #4710, #4686, #4644), 40 MW by 2026-04-01 (claims #4711, #4687) and 99 MW by 2029-05-16 (claims #4712, #4688, #4646). What has not been said is whether these are executed interconnection or electric service agreements or non-binding load letters, and whether the figures are utility-delivered gross MW or critical IT load. With the first two tranches now marked delayed, has Duke actually energized any capacity at NC-1, and on what revised schedule?

The company disclosed the purchase of the NC-1 property in Greensboro, North Carolina, cited as claims #4709 and #4643, with 24 MW energization in Q1 2026 (claims #4713, #4650, #4652) and revenue start in May 2026 (claims #4714, #4690, #4651). What has not been said is the site's street address or parcel, the serving substation, and which tenant will generate that revenue. The revenue start date itself is given as 2026-05-01 in claims #4714 and #4690 but 2026-05-31 in claim #4651. Which date is correct, and what is the current expected date now that every NC-1 milestone has slipped?

The company disclosed guidance of 16 MW of total capacity by end of 2025, cited as claims #4715, #4692 and #4649, which appears to equal MTL-1 (4 MW, claim #4702), MTL-2 (5 MW, claim #4704) and MTL-3 (7 MW, claim #4706). That guidance and the MTL-2 and MTL-3 completion dates (claims #4705, #4707, #4648) are now delayed, yet the 76 MW by Q4 2026 target (claims #4717, #4693, #4654) remains outstanding without restatement. MTL-3 completion is also stated as 2026-01-31 in claim #4707 but 2025-12-31 in claim #4648. How much capacity was actually online at year-end 2025, and is the 76 MW target still a live commitment?

The company disclosed a 5 MW Cerebras colocation agreement at MTL-3, cited as claims #4708, #4684 and #4642. What has not been said is the contract term, pricing per kW, ramp schedule, and whether the 5 MW is IT load within MTL-3's 7 MW. Given the MTL-3 delay, has the Cerebras revenue start shifted, and are there delivery penalties or termination rights?

The company disclosed that Bit Digital would own 80 percent after the IPO (claim #4719), described a Bit Digital contribution of assets (claim #4694), and listed an IPO of 7,812,500 shares plus a 1,171,875-share over-allotment (claims #4655, #4656). Separately it named Sam Tabar as CEO (claims #4701, #4696, #4657) and disclosed the Enovum acquisition (claims #4703, #4678, #4638). What has not been said is the offering price and net proceeds, whether the over-allotment was exercised, how proceeds are allocated to NC-1 capex, the Enovum purchase price, and the terms of any intercompany services, cost-sharing or shared-executive arrangements with the controlling parent. How is executive compensation split between WhiteFiber and Bit Digital, and which related-party agreements govern the contributed assets?

The company disclosed on 2025-07-10 a three-year service agreement with a cloud customer, cited as claim #4718 and repeated on 2025-07-22 as its 'first cloud customer' (claim #4695). What has not been said is who that customer is, what the contract is worth in total or annual revenue, and whether it has termination-for-convenience or minimum-commitment terms. Without these, how should a reader gauge revenue concentration?

The company disclosed on 2025-07-10 a tiered Duke Energy power arrangement of 24 MW by 2025-09-01, 40 MW by 2026-04-01 and 99 MW by 2029-05-16, cited as claims #4710, #4711 and #4712 and restated in claims #4686, #4687, #4688, #4644 and #4646. The first two tranches are now marked delayed. What has not been said is whether these figures are firm interconnection capacity, interruptible service, or a non-binding allocation. Nor has the company said whether the slippage came from Duke's side (substation or transmission upgrades) or its own, or what revised energization dates Duke has committed to in writing.

The company disclosed the NC-1 property acquisition in Greensboro, North Carolina, cited as claims #4709 and #4643. It also disclosed a 24 MW initial phase by 2026-03-31 and 40 MW by 2026-06-30 (claims #4713, #4716, #4652, #4653), with revenue starting in May 2026. That revenue date was given as 2026-05-01 in claims #4714 and #4690 but 2026-05-31 in claim #4651. All of these milestones are now delayed. What has not been said is the site's street address or parcel, the substation it interconnects to, and which tenant, if any, has contracted for NC-1 capacity. Why did the revenue-start date shift within eighteen days of disclosure, and what is the current date?

The company guided on 2025-07-10 to 16 MW of total capacity by end-2025, cited as claims #4715, #4692 and #4649. That figure matches MTL-1 at 4 MW (claim #4702), MTL-2 at 5 MW (claim #4704) and MTL-3 at 7 MW (claim #4706). Yet MTL-2 and MTL-3 completion (claims #4705, #4707, #4648) are delayed, while the 76 MW by Q4 2026 target (claims #4717, #4693, #4654) still stands as stated. What has not been said is whether the MW figures are critical IT load or gross utility capacity. The company also hasn't said how much capacity was actually energized and billing at 2025 year-end, or how the 76 MW target survives when every intermediate milestone has slipped.

The company disclosed a 5 MW Cerebras colocation agreement at MTL-3, cited as claims #4708, #4684 and #4642. MTL-3 itself is delayed past its 2025-12-31 and 2026-01-31 completion dates (claims #4648, #4707). What has not been said is the contract term and pricing, and whether delivery delays trigger credits, penalties or a Cerebras right to terminate. Has Cerebras begun paying for any capacity?

The company disclosed an IPO of 7,812,500 shares plus a 1,171,875-share over-allotment option (claims #4655, #4656). It also disclosed that Bit Digital would retain 80 percent ownership post-IPO (claim #4719) after contributing the business (claim #4694), and named Sam Tabar as CEO (claim #4701). What has not been said is whether the over-allotment was exercised, how net proceeds are allocated across NC-1 and the Montreal sites, and what intercompany services, GPU leasing or cost-sharing agreements exist with Bit Digital. A reader would also want to know whether the CEO holds any concurrent role or compensation at the controlling shareholder.

The company disclosed a three-year cloud services agreement with its first cloud customer on 2025-07-10 and 2025-07-22, cited as claims #4718 and #4695; what has not been said is who that customer is, what the total contract value or annual revenue run-rate is, and what share of revenue that single customer represents. Given that MTL-1 is described as having customers (claim #4639), how concentrated is the revenue base?

The company disclosed a 76 MW total HPC capacity target by Q4 2026 on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4717, #4693 and #4654, and that target is still on record as stated, not revised. What has not been said is how it reconciles with the parts that make it up. The Montreal sites add up to 16 MW (MTL-1 at 4 MW, MTL-2 at 5 MW, MTL-3 at 7 MW; claims #4702, #4704, #4706), and NC-1 is planned at 40 MW (claims #4716, #4653). That leaves about 20 MW with no named site. Which facility supplies it? And has the 76 MW figure been reaffirmed, now that the 16 MW year-end 2025 target (claims #4715, #4692, #4649) and the 40 MW NC-1 Q2 2026 target are both marked delayed?

The company disclosed a phased Duke Energy capacity commitment for NC-1 on 2025-07-28: 24 MW by 2025-09-01, 40 MW by 2026-04-01 and 99 MW by 2029-05-16, cited as claims #4644, #4645 and #4646 (repeated in #4710–4712 and #4686–4688). What has not been said is whether these are utility load figures or critical IT megawatts, which differ materially. The company also has not said whether the 24 MW and 40 MW tranches, both now marked delayed, were actually energized, and if so on what dates. Has the interconnection agreement or substation scope been filed, and does the Greensboro site (claims #4709, #4643) have a disclosed street address and parcel?

The company disclosed completion dates for MTL-2 and MTL-3 that changed within weeks. On 2025-07-28 they were given as 2025-12-31 (claims #4647 and #4648); on 2025-07-10 they were given as 2026-01-31 (claims #4705 and #4707). NC-1's revenue start likewise appears as both 2026-05-01 (claims #4714 and #4690) and 2026-05-31 (claim #4651). Every one of these dates is now marked delayed, and what has not been said is the current energization date and revenue start date for each facility, along with the reason for the slippage.

The company disclosed a 5 MW colocation agreement with Cerebras at MTL-3 on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4708, #4684 and #4642. What has not been said is the contract term, pricing per kW, or whether Cerebras has delivery-date remedies or termination rights tied to MTL-3's delayed completion (claim #4707). The Enovum acquisition (claims #4703 and #4678) was likewise announced without a purchase price or valuation in this record.

The company disclosed that Bit Digital would own 80 percent of the company after the IPO, cited as claim #4719, and that Bit Digital contributed assets, cited as claim #4694. The IPO sold 7,812,500 shares with a 1,171,875-share over-allotment option (claims #4655 and #4656). What has not been said is the terms of the intercompany agreements with Bit Digital, including shared services, transferred assets and any fees paid to it. The company also has not said whether CEO Sam Tabar (claim #4701) holds roles or equity at Bit Digital that create overlapping compensation, and whether the over-allotment option was exercised and with what net proceeds.

The company disclosed a three-year service agreement with its first cloud customer on 2025-07-10 and 2025-07-22, cited as claims #4718 and #4695. What has not been said is who the customer is, what the total contract value is, and how many GPUs or megawatts the agreement covers. Without those figures, a reader cannot tell how concentrated revenue is or how much of the contracted base depends on a single counterparty.

The company disclosed capacity guidance of 16 MW by end-2025 and 76 MW by Q4 2026 on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4715, #4692, #4649, #4717, #4693 and #4654. What has not been said is whether these figures are critical IT load or gross utility draw, and which sites make up the 76 MW. The disclosed sites add up to 16 MW in Montreal (MTL-1, MTL-2 and MTL-3 at 4, 5 and 7 MW) plus 40 MW at NC-1, or 56 MW. Where do the remaining 20 MW come from, and has that site been identified at all?

The company disclosed a phased Duke Energy power commitment on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4710–#4712, #4686–#4688 and #4644–#4646. The phases are 24 MW by 2025-09-01, 40 MW by 2026-04-01 and 99 MW by 2029-05-16. The first two dates have passed and both are flagged delayed. What has not been said is how many megawatts Duke has actually energized at NC-1 today, and whether the 99 MW tranche is a firm service agreement or depends on grid upgrades. Also missing is who bears the interconnection and substation costs.

The company disclosed the purchase of the NC-1 property in Greensboro, North Carolina on 2025-07-10 and 2025-07-28, cited as claims #4709 and #4643, with revenue start promised by May 2026 in claims #4714, #4690 and #4651. What has not been said is the site's physical address, acreage or purchase price, or the substation it connects to. Nor has the company given a revised revenue-start date now that the original one has slipped. How much NC-1 construction capex remains, and is it funded?

The company disclosed a 5 MW Cerebras colocation agreement at the 7 MW MTL-3 Saint-Jérôme facility on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4708, #4684 and #4642. MTL-3's January 2026 completion and revenue start is flagged delayed in claim #4707. What has not been said is the contract's term, its pricing per kW, and whether it carries delivery-date penalties or termination rights that the MTL-3 delay may have triggered. The company also has not said what will fill the remaining 2 MW.

The company disclosed that Bit Digital would own about 80 percent after the IPO, that Bit Digital contributed assets to it, and that 7,812,500 shares were offered with a 1,171,875-share over-allotment, on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4719, #4694, #4655 and #4656. Sam Tabar is named CEO in claims #4701 and #4696. What has not been said is the IPO price, whether the over-allotment was exercised, and how net proceeds were split between NC-1 and the Montreal build-outs. Also absent are the terms of the intercompany services, lease or GPU-supply agreements with Bit Digital. Does the CEO hold any role or compensation at the parent, and how is that time and pay allocated?

The company disclosed on 2025-07-10 and again on 2025-07-22 a three-year service agreement with its first cloud customer (claims #4718 and #4695); what has not been said is who that customer is, what contract value or annual revenue it carries, and what share of total revenue it represents, which matters because a single unnamed counterparty may be carrying the cloud segment.

WhiteFiber disclosed a Duke Energy capacity arrangement stepping from 24 MW to 40 MW to 99 MW (claims #4710, #4711, #4712, repeated as #4644, #4645, #4646). The 24 MW tranche due 2025-09-01 and the 40 MW tranche due 2026-04-01 are both now marked delayed. What has not been said is whether these are firm, signed interconnection or electric service agreements or just capacity reservations, what caused the slippage, and whether the 2029-05-16 date for 99 MW still holds.

The company guided on 2025-07-10 and 2025-07-28 to NC-1 energizing 24 MW by Q1 2026 and generating revenue by May 2026 (claims #4713, #4714, #4650, #4651, #4652), and all of these are marked delayed. What has not been said is the revised energization date, the cause of the delay (power, equipment, permitting, or construction), and whether any tenant signed for NC-1 capacity has termination or penalty rights tied to late delivery. No NC-1 tenant has been named at all.

WhiteFiber disclosed the Cerebras colocation agreement at MTL-3 for 5 MW (claims #4708, #4684, #4642) and a January 2026 completion and revenue start for MTL-3 (claim #4707), which is now delayed. What has not been said is the contract term, pricing per MW, whether Cerebras has walk-away or credit rights because of the delay, and whether billing has started on any of the 5 MW.

The company guided to 16 MW total capacity by end of 2025 (claims #4715, #4692, #4649), which is now marked delayed, while keeping the 76 MW by Q4 2026 target (claims #4717, #4693, #4654) unchanged in the record. What has not been said is how much capacity was actually online at 2025 year-end, whether the figures are gross critical IT load or leasable capacity, and how the 76 MW target survives when both the MTL and NC-1 milestones feeding it have slipped. The company has also given no capex budget or financing plan for that buildout beyond the IPO proceeds (claims #4655, #4656).

WhiteFiber disclosed that Bit Digital would hold roughly 80 percent after the IPO (claim #4719) and that Bit Digital contributed assets to the company (claim #4694), with Sam Tabar as CEO (claims #4701, #4696). What has not been said is the terms of any services, shared-employee, or intercompany agreements with Bit Digital, whether executives are paid by both entities, and how conflicts are handled between the controlling shareholder and minority IPO holders.

The company disclosed a three-year service agreement with its first cloud customer on 2025-07-10 and again on 2025-07-22, cited as claims #4718 and #4695. What has not been said is who that customer is, what the total contract value or annual revenue run-rate is, and how many GPUs or megawatts the agreement covers. Is this customer a meaningful share of revenue, and can it terminate early?

The company disclosed capacity guidance of 16 MW by end of 2025 and 76 MW by end of 2026, cited as claims #4715, #4692, #4649, #4717, #4693 and #4654. The Montreal sites add up to exactly 16 MW (4 MW at MTL-1, 5 MW at MTL-2, 7 MW at MTL-3; claims #4702, #4704, #4706), and NC-1 is guided to 40 MW (claims #4716, #4653). What has not been said is where the remaining 20 MW of the 76 MW figure comes from. It is also unstated whether these figures measure critical IT load or gross utility capacity. With MTL-2 and MTL-3 completion now tracked as delayed past 2026-01-31 (claims #4705, #4707), did the company actually reach 16 MW, and if not, what is the revised number?

The company disclosed a phased Duke Energy power commitment of 24 MW by 2025-09-01, 40 MW by 2026-04-01 and 99 MW by 2029-05-16, cited as claims #4710, #4711, #4712 (repeated in #4686–#4688 and #4644–#4646). What has not been said is whether these tranches are cumulative or incremental. The company has also not said whether the first 24 MW has actually been energized, since both of the first two dates have now passed and are tracked as delayed. Nor has it said what substation, interconnection agreement or capital contribution to Duke underpins the 99 MW figure. Why does 2026 guidance stop at 40 MW for NC-1 when the power path extends to 99 MW?

The company disclosed the purchase of the NC-1 property in Greensboro, North Carolina, with 24 MW completion by 2026-03-31 and revenue starting in May 2026, cited as claims #4709, #4643, #4713, #4714, #4650, #4651 and #4690. What has not been said is the site's street address or acreage, and which tenant or customer is expected to generate that revenue. Is NC-1 being built speculatively? Given that both the completion and revenue dates are now tracked as delayed, what new dates has the company given, and what is the total capex budget for the build?

The company disclosed a 5 MW colocation agreement with Cerebras at MTL-3, cited as claims #4708, #4684 and #4642. What has not been said is the contract term, the pricing per kW, or whether the MTL-3 completion delay (claim #4707) pushed back the Cerebras revenue start or triggered any penalty or renegotiation rights. Who is contracted for the remaining 2 MW of the 7 MW site?

The company disclosed that Bit Digital would own about 80 percent after the IPO, following an asset contribution, with an offering of 7,812,500 shares plus a 1,171,875-share over-allotment, cited as claims #4719, #4694, #4655 and #4656. It named Sam Tabar as CEO in claims #4701, #4696 and #4657. What has not been said is whether the over-allotment was exercised, and what shared-services, intercompany or tax-sharing agreements exist with the controlling parent. It is also unstated how the CEO's time and compensation are split if he also serves in a leadership role at Bit Digital. More broadly, every claim on record dates from July 2025, so the company has not updated any of these delayed commitments in the tracked record since then.

The company disclosed a three-year service agreement with its first cloud customer on 2025-07-10 and again on 2025-07-22, cited as claims #4718 and #4695; what has not been said is who that customer is, the total contract value, the GPU count or megawatts it covers, and how much of the company's revenue it represents.

The company disclosed Duke Energy commitments of 24 MW by 2025-09-01 and 40 MW by 2026-04-01 at NC-1, cited as claims #4710, #4711, #4686, #4687, #4644 and #4645, all now tracked as delayed, along with a 24 MW NC-1 energization targeted for Q1 2026 that has also slipped, cited as claims #4713, #4652 and #4650. What has not been said is the revised energization date, whether the Duke commitments are firm service agreements or load-study reservations, which substation serves the site, and the Greensboro street address of the property acquired under claims #4709 and #4685.

The company guided to 16 MW by end-2025 and 40 MW at NC-1 by Q2 2026, cited as claims #4715, #4692, #4649, #4716 and #4653, all delayed, yet it still shows 76 MW by Q4 2026 as live guidance, cited as claims #4717, #4693 and #4654. What has not been said is whether the 76 MW target has been formally reaffirmed or cut, and whether the figures are critical IT load or gross utility power. It has also not identified the site for the roughly 20 MW left after adding MTL-1, MTL-2, MTL-3 and NC-1's 40 MW (4 + 5 + 7 + 40 = 56 MW).

The company gave MTL-2 and MTL-3 completion dates of 2025-12-31 on 2025-07-28, cited as claims #4647 and #4648, but 2026-01-31 in claims #4705 and #4707, and both are now delayed. What has not been said is which date was the operative target, the current completion date, and whether the 5 MW Cerebras colocation agreement at MTL-3, cited as claims #4708, #4684 and #4642, has started billing, along with its term, pricing and any penalty or termination rights tied to late delivery.

The company disclosed that Bit Digital would own 80 percent after the IPO and contributed assets to WhiteFiber, cited as claims #4719 and #4694, and named Sam Tabar as CEO, cited as claim #4701. What has not been said is the terms of any shared-services, intercompany or tax-sharing agreements with the controlling parent, and whether Tabar holds a concurrent role at Bit Digital. If he does, it is also unclear how his compensation and time are split between the two companies.

The company disclosed an IPO of 7,812,500 shares plus a 1,171,875-share over-allotment option, cited as claims #4655 and #4656, and the NC-1 and Enovum acquisitions, cited as claims #4709, #4703 and #4678. What has not been said is whether the over-allotment was exercised or what net proceeds were raised. It also has not disclosed the purchase consideration for NC-1 and Enovum, or how much capital is still needed and from which sources to fund the NC-1 build-out to 99 MW by 2029, cited as claims #4712, #4688 and #4646.

The company disclosed on 2025-07-10 and again on 2025-07-22 a three-year service agreement with its first cloud customer, cited as claims #4718 and #4695; what has not been said is who that customer is, what the total contract value or annual revenue run-rate is, and how many GPUs or megawatts the contract occupies. For a business with only a few sites, the concentration risk of one unnamed anchor tenant is the first question on any earnings call.

The company disclosed a 5 MW Cerebras colocation agreement placed at MTL-3 (claims #4708, #4684, #4642), while MTL-3's completion and revenue start has slipped past both the 2025-12-31 date (claim #4648) and the 2026-01-31 date (claim #4707); what has not been said is whether Cerebras' service start has moved with it, whether the contract carries delivery penalties or termination rights for late handover, and what pricing per MW or term length the agreement carries. The completion dates also conflict for MTL-2 and MTL-3: filings on the same day cite 2025-12-31 (claims #4647, #4648) and 2026-01-31 (claims #4705, #4707), and neither set has been reconciled.

The company disclosed a phased Duke Energy capacity commitment of 24 MW by 2025-09-01, 40 MW by 2026-04-01 and 99 MW by 2029-05-16 (claims #4710, #4711, #4712, repeated in #4644, #4645, #4646); what has not been said is how many megawatts Duke has actually energized at the Greensboro NC-1 site, the substation or interconnection point involved, and whether the 99 MW phase is a firm contractual commitment or subject to further utility studies. With the first two phases and the NC-1 24 MW completion (claims #4713, #4650, #4652) all marked delayed, investors lack a revised energization date. The NC-1 revenue start is also cited as both 2026-05-01 (claims #4714, #4690) and 2026-05-31 (claim #4651), and has not been updated.

The company disclosed capacity guidance of 16 MW by end of 2025 (claims #4715, #4692, #4649) and 76 MW by Q4 2026 (claims #4717, #4693, #4654); what has not been said is how the 76 MW figure is built. MTL-1, MTL-2 and MTL-3 sum to 16 MW (claims #4702, #4704, #4706) and NC-1 is described as 40 MW (claims #4691, #4653), which leaves roughly 20 MW unassigned to any named site. The disclosures also do not specify whether these figures are critical IT load or gross utility power, and the 76 MW target remains stated even though every intermediate milestone behind it has slipped.

The company disclosed that Bit Digital will hold about 80 percent after the IPO (claim #4719), contributed assets to WhiteFiber (claim #4694), and that Sam Tabar is CEO (claims #4701, #4657); what has not been said is the full set of intercompany agreements with the controlling parent, including services, shared staff and asset transfer pricing. Readers also lack detail on whether the CEO holds a concurrent executive role at Bit Digital, and how compensation and time are split between the two companies.

The company disclosed an IPO of 7,812,500 shares (claim #4655) with a 1,171,875-share over-allotment option (claim #4656), plus the Enovum acquisition (claims #4703, #4678) and the NC-1 Greensboro property purchase (claims #4709, #4643); what has not been said is whether the over-allotment was exercised, how net proceeds were allocated across the delayed build-outs, what was paid for Enovum and NC-1, and whether the capital raised is enough to reach 76 MW without further financing.

The company disclosed a three-year cloud service agreement on 2025-07-10, cited as claim #4718 and repeated as the "first cloud customer" in claim #4695 on 2025-07-22; what has not been said is who the customer is, what the total contract value or annualized revenue is, and how much of the company's GPU fleet or MW footprint the contract consumes. A related question applies to MTL-1's 4 MW (claims #4702, #4679, #4639): which customers occupy it, and how concentrated is that revenue?

The company disclosed a 5 MW Cerebras colocation agreement at MTL-3 on 2025-07-10 and again on 2025-07-22 and 2025-07-28, cited as claims #4708, #4684 and #4642; what has not been said is the term, the pricing per kW or total contract value, and whether the contract has penalty or termination rights now that MTL-3's completion and revenue start (claims #4707, #4648) has slipped past both the 2025-12-31 and 2026-01-31 dates given for it.

The company disclosed a tiered Duke Energy commitment of 24 MW, 40 MW and 99 MW, cited as claims #4710, #4711 and #4712 (repeated as #4686–#4688 and #4644–#4646); what has not been said is whether these are executed interconnection or electric service agreements or non-binding capacity reservations, whether the figures are utility-side or critical IT MW, and what energization date Duke now expects now that both the 2025-09-01 and 2026-04-01 milestones are marked delayed. The NC-1 24 MW completion (claims #4713, #4652), 40 MW by Q2 2026 (claim #4716) and revenue start (claims #4714, #4690, #4651) have all passed without a restated date. Has any MW at NC-1 actually been energized, and is there a signed tenant for it?

The company disclosed 2026 guidance of 76 MW by year-end on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4717, #4693 and #4654, and it remains "stated" while every interim milestone beneath it is delayed. What has not been said is whether that guidance has been formally reaffirmed or withdrawn, and how 76 MW is built up. MTL-1, MTL-2 and MTL-3 (4 + 5 + 7 MW, the basis of the missed 16 MW 2025 target in claims #4715, #4692 and #4649) plus 40 MW at NC-1 sum to only 56 MW. Which site supplies the remaining 20 MW, and is it gross or IT load?

The company disclosed an IPO of 7,812,500 shares and an 1,171,875-share over-allotment option on 2025-07-28, cited as claims #4655 and #4656, alongside the Enovum, MTL-2, MTL-3 and NC-1 Greensboro acquisitions (claims #4703, #4680, #4682, #4709). What has not been said in these claims is the per-share price and net proceeds, whether the over-allotment was exercised, the purchase price of each property, and how much remaining capex is needed to reach 76 MW versus cash on hand. Is additional debt or equity required?

The company disclosed that Bit Digital would own roughly 80 percent post-IPO and contributed the business, cited as claims #4719 and #4694, with Sam Tabar named CEO in claim #4701. What has not been said is the terms of any shared-services, transition, or intercompany agreements with the parent, how executive time and compensation are split if officers serve both entities, and what governance protections exist for minority holders under a controlled-company structure.

The company disclosed a three-year service agreement with its first cloud customer on 2025-07-10 and 2025-07-22, cited as claims #4718 and #4695; what has not been said is who that customer is, the total contract value, the number and generation of GPUs committed, and whether the agreement is take-or-pay or subject to early termination. Without these, investors cannot tell how much of the cloud segment's revenue rests on this one unnamed counterparty.

The company disclosed three Duke Energy capacity commitments of 24 MW, 40 MW and 99 MW on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4710, #4711, #4712, #4644, #4645 and #4646; what has not been said is whether these are executed interconnection or electric service agreements or only load-study reservations, whether the capacity is firm or interruptible, and what caused the 2025-09-01 and 2026-04-01 milestones to slip. The 24 MW tranche is now more than a year past its stated date with no revised date on record.

The company disclosed NC-1 targets of 24 MW energized by 2026-03-31, 40 MW by 2026-06-30, and revenue starting in May 2026, cited as claims #4713, #4652, #4716, #4653, #4714 and #4651; what has not been said is a revised schedule now that every one of these dates is marked delayed. The company also has not named an anchor tenant for NC-1, and it has not explained why the filings give the revenue start as both 2026-05-01 (#4690) and 2026-05-31 (#4651). The Greensboro acquisition, cited as claims #4709 and #4643, was disclosed without a purchase price, street address or building square footage.

The company disclosed a 76 MW total capacity target for end-2026, cited as claims #4717, #4693 and #4654; what has not been said is how that total is built. MTL-1, MTL-2 and MTL-3 add up to 16 MW (#4702, #4704, #4706), and NC-1 adds 40 MW (#4716), for 56 MW. The remaining roughly 20 MW has not been assigned to any named site. The company also has not said whether these figures are critical IT load or gross utility capacity, which matters because the 16 MW end-2025 target (#4715, #4692, #4649) is itself marked delayed and no current energized MW figure is on record.

The company disclosed a 5 MW Cerebras colocation agreement at MTL-3 on 2025-07-10, cited as claims #4708, #4684 and #4642; what has not been said is the contract term, the per-kW pricing, or whether Cerebras revenue has begun. MTL-3's own completion and revenue start of 2026-01-31, cited as claim #4707, is marked delayed. Cerebras would also take about 70% of MTL-3's 7 MW, and it is unclear whether the remaining 2 MW has been leased.

The company disclosed that Bit Digital would own 80% after the IPO and that Bit Digital contributed the business, cited as claims #4719 and #4694, alongside an offering of 7,812,500 shares plus a 1,171,875-share over-allotment, cited as claims #4655 and #4656; what has not been said is the offering price, the net proceeds, whether the over-allotment was exercised, and how much of the proceeds is earmarked for NC-1 capex. The company also has not detailed the related-party arrangements with its controlling shareholder: shared-services fees, intercompany GPU or hosting agreements, and whether CEO Sam Tabar (#4701) holds concurrent roles or compensation at Bit Digital.

The company disclosed a three-year service agreement with its first cloud customer on 2025-07-10 and again on 2025-07-22, cited as claims #4718 and #4695; what has not been said is who that customer is, what the contract is worth, and how many MW or GPUs it covers. Can WhiteFiber name the counterparty, or at least give total contract value and the share of revenue it represents, given that the agreement is presented as a milestone?

The company disclosed a phased Duke Energy capacity commitment on 2025-07-10, 2025-07-22 and 2025-07-28: 24 MW by 2025-09-01, 40 MW by 2026-04-01 and 99 MW by 2029-05-16, cited as claims #4710, #4711, #4712, #4686, #4687, #4688, #4644, #4645 and #4646. What has not been said is whether the 24 MW and 40 MW tranches were actually energized, now that both dates have passed and both are tracked as delayed. The company has also not said whether the delay sits with Duke or with WhiteFiber, whether it has signed an interconnection agreement or only a load letter, and whether all 99 MW is dedicated to the NC-1 Greensboro site acquired under claims #4709 and #4643.

The company reaffirmed a 76 MW total-capacity target for Q4 2026 on three separate dates, cited as claims #4717, #4693 and #4654. It did so while the 16 MW end-2025 target (claims #4715, #4692, #4649), the NC-1 initial 24 MW by Q1 2026 (claims #4713, #4652) and the NC-1 40 MW by Q2 2026 (claims #4716, #4653) have all slipped. What has not been said is how much capacity is actually energized and billing today. The company also has not said whether these MW figures are critical IT load or gross utility capacity, and whether the 76 MW guidance has been formally withdrawn or re-dated.

The company disclosed a 5 MW Cerebras colocation agreement at MTL-3 on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4708, #4684 and #4642. It also disclosed a 7 MW MTL-3 Saint-Jérôme facility (claim #4706) whose completion date is stated inconsistently as 2025-12-31 (claim #4648) and 2026-01-31 (claim #4707) and is now delayed. What has not been said is the Cerebras contract term and pricing, whether Cerebras has any termination or penalty rights tied to the MTL-3 slippage, and who is contracted for the remaining 2 MW. The same question applies to MTL-2, disclosed as 5 MW (claim #4704) with delayed completion (claim #4705): is any customer contracted there at all?

The company disclosed that Bit Digital would own 80 percent of WhiteFiber post-IPO on 2025-07-10, cited as claim #4719, and that Bit Digital contributed assets to the company (claim #4694), with Sam Tabar named CEO (claims #4701, #4696, #4657). What has not been said is what intercompany agreements govern shared services, GPU leasing or asset transfers between parent and subsidiary, and on what pricing terms. The company has also not disclosed whether Tabar holds any concurrent role or equity at Bit Digital, or how his WhiteFiber compensation is structured and approved when a controlling shareholder sits across the table.

The company disclosed an IPO of 7,812,500 shares with a 1,171,875-share over-allotment option on 2025-07-28, cited as claims #4655 and #4656. What has not been said is the final offering price, whether the over-allotment was exercised, and how much of the net proceeds went to NC-1 build-out versus the MTL sites. With several facility timelines now delayed, is the funding for the remaining NC-1 phases fully committed, or will additional debt or equity be needed?

The company disclosed a three-year service agreement with its first cloud customer on 2025-07-10 and 2025-07-22, cited as claims #4718 and #4695; what has not been said is who that customer is, what the contract is worth, how many MW or GPUs it covers, and which facility serves it. Is this customer the same as, or separate from, the 5 MW Cerebras colocation deal (claims #4708 and #4684), and at which site does Cerebras actually sit?

The company disclosed a three-phase Duke Energy capacity commitment on 2025-07-10 and 2025-07-28: 24 MW by 2025-09-01, 40 MW by 2026-04-01, and 99 MW by 2029-05-16, cited as claims #4710, #4711, #4712, #4644, #4645 and #4646. What has not been said is why the first two phases are now past due, what the revised energization dates are, and whether these figures are utility service capacity or billable critical IT load. Has a signed interconnection or electric service agreement been executed, and does it cover substation or transmission upgrades at the Greensboro site?

The company disclosed guidance of 16 MW by end-2025 and 76 MW by Q4 2026 on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4715, #4692, #4649, #4717, #4693 and #4654. The 16 MW figure matches MTL-1, MTL-2 and MTL-3 at 4, 5 and 7 MW (claims #4702, #4704, #4706), and adding NC-1's 40 MW (claim #4691) reaches only 56 MW. What has not been said is where the remaining roughly 20 MW comes from, and whether the 76 MW target still stands now that the 16 MW interim target and every NC-1 milestone have slipped.

The company disclosed an NC-1 revenue start of 2026-05-01 in claims #4714 and #4690, but 2026-05-31 in claim #4651 on 2025-07-28, alongside 24 MW energization by 2026-03-31 (claims #4713 and #4652). All of these are now delayed. What has not been said is which date was the operative target, what the new date is, and whether NC-1 has a contracted tenant at all. The Greensboro property purchase (claims #4709 and #4643) was announced without a street address, parcel size, purchase price or financing source.

The company disclosed MTL-2 and MTL-3 completion by 2025-12-31 in claims #4647 and #4648 on 2025-07-28, but by 2026-01-31 in claims #4705 and #4707. Both are now marked delayed. What has not been said is the current commissioning status of each Montreal site, the reason for the slip, and whether the 7 MW Saint-Jérôme site (claim #4706) has firm Hydro-Québec allocation for its full capacity.

The company disclosed an IPO of 7,812,500 shares plus a 1,171,875-share over-allotment option on 2025-07-28 (claims #4655 and #4656), and Bit Digital's 80% post-IPO ownership and asset contribution (claims #4719 and #4694). What has not been said is the net proceeds, whether the over-allotment was exercised, and how proceeds were allocated between NC-1 and the Montreal build-outs. On the related-party side, what intercompany services, leases or cost-sharing agreements exist with Bit Digital, and how is CEO Sam Tabar (claim #4701) compensated and his time allocated if he holds any concurrent role at the controlling parent?

The company disclosed a three-year cloud services agreement with an unnamed customer on 2025-07-10, cited as claims #4718 and #4695 ("first cloud customer"); what has not been said is who the counterparty is, the contract's total value or annualized revenue, how many MW or GPUs it consumes, whether it is take-or-pay or cancelable, and whether that customer is a related party of Bit Digital. Given that the same filings quantify every megawatt but not a dollar of contracted revenue, the first-year-analyst question is straightforward: what is the remaining performance obligation under this agreement as of the most recent quarter-end, and what percentage of total revenue did this single customer represent?

The company disclosed a 5 MW Cerebras colocation agreement sited at MTL-3 on 2025-07-10 and again on 2025-07-22 and 2025-07-28, cited as claims #4708, #4684 and #4642, alongside MTL-3's total capacity of 7 MW (claims #4706, #4641); what has not been said is the contract term, price per kW-month, escalators, or what happens to that commitment now that MTL-3's completion and revenue start — promised by 2026-01-31 in claim #4707 — carries a delayed status. If a single tenant occupies roughly 70 percent of a facility that has slipped, has Cerebras retained termination or rent-abatement rights, and has any portion of the 5 MW been energized or billed to date?

The company disclosed a tiered Duke Energy arrangement of 24 MW by 2025-09-01, 40 MW by 2026-04-01 and 99 MW by 2029-05-16, cited as claims #4710/#4686/#4644, #4711/#4687/#4645 and #4712/#4688/#4646, with the first two tranches now marked delayed; what has not been said is the legal character of these commitments — executed electric service agreement, conditional interconnection letter, or queue position — and which substation and feeder serve NC-1. The Phase 1 date of 2025-09-01 has passed; what is the revised energization date, what deposits or facilities charges has WhiteFiber already paid to Duke, and are the 40 MW and 99 MW tranches contractually independent of Phase 1 or do they fall away if Phase 1 milestones are missed?

The company disclosed the NC-1 property purchase in Greensboro, North Carolina, cited as claims #4709, #4643 and #4685 (the latter tied to a 2025-05-20 date now marked delayed); what has not been said is the street address, parcel acreage, purchase price, existing building shell versus greenfield status, water source for cooling, or the local permitting timeline. A reader can verify none of the 24 MW / 40 MW buildout at claims #4713, #4650 and #4653 without knowing whether the site has a building permit issued and a general contractor under a fixed-price contract — so which is it, and what is the budgeted capital cost per MW?

The company disclosed that Bit Digital would retain 80.0 percent ownership post-IPO on 2025-07-10, cited as claim #4719, together with a Bit Digital contribution of assets (claim #4694) and the appointment of Sam Tabar as Chief Executive Officer (claims #4701, #4696, #4657), who is also Bit Digital's chief executive; what has not been said is any related-party framework — the master services or shared-services agreement between parent and subsidiary, how Tabar's compensation is allocated between the two public companies, whether he received WhiteFiber equity awards on top of Bit Digital pay, and what the independent-director composition and controlled-company exemptions look like. What were the terms and valuation of the Enovum and data center assets contributed by Bit Digital (claims #4678, #4703, #4638), and was that valuation reviewed by a disinterested committee?

The company disclosed an IPO of 7,812,500 shares plus a 1,171,875-share over-allotment on 2025-07-28, cited as claims #4655 and #4656; what has not been said is the offer price, net proceeds, or the use-of-proceeds allocation across MTL-2, MTL-3 and NC-1 — and, critically, the total capital required to reach the 76 MW target reaffirmed for 2026-12-31 in claims #4717, #4693 and #4654. Since the 16 MW end-of-2025 milestone (claims #4715, #4692, #4649) and essentially every component milestone feeding the 76 MW figure are now delayed while the 76 MW target itself remains merely stated, what is the current MW-by-quarter bridge from today's operational 4 MW at MTL-1 (claims #4702, #4679, #4639) to 76 MW, and how much additional debt or equity beyond IPO proceeds is required to fund it?

The company has repeatedly announced a "first cloud customer" under a three-year service agreement — claims #4718 and #4695, both stated — while naming Cerebras openly as the MTL-3 colocation counterparty in claims #4708, #4684 and #4642. What has not been said is who the cloud customer is, what annual contract value or total contract value the three-year term carries, whether it is the same entity as Cerebras or an affiliate, and whether there are take-or-pay minimums or termination-for-convenience rights. The asymmetry is the tell: if one counterparty can be named, a reader should ask why the revenue-generating one cannot.

The Duke Energy relationship is disclosed as three separate capacity tranches — 24 MW by 2025-09-01 (claims #4710, #4686, #4644), 40 MW by 2026-04-01 (claims #4711, #4687, #4645), and 99 MW by 2029-05-16 (claims #4712, #4688, #4646) — with the first two now marked delayed as of 2026-09-03. What has not been said is what contractual form these "commitments" take: an executed electric service agreement, a load letter, a non-binding interconnection queue position, or a substation study. The question for the next call is which of the three tranches is backed by a signed ESA with a stated in-service date and what deposits or construction contributions the company has posted against each.

NC-1 is described as a Greensboro, North Carolina property acquisition (claims #4709, #4685, #4643) with 24 MW energization in Q1 2026 (claims #4713, #4652) scaling to 40 MW by Q2 2026 (claims #4716, #4653), every one of which flipped to delayed on 2026-09-03. What has not been said is the street address or parcel, the name of the delivering substation and feeder, and the specific gating item behind the delay — transformer lead time, Duke interconnection study, or local permitting. Nine months past the original 24 MW date, a reader is entitled to a revised energization date rather than a restated target.

The 2025 exit-capacity guidance of 16 MW (claims #4715, #4692, #4649) is now delayed, yet the 2026 exit target of 76 MW (claims #4717, #4693, #4654) remains in "stated" status and has never been revised. What has not been said is how 76 MW survives the slippage of every intermediate milestone that was supposed to build to it — MTL-2 and MTL-3 revenue start (claims #4705, #4707, #4648), NC-1 24 MW, and NC-1 40 MW. The arithmetic follow-up: reconcile the 76 MW figure site by site, and state whether it means contracted capacity, energized utility capacity, or IT load available to customers, since the MTL-1 4 MW, MTL-2 5 MW and MTL-3 7 MW figures (claims #4702, #4704, #4706) are not labeled on any consistent basis.

The IPO is disclosed as 7,812,500 shares with a 1,171,875-share over-allotment option (claims #4655, #4656), and Bit Digital is disclosed as retaining 80 percent post-IPO (claim #4719) alongside an unquantified "Bit Digital contribution" (claim #4694). What has not been said is the gross and net proceeds, whether the over-allotment was exercised, what assets and liabilities Bit Digital contributed and at what carrying value, and the total capital required to reach 76 MW against cash on hand. A reader should also ask what portion of remaining buildout is funded versus dependent on future debt or equity, given an 80-percent parent that controls the vote.

Related-party and compensation context is thin against what has been disclosed. CEO Sam Tabar is named (claims #4701, #4696, #4657) and Bit Digital's 80-percent stake is disclosed (claim #4719), but what has not been said is whether Tabar and other executives hold concurrent roles or compensation at Bit Digital, what the terms of any shared-services, transition-services or hosting agreement between WhiteFiber and Bit Digital are, and how much of WhiteFiber's reported revenue is attributable to its controlling shareholder rather than third parties. Similarly, the Enovum acquisition (claims #4703, #4678, #4638) has been announced across three dates without a purchase price, earnout, or seller-continuity disclosure.

Finally, MTL-2 and MTL-3 site acquisitions carry promised-by dates of 2024-12-27 and 2025-04-11 (claims #4680, #4682) that are marked delayed, with completion and revenue start pushed from 2025-12-31 to 2026-01-31 (claims #4648, #4705, #4707). What has not been said is whether MTL-2 and MTL-3 are now energized and billing as of today, what the Saint-Jérôme utility interconnection status is with Hydro-Québec, and whether the Cerebras 5 MW at MTL-3 has commenced service or remains contingent on a facility that has slipped twice.

The company disclosed a three-year cloud services agreement with an unnamed "first cloud customer" on 2025-07-10 and again on 2025-07-22, cited as claims #4718 and #4695; what has not been said is who the counterparty is, the contract's total or annualized value, the committed MW or GPU count underlying it, whether it carries a take-or-pay floor, and what termination or repricing rights exist at the eighteen-month mark. A sophisticated reader will note that a named partner (Cerebras, claims #4708 and #4684) got a MW figure and a name while the revenue-generating customer got neither — why the asymmetry?

The Cerebras colocation agreement was disclosed at 5.0 MW on 2025-07-10 and 2025-07-22, cited as claims #4708 and #4684; what has not been said is which facility houses it, given that MTL-2's entire disclosed capacity is also 5.0 MW (claims #4704, #4680, #4640) and MTL-2's completion has since slipped from the stated 2025-12-31 / 2026-01-31 dates into delayed status (claims #4705, #4647). Is the Cerebras 5 MW incremental to MTL-1's operating 4.0 MW, or is it the whole of a facility that is not yet built — and does the colocation contract contain a delivery-date condition that the MTL-2 delay has now tripped?

The Duke Energy relationship was disclosed as three tranches — 24 MW by 2025-09-01, 40 MW by 2026-04-01, and 99 MW by 2029-05-16 — cited as claims #4710/#4686/#4644, #4711/#4687/#4645, and #4712/#4688/#4646. What has not been said is whether these are executed electric service agreements, load letters, or merely interconnection queue positions; what the contracted rate or rate schedule is; which substation and feeder serve NC-1; and, most pointedly, why the first two tranches have both moved to delayed while the 99 MW 2029 tranche remains carried as "stated" with no revised energization date. If the 24 MW tranche missed 2025-09-01 by over a year, on what basis is the 2029 figure still intact?

The NC-1 Greensboro property acquisition was disclosed with a 2025-05-20 date and is now delayed, cited as claims #4709, #4685 and #4643, with initial 24 MW completion promised by 2026-03-31 and revenue start by 2026-05-01 or 2026-05-31 (claims #4713, #4650, #4652, #4714, #4690, #4651) — all now delayed and all now in the past. What has not been said is the street address or parcel identifier, the purchase price, the total capex per MW to energize the first 24 MW, how much of that capex has been incurred to date, and the current expected first-revenue month. Same question for MTL-3 in Saint-Jérôme at 7.0 MW (claims #4706, #4682, #4641): what is the site address, and is the Hydro-Québec block allocation confirmed and in whose name?

Guidance of 16 MW by 2025-12-31 (claims #4715, #4692) is now delayed and the period has closed, and the 40 MW at NC-1 by 2026-06-30 (claims #4716, #4653) is delayed with that date also past; yet the 76 MW by 2026-12-31 target (claims #4717, #4693, #4654) is still carried as "stated." What has not been said is the actual exit capacity for 2025, the actual energized capacity as of Q3 2026, and the bridge that reconciles a missed 16 MW and a missed 40 MW into an intact 76 MW with roughly one quarter remaining. Are the MW figures critical IT load, utility interconnect capacity, or nameplate — and is the definition consistent across the facility claims and the 76 MW aggregate?

Bit Digital's 80% post-IPO ownership was disclosed on 2025-07-10, cited as claim #4719, alongside a Bit Digital contribution of assets (claim #4694) and the Enovum acquisition (claims #4703, #4678, #4638, the last of which is delayed against a 2024-10-01 date). What has not been said is the related-party operating detail: are there intercompany service, shared-personnel, or GPU-lease agreements with Bit Digital, is Bit Digital itself a customer or the unnamed cloud counterparty, what transfer-pricing basis governs those flows, and which controlled-company governance exemptions are being relied upon? Relatedly, Sam Tabar is disclosed as CEO (claims #4701, #4696, #4657) while also serving at the parent — what has not been said is his time allocation between the two entities, how his compensation is split or borne, and what conflicts committee reviews transactions between them?

The IPO was disclosed as 7,812,500 shares with a 1,171,875-share over-allotment option, cited as claims #4655 and #4656; what has not been said is the offer price, gross and net proceeds, whether the over-allotment was exercised, and the allocation of proceeds across MTL-2, MTL-3 and NC-1 construction. Given that every build milestone tied to those proceeds has since slipped, the first-year-analyst question is simple: how much cash remains, what is the remaining committed capex to reach 76 MW, and is additional financing required before NC-1 first revenue?

The company disclosed a three-year service agreement with its first cloud customer on 2025-07-10 and 2025-07-22, cited as claims #4718 and #4695; what has not been said is who that customer is, the contract value or annual revenue run-rate, the GPU count and generation deployed, and whether the agreement has take-or-pay or early-termination provisions. Is this customer a related party, or does it make up most of the cloud segment's revenue?

The company disclosed a phased Duke Energy power commitment of 24 MW by 2025-09-01, 40 MW by 2026-04-01 and 99 MW by 2029-05-16 for NC-1 in Greensboro, cited as claims #4710, #4711, #4712 (repeated as #4686–#4688 and #4644–#4646), alongside the NC-1 property purchase in claims #4709, #4685 and #4643; what has not been said is the site's street address, the substation or interconnection point, and whether these are executed electric service agreements or non-binding capacity letters. With the first two tranches now marked delayed, how much power has Duke actually energized at NC-1 as of today, and what is the revised date for each phase?

The company disclosed guidance of 16 MW by end of 2025 and 76 MW by end of Q4 2026, cited as claims #4715/#4692/#4649 and #4717/#4693/#4654; what has not been said is how the 76 MW is built up. MTL-1, MTL-2 and MTL-3 (claims #4702, #4704, #4706) plus NC-1's 40 MW (claim #4716) sum to 56 MW, which leaves about 20 MW with no named site. The company also has not said whether these figures are critical IT load or gross utility capacity. Given that the 2025 target, the NC-1 24 MW completion (claim #4713) and the NC-1 40 MW milestone (claim #4716) have all slipped, why is the 76 MW year-end target still unrevised?

The company disclosed a 5 MW Cerebras colocation agreement at MTL-3, cited as claims #4708, #4684 and #4642, while the MTL-3 completion and revenue start promised for 2026-01-31 is marked delayed in claim #4707; what has not been said is the contract term, the price per kW-month, whether any penalties or termination rights are triggered by the delay, and whether Cerebras is billing yet. The same question applies to MTL-2's delayed completion (claims #4705, #4647): does any signed tenant sit behind that 5 MW?

The company disclosed that Bit Digital will own 80 percent after the IPO and contributed assets to WhiteFiber, cited as claims #4719 and #4694, and named Sam Tabar as CEO in claim #4701; what has not been said is the terms of the intercompany services, transition and tax-sharing agreements, whether any executives hold roles at both companies, and how executive compensation is split between the two. Neither has the company said whether the controlled-company exemption is being used to waive independent board committee requirements.

The company disclosed an IPO of 7,812,500 shares with a 1,171,875-share over-allotment option, cited as claims #4655 and #4656, and the Enovum acquisition in claims #4703, #4678 and #4638; what has not been said is whether the underwriters exercised the over-allotment, what net proceeds were raised, and how much remaining capex NC-1 needs to reach 40 MW and 99 MW compared with cash on hand. The Enovum purchase price and any earn-out or assumed debt also remain undisclosed.

The company disclosed a three-year service agreement with a cloud customer on 2025-07-10 and again on 2025-07-22, cited as claims #4718 and #4695; what has not been said is who that customer is, what the contract is worth, and how much GPU or MW capacity it covers. The same applies to the unnamed customers at MTL-1 referenced in claim #4639. Is revenue concentration in one undisclosed counterparty material enough to warrant naming it?

The company disclosed a Cerebras colocation agreement of 5.0 MW at MTL-3 on 2025-07-10, 2025-07-22 and 2025-07-28, cited as claims #4708, #4684 and #4642, against a 7.0 MW site (claims #4706 and #4641); what has not been said is whether the 5 MW is critical IT load or gross utility capacity, what the term and pricing are, and whether the slip of MTL-3 completion past both 2025-12-31 (claim #4648) and 2026-01-31 (claim #4707) triggered delay penalties, renegotiation or a revised start date for Cerebras.

The company disclosed Duke Energy capacity commitments of 24 MW by 2025-09-01, 40 MW by 2026-04-01 and 99 MW by 2029-05-16 on 2025-07-10, cited as claims #4710, #4711 and #4712 and restated in #4686–#4688 and #4644–#4645; what has not been said is whether these are executed interconnection or electric service agreements or only load letters. Why was the 24 MW power date set roughly seven months ahead of NC-1's 24 MW build completion (claims #4713 and #4652)? And has any of that power actually been energized now that both near-term tranches are marked delayed?

The company guided to 16 MW by end-2025 and 76 MW by end-Q4 2026 on 2025-07-10, cited as claims #4715 and #4717 and repeated in #4692, #4649, #4693 and #4654; what has not been said is a revised date for the missed 16 MW target, or whether the 76 MW figure has been reaffirmed or cut. NC-1's 24 MW (claim #4713), 40 MW (claim #4716) and May 2026 revenue start (claims #4714, #4690 and #4651) have all slipped, and NC-1 makes up most of that 76 MW. The company also moved MTL-2 and MTL-3 completion from 2025-12-31 (claims #4647 and #4648) to 2026-01-31 (claims #4705 and #4707) without explaining the change.

The company disclosed an IPO of 7,812,500 shares with a 1,171,875-share over-allotment option on 2025-07-28, cited as claims #4655 and #4656, and Bit Digital retaining about 80% ownership post-IPO (claim #4719); what has not been said in this record is the offering price, net proceeds, whether the over-allotment was exercised, and how proceeds are allocated between NC-1 construction and the Enovum/MTL build-out (claims #4703, #4709 and #4680). Given the delays, is NC-1 fully funded to 40 MW, or is further debt or equity required?

The company disclosed Sam Tabar as CEO on 2025-07-10, cited as claim #4701, alongside a Bit Digital asset contribution (claim #4694) and Bit Digital's roughly 80% controlling stake (claim #4719); what has not been said is whether the CEO holds a concurrent role at the parent, how compensation and time are split between the two companies, and on what terms intercompany services, contributed assets and any shared GPU or customer arrangements are priced. These related-party details are what a controlled-company proxy would be expected to address.

The company disclosed a three-year service agreement with a cloud customer on 2025-07-10 and again on 2025-07-22, cited as claims #4718 and #4695. What has not been said is who that customer is, what the contract is worth in total or per year, and how many GPUs or megawatts it covers. Without those figures a reader cannot tell how concentrated the cloud revenue is.

The company disclosed Duke Energy power commitments of 24 MW by 2025-09-01, 40 MW by 2026-04-01 and 99 MW by 2029-05-16, cited as claims #4710/#4686/#4644, #4711/#4687/#4645 and #4712/#4688. The first two are now delayed. What has not been said is whether the delay sits with Duke's interconnection work or with the company's own construction, what the revised energization dates are, and whether the 99 MW is a firm allocation or subject to further studies. The filings also leave unclear whether the 40 MW is cumulative or on top of the 24 MW, and why the 99 MW commitment exceeds the 40 MW phased NC-1 build described in claim #4691.

The company guided to 16 MW by end of 2025 and 76 MW by Q4 2026, cited as claims #4715/#4692/#4649 and #4717/#4693/#4654. The 2025 target is now delayed, and the NC-1 40 MW milestone in claims #4716/#4653 has also slipped, yet the 76 MW target still stands. What has not been said is where the remaining capacity comes from. MTL-1, MTL-2 and MTL-3 at 4, 5 and 7 MW (claims #4702, #4704, #4706) plus 40 MW at NC-1 total only 56 MW, which leaves roughly 20 MW with no named site. The filings also do not say whether these figures are gross utility power or critical IT load.

The company disclosed a 5 MW Cerebras colocation agreement at MTL-3, cited as claims #4708, #4684 and #4642. MTL-3 completion was given as 2026-01-31 in claim #4707 but as 2025-12-31 in claim #4648, and both dates are now missed. What has not been said is the contract term and pricing, whether late delivery triggers penalties or termination rights, and whether Cerebras has received any revised handover date. The MTL-2 timeline shows the same unexplained date change between claims #4705 and #4647.

The company disclosed the NC-1 property purchase in Greensboro, North Carolina, cited as claims #4709, #4643 and #4685. It also gave NC-1 revenue start dates of 2026-05-01 in claims #4714/#4690 and 2026-05-31 in claim #4651, both now delayed. What has not been said is the site's street address, the purchase price, the substation or voltage of the Duke interconnection, or which delayed date the company now considers current.

The company disclosed that Bit Digital will own 80 percent after the IPO, that Bit Digital contributed assets, and that Sam Tabar is CEO. It also disclosed an offering of 7,812,500 shares with a 1,171,875-share over-allotment option. These are cited as claims #4719, #4694, #4701, #4655 and #4656. What has not been said is the offering price and net proceeds, and whether the over-allotment was exercised. The filings also omit the terms of any intercompany services, lease or GPU agreements with the controlling parent. Finally, they do not say how the CEO's time and pay are split between WhiteFiber and Bit Digital, or who on the board is independent of that parent.

The company disclosed a three-year service agreement with its first cloud customer on 2025-07-10 and again on 2025-07-22, cited as claims #4718 and #4695. What has not been said is who that customer is, the total contract value, and how many GPUs or megawatts the agreement covers. Without those figures, a reader cannot tell how concentrated revenue is or whether the contract is material to the build-out it is meant to justify.

The company disclosed a phased Duke Energy power commitment of 24 MW by 2025-09-01, 40 MW by 2026-04-01 and 99 MW by 2029-05-16, cited as claims #4710, #4711, #4712 (repeated as #4686, #4687, #4688, #4644 and #4646). Both near-term dates have now lapsed and are marked delayed. What has not been said is whether Duke has actually energized any of that capacity and what the revised utility dates are. The company also has not said whether the megawatt figures are utility-delivered power or critical IT load, or what interconnection point or substation serves the site.

The company disclosed the purchase of the NC-1 property in Greensboro, North Carolina, cited as claims #4709 and #4643. It guided to 24 MW by Q1 2026 and 40 MW by Q2 2026, cited as claims #4713, #4652 and #4653, and to revenue starting at NC-1. The revenue start appears as 2026-05-01 in claims #4714 and #4690 but as 2026-05-31 in claim #4651. Every one of these dates is now marked delayed. What has not been said is the site's street address, the reason the dates slipped, and which of the two revenue-start dates was the real commitment. The company also has not said whether any customer has signed for NC-1 capacity or whether it is being built speculatively.

The company guided to 16 MW of total capacity by the end of 2025, cited as claims #4715, #4692 and #4649. That figure equals the three Montreal sites combined: MTL-1 at 4 MW (#4702), MTL-2 at 5 MW (#4704) and MTL-3 at 7 MW (#4706). MTL-2 and MTL-3 completion, cited as claims #4705, #4707 and #4648, is marked delayed. What has not been said is how many megawatts are actually energized and billing today. The company also has not said whether the Cerebras 5 MW colocation at MTL-3, cited as claims #4708, #4684 and #4642, carries delivery deadlines, penalties or termination rights that the MTL-3 slip could trigger.

The company reaffirmed a 76 MW capacity target for the end of Q4 2026, cited as claims #4717, #4693 and #4654. The identified sites add up to only about 56 MW: 16 MW in Montreal plus 40 MW at NC-1. What has not been said is which site supplies the remaining roughly 20 MW. The company also has not said whether the 76 MW target still stands after every intermediate milestone behind it slipped, or whether it is measured as gross or IT capacity.

The company disclosed an IPO of 7,812,500 shares with a 1,171,875-share over-allotment option, cited as claims #4655 and #4656. It also disclosed that Bit Digital would own 80 percent after the IPO and contributed the business, cited as claims #4719 and #4694. What has not been said is the net proceeds, whether the over-allotment was exercised, and how the proceeds compare with the capital needed for NC-1. The company also has not described the related-party arrangements with its controlling parent, such as services agreements, cost allocations or intercompany debt. Nor has it said how CEO Sam Tabar's compensation and time are divided between WhiteFiber and Bit Digital, cited as claims #4701, #4696 and #4657.

The company disclosed a three-year service agreement with its first cloud customer on 2025-07-10 and 2025-07-22, cited as claims #4718 and #4695. What has not been said is who the customer is, what the contract is worth in total or per year, how many GPUs or MW it covers, and what share of revenue this one counterparty represents.

The company disclosed Duke Energy commitments of 24 MW, 40 MW and 99 MW, cited as claims #4710/#4686/#4644, #4711/#4687 and #4712/#4688/#4646. The first two are now past their September 2025 and April 2026 dates and marked delayed. What has not been said is whether these figures are utility interconnection capacity or billable critical IT load, what the revised energization dates are, and whether the 99 MW phase by May 2029 is firm or contingent on grid upgrades or customer demand.

The company guided to 16 MW of total capacity by end of 2025, cited as claims #4715, #4692 and #4649. That figure is exactly the sum of MTL-1, MTL-2 and MTL-3 (claims #4702, #4704, #4706). Within the same month, MTL-2 and MTL-3 completion was given as 2026-01-31 (claims #4705, #4707) and as 2025-12-31 (claims #4647, #4648), and all are now delayed. What has not been said is how many MW were actually energized and revenue-generating at year-end 2025, and which date investors were meant to rely on.

The company disclosed the NC-1 property purchase in Greensboro, cited as claims #4709, #4685 and #4643. It targeted 24 MW by Q1 2026 and revenue starting on 2026-05-01 or 2026-05-31, depending on the filing (claims #4713, #4652, #4714, #4651). All of these are now delayed. What has not been said is the purchase price, the site address and substation or interconnection point, whether any NC-1 capacity has a signed tenant, and whether the 76 MW Q4 2026 target (claims #4717, #4693, #4654) is still reaffirmed given that NC-1 supplies most of it.

The company disclosed a 5 MW Cerebras colocation agreement at MTL-3, cited as claims #4708, #4684 and #4642. What has not been said is the contract term, pricing per MW, and committed start date. It also has not said whether the MTL-3 delay deferred Cerebras revenue or triggered service credits, penalties or termination rights.

The company disclosed that Bit Digital would own 80 percent after the IPO and contributed assets to WhiteFiber, cited as claims #4719 and #4694. It also disclosed an IPO of 7,812,500 shares with a 1,171,875-share over-allotment (claims #4655, #4656) and named Sam Tabar as CEO (claim #4701). What has not been said is the value and terms of the asset contribution and Enovum acquisition (claims #4703, #4678), the intercompany service agreements with the controlling parent, whether the over-allotment was exercised, how proceeds were split between NC-1 and Montreal, and how the CEO's pay and time are allocated if they also hold a role at Bit Digital.

The company disclosed IPO share offerings of 7.8 million shares with a 1.2 million share overallotment on 2025-07-28, cited as claims 4655 and 4656; what has not been said is the IPO price per share and total capital proceeds, rendering the implied valuation and capitalization structure opaque to investors.

The company disclosed Bit Digital's 80% post-IPO ownership stake on 2025-07-10, cited as claim 4719, and referenced Bit Digital's "contribution" on 2025-07-22, cited as claim 4694; what has not been said is the consideration paid, timing of the transfer, vesting or clawback conditions, earnout provisions, or the valuation basis for the 80% figure.

The company announced the Enovum data center acquisition on 2025-07-10 (claim 4703), 2025-07-22 (claim 4678), and 2025-07-28 (claim 4638) with a promised close of 2024-10-01; what has not been said is the acquisition price, financing structure, integration status, or revised timeline given the "delayed" designation—investors cannot model the financial impact or timing of capacity realization.

The company disclosed an unnamed "cloud customer" with a three-year service agreement on 2025-07-10, cited as claim 4718, and separately mentioned a "first cloud customer" on 2025-07-22, cited as claim 4695; what has not been said is whether these are the same counterparty, the annual utilization or revenue implied, and pricing escalation terms—details required to validate the capacity guidance.

The company disclosed three Duke Energy capacity tranches (24 MW, 40 MW, 99 MW) across claims 4710–4712 (2025-07-10), 4686–4688 (2025-07-22), and 4644–4646 (2025-07-28); what has not been said is the levelized power price, minimum take-or-pay terms, contract duration, renewal and termination conditions, exclusivity restrictions, and any force majeure carve-outs—inputs essential for revenue forecasting.

The company identified three data center properties (NC-1 Greensboro, MTL-2 Montreal, MTL-3 Saint-Jérôme) in claims 4709, 4704, and 4706; what has not been said is the specific street address, the primary utility substation interconnection point, transmission voltage class, or the grid operator identity for each site—details investors would use to independently verify available capacity and power costs.

The company disclosed a 'cloud customer three-year service agreement,' cited as claim 4718 on 2025-07-10 (also referenced as 'first cloud customer' in claim 4695 on 2025-07-22); what has not been said is which customer has signed this agreement and why the company cannot disclose its identity, given that Cerebras and Duke Energy are named.

The company disclosed that Bit Digital maintains 80% ownership post-IPO, cited as claim 4719 on 2025-07-10; what has not been said is the governing terms, voting rights, transfer restrictions, anti-dilution provisions, board composition arrangements, or economic participation rights of Bit Digital's controlling stake.

The company disclosed Cerebras colocation agreement for 5 MW capacity, cited as claims 4708, 4642, and 4684; what has not been said is the contract term length, pricing per MW, take-or-pay obligations, or revenue timing, despite naming Cerebras as a customer whereas other customer agreements remain unnamed.

The company disclosed Duke Energy capacity commitments totaling 99 MW (Phase 1: 24 MW by Sept 2025, Phase 2: 40 MW by April 2026, Phase 3: 99 MW by May 2029), cited as claims 4710, 4711, and 4712; what has not been said is the pricing per MW, commercial structure (fixed vs. variable), take-or-pay minimums, or pricing escalation terms for these capacity commitments.

The company disclosed an Enovum data center acquisition promised by October 2024, cited as claims 4703 and 4638; what has not been said is whether this acquisition was renegotiated to a new close date, abandoned entirely, or remains pending, given that it is marked as delayed more than 20 months past the promised date.

The company disclosed NC-1 will expand from 24 MW initial to 40 MW total via phased expansion, cited as claims 4691 and 4713 (initial 24 MW promised by March 2026 per claims 4650 and 4652); what has not been said is the specific completion timeline, capex requirements, and trigger events for Phase 2 expansion, or whether it is contingent on Duke Energy Phase 2 commitment (claim 4687).

The company disclosed a 5 MW Cerebras colocation agreement (claims 4708, 4684, 4642) but has not stated the revenue rate, contract duration, or service model. Given Cerebras is the only named customer besides Duke Energy, the absence of annualized revenue is striking; what is the expected revenue per MW and multi-year ARR impact of the Cerebras contract?

The company announced a three-year cloud customer service agreement (claim 4718) and separately a first cloud customer (claim 4695), but disclosed neither customer name nor revenue or capacity. While megawatt targets are quantified precisely, customer revenue is entirely withheld; what revenue amount is the three-year cloud customer agreement worth?

The Enovum data center acquisition (claims 4703, 4678, 4638) was promised by October 2024 and remains unexplained 22 months later, with zero disclosure of acquisition price, assets acquired, financing terms, or delay reason. What was the Enovum purchase price, what capacity did it add, and has the acquisition closed?

The NC-1 property in Greensboro, North Carolina (claims 4709, 4643) was announced without street address, zoning status, or utility interconnection details—power feed capacity, water availability, fiber connectivity. For a $10M+ HPC real estate investment, what is the actual grid capacity available at NC-1, and what is the utility interconnection timeline?

The Duke Energy partnership comprises three tiers (24 MW delayed, 40 MW delayed, 99 MW on track to 2029, claims 4710–4712) but the company has not explained whether these are sequential or parallel, what contractual triggers phase them, or whether earlier delays affect the 99 MW timeline. What are the conditions precedent for Duke Energy's 99 MW commitment, and how does it depend on delivery of the earlier phases?

The company disclosed a Cerebras colocation agreement (5 MW, claims 4642/4684) and a cloud customer three-year service agreement (claims 4695, 4718) but has not named any other customers despite 76 MW capacity guidance by end-2026 (claim 4654). What has not been said is the customer concentration and identities of counterparties representing the remaining 66 MW of capacity—how many customers are signed, what is the largest customer's share, and are any revenue-critical customers withheld from disclosure?

Three Duke Energy tranches are disclosed (24 MW by 2025-09-01, 40 MW by 2026-04-01, 99 MW by 2029-05-16; claims 4644, 4645, 4688), but the first phase is already delayed as of September 2026. What has not been said is whether these commitments contain termination rights, liquidated damages provisions, or force majeure language, and whether Duke Energy can unilaterally reduce or exit remaining tranches if WhiteFiber continues to miss delivery deadlines.

The company disclosed Bit Digital's 80% post-IPO ownership (claim 4719) and an underwriter over-allotment option (claim 4656), indicating a recent IPO, but has not disclosed the total capital expenditure required to build out 76 MW by end-2026 (claim 4654), cumulative deployment to date, or the funding roadmap beyond Bit Digital. What has not been said is the capex budget, monthly burn rate, and whether debt facilities or secondary equity raises are planned to fund the remaining buildout.

Enovum data center acquisition is stated as completed (claims 4638, 4678) with a target close of October 2024, but no purchase price, cash-or-stock consideration, or Enovum's existing customer base and revenues are disclosed. What has not been said is whether Enovum's assets and customers contribute meaningfully to WhiteFiber's stated capacity and revenue targets, or if the acquisition was primarily strategic.

NC-1 Greensboro property acquisition (claims 4643, 4685) targets 24 MW energization by Q1 2026 (claims 4650, 4652) and phased expansion to 40 MW total (claim 4691), but no disclosure addresses grid interconnection status, available transmission capacity from the utility, or the capital cost for cooling and site infrastructure. What has not been said is whether the Duke Energy capacity commitments are contingent on successful grid interconnection and what the site-level capex breakdown is for electrical infrastructure, cooling systems, and construction.

Sam Tabar is disclosed as CEO (claims 4657, 4701), and Bit Digital holds 80% ownership (claim 4719), but no executive compensation, equity incentive structures, or governance protections for minority shareholders are disclosed in the record. What has not been said is Sam Tabar's base salary, bonus metrics, and equity vesting schedules, as well as what board representation Bit Digital holds and what safeguards prevent related-party transactions or conflicts of interest.

The company disclosed a three-year cloud customer service agreement (claim 4718) and reference to a first cloud customer (claim 4695), but has not disclosed the customer's identity or any commercial terms beyond the multi-year commitment. For an infrastructure provider in capital raise mode, anchor customer identity typically features prominently in pitch materials; what is the name of this customer, and what annual revenue is committed under the three-year agreement?

The company announced three separate Duke Energy partnerships (24 MW by 2025-09-01 per claims 4710/4686/4644—now 12 months overdue; 40 MW by 2026-04-01 per claims 4711/4687—now 5 months overdue; 99 MW by 2029-05-16 per claims 4712/4688/4646) but has not disclosed whether these are power purchase agreements, capacity reservations, or interconnection commitments, nor the pricing, take-or-pay terms, or ramp schedules. Why are the first two phases significantly delayed, and what is Duke Energy's recourse if timelines slip further?

The Cerebras colocation agreement for 5.0 MW (claims 4708, 4684, 4642) lacks disclosure of pricing per MW, contract duration, minimum utilization commitments, or expansion options. Is this a fixed-price lease, and does Cerebras have penalties or early-termination rights if WhiteFiber's facility underperforms?

The NC-1 Greensboro property acquisition (promised by 2025-05-20 per claims 4685/4643), initial 24 MW energization (promised by 2026-03-31 per claims 4713/4650/4652), and revenue generation (promised by 2026-05-01 per claims 4714/4690) are all now significantly overdue as of September 2026, yet the company has not disclosed the current status—whether the property is acquired, what permitting or interconnection bottlenecks exist, or the revised timeline. How many MW is currently operational at NC-1, and what is the breakdown of delays between land acquisition, permitting, construction, and grid interconnection?

The company disclosed Bit Digital's 80% post-IPO ownership (claim 4719) and an IPO raise of 7.8M primary shares plus 1.2M over-allotment option (claims 4655/4656), but has not disclosed the IPO share price, total capital raised, or how an 80% majority stake is mechanically maintained post-public-offering. Are there ongoing related-party service agreements, management fees, or contingent payments to Bit Digital, and what governance safeguards protect minority shareholders from self-dealing?

The company identified CEO Sam Tabar (claims 4701/4696/4657) but has not disclosed executive compensation, equity holdings, or related-party transactions. What are the CEO's base salary, bonus structure, and equity awards, particularly any change-of-control provisions or incentives tied to the delayed facility milestones?

WhiteFiber disclosed a cloud customer with a three-year service agreement (claim 4718) and a 'first cloud customer' (claim 4695) without naming either party, while simultaneously naming Duke Energy and Cerebras—what has not been said is the revenue contribution, PPA rate per kWh, term length, and renewal optionality of this unnamed customer, which a first-year analyst would request to reconcile the claimed 5 MW Cerebras commitment with total facility utilization. The company announced the Enovum data center acquisition (claims 4703, 4678, 4638) without stating the purchase price, the physical location of the acquired asset (is it one of MTL-1, MTL-2, MTL-3, or NC-1?), its capacity specifications, or the integration timeline—material omissions when assessing organic vs. acquisition-driven capacity growth. MTL-2 and MTL-3 both carry stated operational deadlines of 2026-01-31 (claims 4705, 4707) yet show 'delayed' status as of 2026-09-06; what has not been disclosed is whether these facilities remain on track, what the revised timelines are, or whether construction or financing constraints have indefinitely paused the projects. NC-1 in Greensboro is promised at 24 MW (Q1 2026), 40 MW (Q2 2026), and three tranche Duke Energy commitments (24/40/99 MW across 2025–2029), yet the company has not disclosed the maximum site capacity, the utility interconnection point and capacity ceiling at the specific property address, or whether the Duke Energy phases are sequenced by demand pull or facility readiness—critical for assessing whether the guidance of 76 MW by end-2026 is achievable. Bit Digital is disclosed to retain 80 percent ownership post-IPO (claim 4719), but the company has not stated what Bit Digital's pre-IPO ownership was, whether the IPO diluted from 100 percent, the current investor cap table, or any management overlap or related-party arrangements that could create conflicts of interest. The IPO raised capital via 7,812,500 shares plus a 1,171,875-share over-allotment option (claims 4655, 4656), but the company has not disclosed the IPO price per share, total capital raised, or the allocation of proceeds across facility build-out, debt repayment, and working capital—essential for validating whether the capital base is sufficient to deliver the stated 76 MW guidance by 2026.

The company disclosed a "three-year service agreement" with a cloud customer on 2025-07-10 (claim 4718) and again on 2025-07-22 (claim 4695); what is the customer's name, and what is the annual recurring revenue and total contract value of this agreement? The company promised completion of the Enovum data center acquisition by 2024-10-01 (claim 4638), but it remained marked "delayed" as of 2026-09-03, more than twenty months overdue; has this acquisition been terminated, is it still pending, or has it completed with a revised timeline? The company announced Duke Energy capacity commitments in three phases totaling 163 MW (claims 4710, 4711, 4712 from 2025-07-10; repeated in claims 4686, 4687, 4688 from 2025-07-22), but never disclosed whether these are firm purchase orders with penalty clauses, framework agreements, or preliminary commitments, nor the pricing per MW or revenue recognition model for each phase. The company disclosed that Bit Digital retained 80% ownership after its IPO (claim 4719 from 2025-07-10) and announced an IPO offering of 7,812,500 shares (claim 4655 from 2025-07-28); how is this 80% ownership structure reconciled in a public company, and what governance disclosures—including board composition, related-party transaction policies, and CEO compensation arrangements—were made given Bit Digital's controlling stake? The company guided to 76 MW total capacity by end 2026 (claims 4717, 4654), but disclosed only four facilities totaling approximately 56 MW (MTL-1: 4 MW, MTL-2: 5 MW, MTL-3: 7 MW, NC-1: 40 MW); what additional capacity, facilities, or acquisitions comprise the remaining ~20 MW? The company announced facility revenue start dates for MTL-2 (by 2026-01-31, claim 4705), NC-1 (by 2026-05-01, claims 4714, 4690), and MTL-3 (claim 4707), but provided no facility-level revenue guidance, pricing per MW, or blended pricing assumptions; what are the expected annual revenues or revenue run-rates for each facility, and what pricing per MW underlies the Duke Energy and cloud customer commitments?

The company disclosed a "first cloud customer" and a "cloud customer three-year service agreement" on July 2025, cited as claims 4695 and 4718; what has not been said is the customer name, capacity allocation, revenue amount, or pricing terms. For a company scaling to 76 MW with delayed facility timelines, the absence of publicly disclosed customer contracts leaves material uncertainty about demand validation and revenue visibility.

The company disclosed Enovum data center acquisition with a close promised by October 2024 on July 2025, cited as claim 4638; what has not been said is the actual close date, integration status, or explanation for the 22-month delay. A transaction of this scale typically requires clear disclosure of close date and operational progress milestones.

The company disclosed MTL-2 and MTL-3 facility completion promised by December 2025 on July 2025, cited as claims 4705 and 4707; what has not been said is the current construction status, revised timelines, or revenue generation start dates, particularly given that MTL-3 is committed to power the Cerebras colocation customer. Nine months past the original deadline with no timeline update suggests material execution risk.

The company disclosed Duke Energy capacity commitments of 24 MW, 40 MW, and 99 MW on July 2025, cited as claims 4710, 4711, and 4712; what has not been said is whether these are binding contracts or non-binding LOIs, the pricing structure per MW, take-or-pay obligations, or the reason why the first two tranches are delayed while the third is stated. The commercial and legal structure of these partnerships is essential for understanding revenue certainty.

The company disclosed NC-1 Greensboro facility with promised capacity delivery in Q1 and Q2 2026 on July 2025, cited as claims 4709, 4713, and 4714; what has not been said is the street address, interconnecting utility, grid interconnection application status, or remaining permitting milestones. These facility-level details are standard for evaluating data center delivery timeline credibility and power supply certainty.

The company disclosed Bit Digital's 80% post-IPO ownership and a 1.17M share underwriter over-allotment on July 2025, cited as claims 4719 and 4656; what has not been said is the total post-IPO share count, IPO pricing per share, resulting dilution percentage, and governance terms such as board representation or veto rights for the controlling shareholder. These cap table and related-party disclosures are standard proxy-level transparency for companies with controlling shareholders.

The company disclosed a Cerebras colocation agreement for 5.0 MW capacity (claims #4708, 4642) but has not disclosed the contract end date, renewal conditions, or annual revenue contribution; what material terms govern this customer relationship and how much revenue does Cerebras generate annually?

A three-year cloud customer service agreement is announced (claims #4718, 4695) without the customer's name, capacity allocation, or contract value; what is the customer identity and what annual revenue does this commitment represent?

NC-1 property acquisition in Greensboro, North Carolina is announced (claims #4709, 4643) but lacks disclosure of the specific utility provider, grid interconnection point, or primary network carrier; what is the interconnection topology and who are the utility and network counterparties serving this facility?

Enovum acquisition (claims #4703, 4638) and MTL-2/MTL-3 property acquisitions (claims #4704, 4640, 4706, 4641) are announced with capacity targets but without disclosing purchase prices, retrofit costs, or integration expenses; what is the total capex required to bring each facility to revenue-generating status, and what was paid for the Enovum acquisition?

The IPO issuance of 7.8M shares (claim #4655) plus 1.17M share over-allotment (claim #4656) is announced without disclosing offer price, total capital raised, or capital allocation plan; how much capital was raised and how will proceeds be deployed across construction, acquisitions, working capital, and debt service?

Duke Energy phase commitments are announced for 24 MW (2025-09-01), 40 MW (2026-04-01), and 99 MW (2029-05-16) (claims #4710–4712, #4686–4688), with two phases now delayed; what penalties or renegotiation rights does Duke Energy have if completion dates slip further, and what is the company's contingency plan if phases are cancelled or further delayed?

The company disclosed a 'cloud customer three-year service agreement' on 2025-07-10, cited as claim 4718; what has not been said is the customer's identity and the quantified revenue impact of this multi-year contract. The company promised Duke Energy 24 MW capacity by 2025-09-01, disclosed on 2025-07-10, cited as claims 4710, 4686, 4644; what has not been said is the current status of this capacity, which is marked delayed as of September 2026, contradicting its inclusion in 2025 guidance of 16 MW total. The company announced the Enovum data center acquisition on 2025-07-10, cited as claim 4703; what has not been said is the purchase price, the capacity or revenue acquired, or whether Bit Digital (which owns 80% post-IPO per claim 4719) facilitated or funded this transaction. The company disclosed acquisition of NC-1 property in Greensboro, North Carolina on 2025-07-28, cited as claim 4643; what has not been said is the specific street address, the utility interconnection agreement status, permitting timeline, or grid capacity confirmation from Duke Energy, despite promising 24 MW completion by Q1 2026 and revenue generation by May 2026. The company disclosed a Cerebras colocation agreement for 5 MW on 2025-07-28, cited as claim 4642; what has not been said is the contract term duration, revenue model, and whether this represents a pilot deployment or a multi-year committed arrangement. The company disclosed Bit Digital's 80% post-IPO ownership on 2025-07-10, cited as claim 4719, and a 7.8M share IPO offering on 2025-07-28, cited as claim 4655; what has not been said is the total shares outstanding post-IPO, voting rights allocation, and the dilution impact on public minority shareholders resulting from 80% retention.

The company disclosed a "Cloud customer three-year service agreement" on 2025-07-10, cited as claim 4718, positioned as material customer traction, but has not named the customer or disclosed contract value, revenue model, or capacity allocation. What has not been said is: which customer is this, what are the financial terms, and how many MW are reserved, so investors can assess revenue quality and capacity availability?

The company announced acquisition of "NC-1 Greensboro property" on 2025-07-10, cited as claim 4709, and later targeted 40 MW capacity by Q2 2026 (claims 4650, 4652, 4653), yet provided no street address, acreage, grid interconnection point, or utility provider. What has not been said is: where exactly is the site, what is the transmission line or substation interconnection, and what are the infrastructure details that support the 40 MW capacity and timeline claim?

The company stated "Bit Digital post-IPO ownership 80.0 percent" on 2025-07-10, cited as claim 4719, and announced a 7,812,500 share IPO offering (claim 4655), but the ownership structure remains opaque. What has not been said is: what is the total shares outstanding post-IPO, the IPO price, and the capitalization table showing Bit Digital's pre-IPO and post-IPO stakes, so investors understand control and dilution?

The company announced a "Duke Energy capacity agreement" on 2025-07-10 with three phases (24 MW by Sept 2025, 40 MW by April 2026, 99 MW by May 2029), cited as claims 4710, 4711, 4712, but the first two phases are now delayed as of September 2026. What has not been said is: what are the renegotiation triggers, termination clauses, or penalty provisions if capacity milestones are missed, and does the contract include force majeure or customer-at-fault provisions?

The company disclosed "Cerebras colocation agreement 5.0 MW" on 2025-07-10, cited as claims 4708 and 4642, at MTL-3 (7.0 MW total capacity per claim 4706), but did not clarify whether the 5 MW is exclusive or whether remaining capacity is available. What has not been said is: how much of the 7 MW is actually available to other customers, and are there other pre-commitments or oversubscription arrangements that affect the stated capacity?

MTL-2 (5 MW) and MTL-3 (7 MW) were promised operational by December 2025 (claims 4647, 4648) and now show delayed status as of September 2026, yet the company has not disclosed whether these remain in development or have been abandoned. What has not been said is: whether these facilities are included in the 76 MW Q4 2026 capacity guidance (claims 4654, 4717), and if the remaining 17 MW gap is expected from these sites or new facilities?

The company disclosed a "Cloud customer three-year service agreement" (claim 4718) and referenced a "First cloud customer" (claim 4695). What is this customer's name, what is the annual revenue amount, and what capacity and SLA commitments has WhiteFiber guaranteed?

The company stated Bit Digital maintains 80% ownership post-IPO (claim 4719) while issuing 7.8 million IPO shares (claim 4655) plus 1.2 million over-allotment shares (claim 4656). Given this capital structure, what percentage of fully diluted shares does the public float represent, and how is board control apportioned between Bit Digital and public shareholders?

WhiteFiber announced the Enovum data center acquisition (claims 4703, 4678) with an October 2024 completion deadline (claim 4638), now marked delayed as of September 2026. What was the transaction price, what are the facility's actual power capacity and square footage, and what is the revised completion timeline?

The company disclosed three phased Duke Energy partnerships: 24 MW by September 2025 (claims 4710, 4686), 40 MW by April 2026 (claims 4711, 4687), and 99 MW by May 2029 (claims 4712, 4688). What are the per-MW unit economics (monthly or annual price), are there take-or-pay or volume minimums, and what are the contract termination and exclusivity terms?

Across four facilities (MTL-1, MTL-2, MTL-3, NC-1), the company disclosed capacity reaching 76 MW by end-2026 (claims 4702-4706, 4716-4717). What are the street addresses and interconnection points, what is the regulatory approval status for each site, and on what basis (owned, leased, optioned) does WhiteFiber control the properties?

The company announced a Cerebras colocation agreement for 5 MW at MTL-3 (claims 4642, 4684). What is the contract term, what is the pricing per MW, and does the agreement restrict WhiteFiber's ability to sign competing tenants at the same facility?

The company disclosed a 'Cloud customer three-year service agreement' (claim 4718) and 'First cloud customer' (claim 4695) at its July 2025 IPO announcements but never disclosed the customer's identity, contract value, or revenue start date. For a company highlighting customer wins during IPO roadshow, a sophisticated reader would expect the customer name and commercial terms, not a placeholder title—particularly given the company announced material customer relationships but only Cerebras is named.

WhiteFiber announced the Enovum data center acquisition as completed status in July 2025 (claims 4703, 4678) yet claim 4638 shows the same acquisition was promised by 2024-10-01 and remains marked delayed. The inconsistency raises a critical question: did this asset actually close before the IPO on 2025-07-28, and if so, by what date? Without confirmation, investors cannot verify core assets were in WhiteFiber's ownership at IPO close.

The NC-1 Greensboro property acquisition (claim 4709) was disclosed as stated but claim 4685 reveals it was promised by 2025-05-20 and is delayed—yet this property is foundational to the 24-40 MW North Carolina phased build-out (claims 4689, 4691, 4713, 4716). Did WhiteFiber actually acquire and control this site before IPO, or did the company go public without owning one of its three core properties?

CEO Sam Tabar is named (claim 4701) but no proxy disclosures of compensation, equity ownership, or related-party governance appear. Coupled with Bit Digital's disclosed 80% post-IPO ownership (claim 4719), the claim set lacks clarity on public float percentage, board independence, and whether Tabar is Bit Digital's appointee—standard disclosures for an IPO with 80% insider retention.

Three Duke Energy partnership phases totaling 163 MW (claims 4710, 4711, 4712) are disclosed with promised-by dates spanning 2025-09-01 through 2029-05-16, but no contract terms, pricing per MW, binding vs. optional commitment status, or revenue recognition timing are revealed. Investors have no basis to assess whether these 'commitments' represent firm orders or conditional capacity reservations.

The company disclosed a "Cloud customer three-year service agreement" on 2025-07-10, cited as claim 4718; what has not been said is the customer's name, the annual revenue commitment, the capacity allocation, or the pricing terms.

The company announced NC-1 Greensboro property acquisition on 2025-07-10, cited as claim 4709, with promised 24 MW completion by 2026-03-31 (claim 4713); what has not been said is the facility's street address, the specific grid interconnection point, or the utility approval status.

The company disclosed three Duke Energy capacity commitments on 2025-07-10—24 MW (claim 4710), 40 MW (claim 4711), and 99 MW (claim 4712)—totaling 163 MW; what has not been said is whether these are binding long-term power purchase agreements or nonbinding letters of intent, what per-MW revenue rate applies, or what contractual penalties WhiteFiber faces if it misses capacity completion deadlines (particularly relevant given all three phases are currently delayed).

The company announced MTL-1 Montreal is "operational" with 4.0 MW capacity on 2025-07-10, cited as claim 4702; what has not been said is the annual revenue generated, the names or count of customers, or the current capacity utilization percentage.

The company disclosed Bit Digital's 80% post-IPO ownership stake on 2025-07-10, cited as claim 4719; what has not been said is the terms of related-party transactions between WhiteFiber and Bit Digital, whether Bit Digital receives preferential pricing on capacity, what governance or board control rights Bit Digital retains, or the history of how this ownership stake was initially acquired.

The company disclosed IPO terms of 7.8 million shares offered and 1.17 million share overallotment on 2025-07-28, cited as claims 4655 and 4656; what has not been said is the IPO share price, the implied post-IPO enterprise valuation, or the lock-up expiration dates for key shareholders including the 80% Bit Digital stake.

The company disclosed a 'cloud customer three-year service agreement' on 2025-07-10 (claims 4718, 4695) but provided no customer name, capacity level, or pricing terms. Given that Cerebras is disclosed by name at 5 MW capacity, what has not been said is the identity and contract economics of this material customer agreement.

The company announced a Duke Energy Phase 1 capacity commitment of 24 MW promised by 2025-09-01 (claims 4710, 4644, 4686) but this is now delayed as of 2026-09-03—over one year past the stated deadline. What has not been said is the current status of this major partnership, whether delivery terms have been renegotiated, any impacts on the Phase 2 (40 MW April 2026) and Phase 3 (99 MW May 2029) commitments, or contractual consequences of the delay.

The company promised NC-1 Greensboro at 24 MW energized by Q1 2026 and 40 MW by Q2 2026 (claims 4713, 4653, 4716), both now delayed as of 2026-09-03. What has not been said is the status of interconnection agreements with Duke Energy or the relevant utility, the Point of Interconnection assignment, any grid-related permitting bottlenecks, or whether utility coordination is the root cause of facility timeline slippage.

The company disclosed an Enovum data center acquisition promised for 2024-10-01 on 2025-07-10 (claims 4703, 4638, 4678), now delayed as of 2026-09-03. What has not been said is the purchase price paid, whether the transaction was cash- or debt-financed, what specific assets or operating contracts were acquired, or whether earnout obligations are tied to facility performance.

The company disclosed Bit Digital's 80% post-IPO ownership on 2025-07-10 (claim 4719) but provided no detail on capitalization. What has not been said is how much capital Bit Digital contributed pre- and post-IPO, what voting rights or preferred returns Bit Digital holds, whether Bit Digital has reserved or discounted capacity allocations across facilities, or what affiliation Sam Tabar (the named CEO) has to Bit Digital's principals.

The company promised MTL-2 and MTL-3 completion by 2025-12-31 and 2025-04-11 respectively (claims 4681, 4682, 4683, 4705, 4707) but both are now delayed as of 2026-09-03, over eight months past promised dates. What has not been said is the root cause of these delays (permitting, supply chain, financing, construction), the revised completion timelines, whether capex has overrun budget, or how the delays cascade into the 40 MW and 76 MW 2026 guidance targets.

The company disclosed a three-year service agreement with a "Cloud customer" (claims 4718, 4695, announced July 2025) but withheld the customer's identity and revenue contribution. Given that the company explicitly named Cerebras Systems in parallel colocation announcements (claims 4708, 4642, 4684), the omission is conspicuous: what is not being said is who this first major cloud customer is and what revenue they represent under the three-year agreement.

The company promised Duke Energy Phase 1 capacity of 24 MW by 2025-09-01 (claims 4710, 4686, 4644, all July 2025), a deadline now passed with status marked "delayed." What has not been disclosed is the reason for missing this commitment, the revised completion timeline, and whether Duke Energy retains termination, penalty, or price-adjustment rights—material context since Phase 2 (40 MW, April 2026) and Phase 3 (99 MW, May 2029) depend on Phase 1 credibility (claims 4711, 4712, 4687, 4688).

Across four July 2025 announcements, the company disclosed major acquisitions: Enovum (claims 4703, 4638), MTL-2 5.0 MW (claims 4680, 4640), MTL-3 7.0 MW (claims 4682, 4641), and NC-1 Greensboro (claims 4709, 4685, 4643), yet disclosed neither aggregate acquisition cost nor capital structure—whether funded by bank debt, IPO proceeds, or equity. Given Bit Digital's 80% post-IPO ownership (claim 4719), what has not been said is the total debt-to-equity ratio and whether these properties carry secured debt.

The company repeatedly refers to MTL-1 as having "4.0 MW operational capacity" (claims 4702, 4639, 4679, July 2025) without clarifying whether this means power installed, grid-connected, or actively generating revenue. For a facility claimed as pre-existing and operational, what has not been disclosed is the current revenue run rate from MTL-1 and its capacity utilization?

The Cerebras colocation agreement (5.0 MW, claims 4708, 4684, 4642, July 2025) lacks disclosure of contract term length, pricing per MW, and contingency clauses. Given NC-1 Greensboro's initial 24 MW completion was promised by 2026-03-31 (claims 4713, 4650) and is now delayed, what has not been said is whether Cerebras is contingent on NC-1 delivery or backed by alternate capacity such as MTL-3?

On 2026-08-16, WhiteFiber disclosed a $60M real estate acquisition in Yadkin County, North Carolina (claim 29) and announced electrical power capacity targets of 60 MW minimum (claim 31) and 198 MW potential expansion (claim 32). What has not been said is the specific property address or parcel information, total acreage, existing on-site infrastructure status, and critical details on electrical grid interconnection—information essential for a sophisticated investor evaluating a material real estate transaction.

Claim 31 promises 60 MW of minimum electrical power capacity at the acquired properties by 2026-10-30, the same deadline as property acquisition closing (claim 29). What has not been said is whether this 60 MW represents installed nameplate capacity available immediately upon closing, permitted capacity pending construction, or theoretical maximum—a distinction that determines whether the timeline is achievable and what capital will be required beyond the $60M purchase price.

Claim 25, disclosed on 2026-08-20, announces Energy Service Agreements for Data Centers without naming counterparties, disclosing pricing terms, specifying contract duration, or indicating whether these are long-term take-or-pay commitments. What has not been said is the identity of these agreement counterparties, pricing structure, and whether they represent guaranteed revenue streams or optional arrangements.

Claim 26, also disclosed 2026-08-20, announces GPU Servers for Cloud Business with no disclosure of quantity of units, specific GPU model specifications, the supplier or manufacturer, expected deployment timeline, or total capital allocation. A first-year analyst would ask: how many units, which GPU architecture, when will they be delivered, and what is the total investment magnitude?

Claims 22 and 8, dated 2026-08-20 and 2026-08-21, report a convertible note exchange of $198.15M for $118.5M plus 6.3M shares, yet the encoded amount appears as 118.5 USD, creating ambiguity in whether net proceeds figures in claims 21 and 7 ($298.5 USD) represent millions or another denomination. What has not been said is the precise cash-to-shares conversion mechanics and complete accounting of settlement.

Claim 19, disclosed 2026-08-25, identifies Unifi Manufacturing as counterparty to a real estate purchase agreement but discloses no purchase price, expected closing date, specific property address or acreage, confirmation of related-party status, or detail of business assets transferring with the real estate. What has not been said is whether this Unifi transaction is material and on what timeline investors should expect closing.

The company's 16 MW year-end 2025 capacity target (claims 4715 and 4692) is the clearest silent lapse. It was stated on 2025-07-10 and repeated on 2025-07-22, with a promised date of 2025-12-31. The automated ledger moved it to delayed on 2026-09-03 and then to not_yet_verifiable on 2026-10-09. The figure was the sum of MTL-1 (4 MW, claims 4702 and 4679), MTL-2 (5 MW, claims 4704 and 4680) and MTL-3 (7 MW, claims 4706, 4682 and 4641), so meeting it depended on both newer Montreal sites. Nothing on record says whether the 16 MW was reached, by when, or that the target was revised.

MTL-2's completion and revenue start, promised for 2026-01-31 (claim 4705), is now marked delayed. The ledger moved it to delayed on 2026-09-03 and again on 2026-10-10. The site was last referenced on 2025-07-28, as a property acquisition (claim 4640). The 2025-07-22 disclosure described an MTL-2 operational timeline without restating a date (claim 4681), which may have been the first softening of the January commitment. The record contains no revised date, no announcement that the site began producing revenue, and no explanation for the slip.

MTL-3 in Saint-Jérôme was presented as a 7 MW data center on 2025-07-10 (claim 4706). By 2025-07-22 it had become a site acquisition (claim 4682), and by 2025-07-28 a site secured (claim 4641). No completion date or construction milestone was ever attached to it, and it does not appear after July 2025. Its status has moved back and forth between stated, delayed and not_yet_verifiable without any new company statement. What has not been said is whether MTL-3 is being built, on what schedule, and whether its 7 MW still counts toward the 76 MW end-2026 guidance (claims 4717, 4693 and 4654), which remains on record unchanged.

The Cerebras colocation arrangement (claims 4708 and 4684) appeared as a 5 MW partnership on 2025-07-10. Twelve days later it was recast as a customer relationship, with no change in stated terms. The unnamed cloud customer's three-year service agreement (claims 4718 and 4695) followed a similar path, from a specific contract term to a generic first cloud customer. Neither is mentioned after 2025-07-22 in the claims on record. There has been no disclosure of which facility hosts Cerebras, whether the 5 MW is live, or whether the cloud agreement is still in effect.

By contrast, the North Carolina commitments have been met as stated. These are the Duke Energy 24 MW and 40 MW tranches (claims 4710, 4686, 4711 and 4687) and the NC-1 milestones (claims 4713, 4714, 4716, 4690, 4650 and 4653), all now verified. The gaps above sit in Montreal and in the customer relationships. The claims ledger as supplied ends partway through claim 4649, so any later disclosures beyond that point were not reviewed.

The MTL-2 expansion in Montreal is the clearest case of a deadline that passed without comment. In the 2025-07-10 disclosure, management tied MTL-2 completion and the start of revenue to 31 January 2026 (claim 4705). Eighteen days later, the 2025-07-28 filing gave an earlier target of 31 December 2025 (claim 4647), and the 2025-07-22 materials referred to an operational timeline without a date (claim 4681). Neither date was met in the record we hold. Claim 4705 has been classified as delayed since 2026-09-03 and was most recently confirmed as delayed on 2026-10-10. Claim 4647 moved to delayed on 2026-10-10. The last mention of MTL-2's schedule was 2025-07-28. We have found no revised completion date, no explanation of why the target moved by a month between two filings issued less than three weeks apart, and no statement on whether the 5 MW expansion (claims 4704, 4680, 4640) is generating revenue.

The 16 MW year-end 2025 capacity guidance has gone quiet in a way that follows from the MTL-2 slip. Claims 4715 (2025-07-10) and 4692 (2025-07-22) committed to 16 MW of total capacity by 31 December 2025. That figure appears to equal MTL-1's 4 MW, MTL-2's 5 MW and MTL-3's 7 MW combined. The guidance was last repeated on 2025-07-22. Both claims were flagged delayed on 2026-09-03 and have been not_yet_verifiable since 2026-10-09, more than nine months after the target date. Management has not said whether 16 MW was reached at year-end, what capacity was actually energized, or how the MTL-2 delay affected the total.

The headline 76 MW target for the end of 2026 is still on the record but has not been restated. It appears in claims 4717 (2025-07-10), 4693 (2025-07-22) and 4654 (2025-07-28), with no later mention. The NC-1 milestones that form most of the build-out have been verified (claims 4713, 4716, 4714, 4690), but the target still depends on Montreal capacity that is behind schedule. With less than three months until the 31 December 2026 deadline, the ledger holds no reaffirmation, no breakdown by site, and no acknowledgement of whether the earlier Montreal shortfall carries into this figure.

The Cerebras colocation agreement and the first cloud customer contract were disclosed as signed commercial anchors and have not been mentioned since. The 5 MW Cerebras arrangement at MTL-3 appears in claims 4708, 4684 and 4642. The three-year cloud service agreement appears in claims 4718 and 4695. All were last mentioned in July 2025 and remain at their original stated status. The MTL-3 site itself (claim 4641) passed through delayed and not_yet_verifiable classifications before returning to stated on 2026-10-09. Management has not said whether Cerebras has occupied the space, when billing began, or whether the cloud agreement remains in force on its original terms.

More generally, every claim on file for WhiteFiber dates from a three-week window in July 2025 around the IPO. The ledger shows no forward-looking statement from management after 2025-07-28, so none of the slips above has been addressed in a disclosure we have captured.

MTL-2, the 5 MW Montreal expansion, is the clearest case. On 2025-07-10 the company tied MTL-2 completion and revenue start to 2026-01-31 (claim 4705). The 2025-07-22 disclosure described an MTL-2 operational timeline without a date (claim 4681). The 2025-07-28 prospectus put completion at 2025-12-31 (claim 4647). That is the last mention on record. Both dated claims have now passed and the ledger marks them delayed. Nothing in the record gives a revised completion date, reconciles the conflicting December and January targets, or says whether MTL-2 is generating revenue. The capacity itself is still carried as stated (claims 4704, 4680, 4640).

The 16 MW year-end 2025 capacity guidance has also gone quiet. It appears on 2025-07-10 (claim 4715) and 2025-07-22 (claim 4692). By 2025-07-28 it had dropped out of the guidance language, which refers only to the 76 MW Q4 2026 target. The figure matches MTL-1 at 4 MW, MTL-2 at 5 MW and MTL-3 at 7 MW. With MTL-2 delayed, the guidance probably depended on a facility that did not arrive on schedule. The deadline passed more than nine months ago and both claims sit at not_yet_verifiable. No year-end capacity figure, confirmation or downward revision has been disclosed.

The 76 MW total capacity target for end of Q4 2026 was repeated across all three July 2025 disclosures (claims 4717, 4693, 4654). It has not been restated since 2025-07-28. The deadline is now about eleven weeks away. The target was built on the same Montreal sites, including the delayed MTL-2, plus NC-1. The record has no reaffirmation, bridge to the figure or acknowledgment of the MTL-2 slip's effect on it.

The Cerebras colocation agreement for 5 MW at MTL-3 was described as a partnership on 2025-07-10 (claim 4708) and as a customer agreement on 2025-07-22 and 2025-07-28 (claims 4684, 4642). The July 28 language attached it to MTL-3, the Saint-Jérôme site that was then described as secured rather than built (claim 4641). There has been no mention since. No commencement date, energization milestone or revenue contribution has been disclosed. The unnamed cloud customer's three-year service agreement (claims 4718, 4695) has the same gap: it was last referenced on 2025-07-22, with no term start, counterparty or scope given afterward.

More generally, every claim on record for WhiteFiber dates from a three-week window around the July 2025 IPO, with nothing after 2025-07-28. The NC-1 and Duke Energy milestones have since been marked verified (claims 4713, 4714, 4710, 4711 and related). Those are automated ledger determinations made in October 2026, not new management statements. The record contains no post-IPO disclosure updating any of the timelines set out at listing.

Every claim on file for WhiteFiber dates from a three-week window around its July 2025 IPO, between 10 July and 28 July. Nothing in the record comes from later disclosures. The silences below should be read against that gap. They are topics management put dates and numbers on at listing and has not revisited in anything captured since.

The MTL-2 Montreal expansion is the clearest slip. The 10 July filing put completion and revenue start at 31 January 2026 (claim 4705). The 28 July document put completion at 31 December 2025 (claim 4647), and the 22 July filing described an operational timeline without a date (claim 4681). The target therefore already moved between documents in the same month without explanation. Both dated claims now sit at delayed, last mentioned on 28 July 2025, and their deadlines passed more than eight months ago. No revised date, cause, or confirmation that the 5 MW (claims 4704, 4680, 4640) is energized appears anywhere in the record.

The guidance of 16 MW total capacity by the end of 2025 was last stated on 22 July 2025 (claims 4715, 4692). Its status is not_yet_verifiable more than nine months after the deadline. This figure appears to depend partly on MTL-2 coming online. Management has not said whether the 16 MW was reached at year-end, nor has it restated the number lower.

The Cerebras colocation agreement for 5 MW at MTL-3 in Saint-Jérôme was described in all three IPO-period documents (claims 4708, 4684, 4642), most recently on 28 July 2025. The 7 MW MTL-3 site was described as acquired or secured (claims 4706, 4682, 4641). None of these claims carries a delivery date, and the record contains no later word on build progress, energization, or whether the Cerebras contract has started billing. The same applies to the three-year agreement with the first cloud customer (claims 4718, 4695). It was presented as an anchor contract at listing and has not been referenced since 22 July 2025.

The headline target of 76 MW total capacity by the end of Q4 2026 (claims 4717, 4693, 4654) was last stated on 28 July 2025 and still reads only as stated. The North Carolina pieces of that figure have since verified (claims 4713, 4714, 4716). The Montreal pieces have not. With under three months left to the deadline, the record holds no reaffirmation of the 76 MW, no updated breakdown by site, and no acknowledgment of how the MTL-2 delay affects it.

Every claim on file for WhiteFiber comes from three disclosures issued within eighteen days of each other in July 2025: 10 July, 22 July and 28 July. The record holds nothing after 28 July 2025. Each lapsed date below was therefore set in the IPO-era filings and has since passed without a revision on record. The status changes in the ledger were made by automated extraction, not by the company. Several acquisition items flagged not_yet_verifiable, such as Enovum (4638), MTL-2 (4640, 4680), MTL-3 (4641, 4682) and the NC-1 purchase (4643, 4685), carry dates that were already in the past when first disclosed. They describe completed transactions rather than open promises and are not treated as walk-backs here.

The clearest lapse is the 16 MW total capacity target for year-end 2025. It appears in all three July 2025 disclosures (4715, 4692, 4649) and was last stated on 28 July 2025. The deadline passed more than nine months ago, and the ledger moved the claim to delayed on 3 September 2026 and to not_yet_verifiable on 9 October 2026. The record contains no statement confirming the target was met, revising it, or explaining a shortfall. The 76 MW target for end-2026 (4717, 4693, 4654) still stands as stated. That longer-dated target is built on the shorter one, which has gone unaddressed.

The NC-1 Greensboro schedule has slipped in pieces. The initial 24 MW was due by 31 March 2026 (4713), the expansion to 40 MW by 30 June 2026 (4716), and the matching Duke Energy power tranches by 1 September 2025 for 24 MW (4710, 4686, 4644) and 1 April 2026 for 40 MW (4711, 4687). All of these dates have passed, and none has a completion or revision notice on record. The revenue start date also moved within the July filings themselves. On 10 and 22 July it was 1 May 2026 (4714, 4690); on 28 July it was 31 May 2026 (4651). Nothing on record explains that month's shift. The 28 July filing describes Duke Phase 1 (4644) and Phase 3 (4646) but contains no Phase 2 claim corresponding to the 40 MW tranche. Its absence may reflect drafting rather than intent, but it has not been accounted for.

MTL-2 in Montreal was given a completion and revenue start of 31 January 2026 in the 10 July 2025 disclosure (4705). On 22 July the timeline (4681) appears without a date, and by 28 July no MTL-2 timeline claim was extracted at all, only the historical property acquisition (4640). The January 2026 date has since lapsed. The record holds no confirmation that MTL-2 is energised or earning revenue, and no statement replacing the date.

The first cloud customer's three-year service agreement was cited on 10 July (4718) and 22 July (4695). It does not appear among the 28 July claims, which name Cerebras at MTL-3 (4642) as the customer relationship. The customer was never identified, and no later disclosure on record says whether the agreement began, was amended, or is still in force.

The record on file for WhiteFiber consists entirely of disclosures dated between July 10 and July 28, 2025, around the company's IPO. No later statement appears in this ledger, so every item below has a last-mention date of 2025-07-28 or earlier. Each "delayed" status was assigned automatically on 2026-09-03 because a promised date had passed. None reflects a revision the company announced. Because the ledger shows no later disclosures at all, part of this silence may be a gap in collection rather than in what the company has said.

The two Montreal expansions, MTL-2 (5 MW) and MTL-3 (7 MW), carried completion dates that did not agree across the IPO-period filings. The July 10 disclosure tied completion and revenue start for both to January 31, 2026 (claims 4705, 4707). The July 28 disclosure gave December 31, 2025 instead (claims 4647, 4648). The July 22 timeline entries gave no date at all (claims 4681, 4683). Both dates have now passed, and the record contains no statement that either facility was completed or energized. It also contains no explanation of which date was the operative one or what moved it.

The headline guidance of 16 MW of total capacity by end-2025 was repeated in all three filings (claims 4715, 4692, 4649). Its deadline passed nine months ago. Nothing on record confirms the figure was reached, and nothing restates it as missed or revised. Since that total depends on the MTL-2 and MTL-3 build-outs above, it appears to have lapsed along with them without comment.

The North Carolina program at NC-1 Greensboro has the densest cluster of lapsed dates. The Duke Energy Phase 1 power commitment of 24 MW was due September 1, 2025 (claims 4710, 4686, 4644). The 24 MW initial energization was due in Q1 2026 (claims 4713, 4652, 4650). Phase 2 power of 40 MW was due April 1, 2026 (claims 4711, 4687, 4645), and 40 MW of cumulative site capacity was due by the end of Q2 2026 (claims 4716, 4653). NC-1 revenue was to start May 1, 2026, or May 31 in the July 28 version (claims 4714, 4690, 4651). All of these have passed with no update on the record. No statement addresses whether the delay sits with the utility interconnection, construction, or customer contracting. The longer-dated 99 MW Phase 3 commitment for 2029 is still described as stated (claims 4712, 4688, 4646).

The 76 MW year-end 2026 capacity target is still nominally intact (claims 4717, 4693, 4654). However, the near-term milestones it was built from have each lapsed without restatement. Management has not said whether the 76 MW figure still stands. The named commercial anchors have also gone unmentioned since July 2025. These are the 5 MW Cerebras colocation at MTL-3 (claims 4708, 4684, 4642) and the three-year cloud customer agreement (claims 4718, 4695). There is no word on whether either has moved to delivery or been affected by the MTL-3 slip.

The schedule for NC-1, the Greensboro, North Carolina site, carries the most dated commitments in the record and has gone the longest without a revised date. In July 2025 disclosures, management said the initial 24 MW would be energized and complete by the end of the first quarter of 2026 (claims 4713, 4650, 4652). It said revenue would begin in May 2026, dated May 1 in claims 4714 and 4690 and May 31 in claim 4651. It said cumulative capacity would reach 40 MW by the end of the second quarter of 2026 (claims 4716, 4653). The last mention on record is 2025-07-28. All of these passed their deadlines and were moved to delayed by automated transition on 2026-09-03, not by any company revision. Nothing on record says whether NC-1 energized, whether revenue began, or what the new dates are.

The Montreal expansions MTL-2 (5 MW) and MTL-3 (7 MW) were given two slightly different finish lines within the same month. On 2025-07-10, management tied both to completion and revenue start by January 31, 2026 (claims 4705, 4707). On 2025-07-28, it described completion by the end of 2025 (claims 4647, 4648). The 2025-07-22 timeline claims (4681, 4683) carry no date at all. Both sets of deadlines have lapsed and were auto-marked delayed. The 5 MW Cerebras colocation agreement placed at MTL-3 (claims 4708, 4684, 4642) still reads as stated, but it depends on a facility whose completion has never been confirmed in the record.

The Duke Energy power arrangement was described in three phases. The first two have passed their dates without comment: 24 MW by September 1, 2025 (claims 4710, 4686, 4644) and 40 MW by April 1, 2026 (claims 4711, 4687, 4645). The 99 MW phase dated May 2029 (claims 4712, 4688, 4646) is still carried as stated. The record does not say whether either earlier tranche of power was actually delivered. That question also governs NC-1's energization dates above.

The capacity guidance has the same gap. The target of 16 MW total by the end of 2025 (claims 4715, 4692, 4649) lapsed and was auto-marked delayed, with no year-end figure disclosed. The larger target of 76 MW by the fourth quarter of 2026 (claims 4717, 4693, 4654) still stands as stated. Yet its building blocks, the Montreal expansions and the first 40 MW at NC-1, have all slipped. No revised guidance or reconciliation appears on record, and the last restatement of either figure was 2025-07-28.

Two caveats apply. Every delayed status here is a deadline-driven automated transition, not evidence of a confirmed miss. And the claims ledger as supplied ends partway through the 2025-07-28 entries, so later disclosures may exist that are not reflected. Separately, several acquisition claims are flagged delayed only because their reference dates fell before the filing: the MTL-2 acquisition (4680), the MTL-3 site (4682, 4641), and the NC-1 purchase (4685, 4643). These read as completed events rather than slipped promises.

The 16 MW year-end 2025 capacity target is the clearest lapsed commitment. Management stated it three times in July 2025: in the 2025-07-10 filing (claim 4715), on 2025-07-22 (claim 4692), and on 2025-07-28 (claim 4649). The record contains no mention after 2025-07-28. The December 31, 2025 deadline passed, and the claims were automatically moved to delayed on 2026-09-03. Nothing on record says whether 16 MW was reached, revised, or dropped, and no replacement year-end figure has been disclosed.

The Montreal expansions, MTL-2 (5 MW) and MTL-3 (7 MW), carry completion dates that shifted within the July 2025 filings themselves. The 2025-07-10 disclosure tied completion and revenue start for both to January 31, 2026 (claims 4705 and 4707). The 2025-07-22 filing described their operational timelines without a date (claims 4681 and 4683). The 2025-07-28 filing then gave December 31, 2025 (claims 4647 and 4648). Both dates have now passed, and every dated claim was auto-transitioned to delayed on 2026-09-03. The record does not reconcile the two dates, confirm that either site is energized or earning revenue, or say whether the Cerebras 5 MW colocation at MTL-3 (claims 4708, 4684 and 4642) has started.

The NC-1 Greensboro schedule has lapsed at each milestone. In July 2025 management promised the initial 24 MW by March 31, 2026 (claims 4713, 4650 and 4652) and 40 MW cumulative by June 30, 2026 (claims 4716 and 4653). The revenue start date moved from May 1, 2026 (claims 4714 and 4690, last stated 2025-07-22) to May 31, 2026 (claim 4651, 2025-07-28), a one-month slip within eighteen days that no filing on record explains. All of these dates have passed, all were marked delayed on 2026-09-03, and no disclosure confirms energization, a first tenant, or a revised schedule.

The Duke Energy power commitments that underpin NC-1 have also gone unaddressed past their dates. The 24 MW tranche due September 1, 2025 (claims 4710, 4686 and 4644) and the 40 MW tranche due April 1, 2026 (claims 4711, 4687 and 4645) were last described on 2025-07-28, then as phases of a capacity agreement. Both are now delayed. Nothing on record states whether Duke delivered the power, whether the phasing was renegotiated, or whether the delay sits with the utility or with WhiteFiber.

The 76 MW year-end 2026 target remains formally stated (claims 4717, 4693 and 4654), as does the 99 MW Duke Energy phase for 2029 (claims 4712, 4688 and 4646). Both were last reaffirmed on 2025-07-28. Every interim milestone that 76 MW depends on has since lapsed without comment. Management has not said whether the target still stands with less than three months left in 2026.

Every claim on record for WhiteFiber dates from a three-week window in July 2025: disclosures dated July 10, July 22 and July 28, around the IPO. Nothing in the ledger shows a later company statement. Each 'delayed' status below was applied automatically on 2026-09-03 after the promised date passed. None reflects a revision announced by management. The record also contains no reaffirmation, revised date or explanation for any of them.

The NC-1 Greensboro build-out is the largest of these gaps. Management said the initial 24 MW would be completed and energized by the end of Q1 2026 (claims 4713, 4650, 4652). It said revenue would start in May 2026 (claims 4714, 4690, 4651). It said the site would reach 40 MW cumulative by the end of Q2 2026 (claims 4716, 4653). The revenue start date itself shifted within the July filings, from May 1 (4714, 4690) to May 31 (4651), without comment. The last mention of all three milestones was July 28, 2025, and all three passed their dates. The record does not say whether NC-1 was energized, whether it produced revenue, or what the new schedule is.

The Duke Energy power agreement underpins NC-1. It was described in phases: 24 MW by September 1, 2025 (claims 4710, 4686, 4644), 40 MW by April 1, 2026 (claims 4711, 4687, 4645), and 99 MW by May 16, 2029 (claims 4712, 4688, 4646). The first two dates have passed and were marked delayed. The 99 MW phase still carries the original 'stated' language from July 2025. Nothing on record confirms that Duke delivered the Phase 1 or Phase 2 capacity, or that the phases were rescheduled. The 2029 commitment has not been restated against that background.

The Montreal expansions show a similar pattern. On July 10, MTL-2 (5 MW) and MTL-3 (7 MW, Saint-Jérôme) were each said to reach completion and revenue start by January 31, 2026 (claims 4705, 4707). On July 22 they appeared as undated 'operational timelines' (claims 4681, 4683). On July 28 they were given an end-of-2025 completion target (claims 4647, 4648). Both later dates have passed. The Cerebras 5 MW colocation agreement at MTL-3 (claims 4708, 4684, 4642) has not been mentioned since July 28, 2025. The record does not say whether Cerebras is live at the site.

The headline capacity guidance follows from these delays. The target of 16 MW total by the end of 2025 (claims 4715, 4692, 4649) passed without any statement of the capacity actually reached. The 76 MW target for the end of 2026 (claims 4717, 4693, 4654) is still marked 'stated' because its date has not arrived. However, it depends on the NC-1 and Montreal milestones above, and it has not been reaffirmed or revised since July 2025.

For completeness, several acquisition claims were also flagged 'delayed': MTL-2 (4680), MTL-3 (4682), NC-1 (4685, 4643) and Enovum (4638). Their 'promised by' dates fall before the July 2025 disclosures, so they appear to describe transactions already closed rather than walked-back commitments. They should not be read as evidence of slippage.

The North Carolina build-out is the clearest case. In the July 2025 disclosures, management tied NC-1's first 24 MW to a Duke Energy Phase 1 commitment due by September 1, 2025 (claims 4710, 4686, 4644), with energization in Q1 2026 (4713, 4650, 4652) and revenue starting in May 2026 (4714, 4690, 4651). The revenue date was given as May 1 in two filings and May 31 in the July 28 filing, and that gap was never explained. These milestones were last mentioned on July 28, 2025. All of them passed their stated dates and were automatically marked delayed on September 3, 2026. In the available record, nothing confirms that the 24 MW was energized, that Duke delivered the Phase 1 power, that NC-1 has produced revenue, or what the new timetable is.

The Montreal expansions show the same pattern, with an inconsistency added. The July 10 disclosure gave January 31, 2026 for completion and revenue at both MTL-2 and MTL-3 (4705, 4707). The July 28 filing instead gave December 31, 2025 for both (4647, 4648). Both dates have passed and the claims now read delayed. The record does not say whether either facility was finished. This also affects the Cerebras colocation agreement (4708, 4684, 4642). Management placed that 5 MW at MTL-3, and it is still carried as stated, but nothing has been said about whether the customer was ever served from a facility that has no confirmed completion date.

The interim capacity guidance has lapsed while the headline target stays in place. The end-2025 target of 16 MW total capacity (4715, 4692, 4649) and the 40 MW at NC-1 by the end of Q2 2026 (4716, 4653) have both passed, as has the Duke Phase 2 40 MW commitment due April 1, 2026 (4711, 4687, 4645). All are now marked delayed. The 76 MW by end-2026 target (4717, 4693, 4654) and the 99 MW Duke Phase 3 commitment for 2029 (4712, 4688, 4646) still appear as stated. The record shows no reaffirmation and no revision of them since July 28, 2025. Management has not addressed how the end-2026 figure can hold when every intermediate step toward it has been missed or has gone unreported.

The unnamed first cloud customer and its three-year service agreement (4718, 4695) were last mentioned on July 22, 2025. Nothing in the record discloses the counterparty, the contract value, or whether the agreement is still in force.

Several acquisition claims are also flagged delayed: Enovum, the MTL-2 and MTL-3 site purchases, and the NC-1 property (4638, 4680, 4682, 4685, 4643). Their promised dates fall before the filings that mention them, so they most likely describe transactions that were already completed rather than promises that slipped. They are not treated here as walk-backs. More broadly, every delayed status in this ledger comes from automatic date expiry, not from any company statement. The silence is the finding: the record holds no WhiteFiber disclosure after July 2025 that updates any of these timelines.

Every claim on file for WhiteFiber comes from three documents dated 10, 22 and 28 July 2025, the period around the IPO. The record holds no later disclosure that restates, revises or retires any of them. Each delayed status below was assigned automatically on 3 September 2026 because a promised date passed. None reflects a company statement. The ledger supplied here also ends partway through claim 4651, so later entries were not reviewed.

The Duke Energy power arrangement for NC-1 was described in all three July 2025 documents as a phased commitment. The first tranche was 24 MW by 1 September 2025 (claims 4710, 4686, 4644). The second was 40 MW by 1 April 2026 (claims 4711, 4687, 4645). The third was 99 MW by 16 May 2029 (claims 4712, 4688, 4646). The record was last updated on 28 July 2025. Both near-term dates have since passed and are now marked delayed, while the 2029 tranche still reads as stated. The record contains nothing on whether the 24 MW or 40 MW was energized on schedule, late, or on revised terms.

The NC-1 Greensboro build-out has the same gap. The 24 MW initial phase was due by 31 March 2026 (claims 4713, 4652, 4650). Revenue was to begin by 1 May 2026 (claims 4714, 4690). Cumulative capacity was to reach 40 MW by 30 June 2026 (claims 4716, 4653). These were last mentioned on 28 July 2025, and all three dates have passed and are marked delayed. No revised schedule, revenue-start confirmation or explanation appears in the record. NC-1 was the core of the growth case at the IPO.

The Montreal expansions MTL-2 (5 MW) and MTL-3 (7 MW, Saint-Jérôme) were given different completion dates in documents weeks apart. The 10 July material said 31 January 2026 (claims 4705, 4707). The 28 July material said end of 2025 (claims 4647, 4648). The 22 July filing gave operational timelines without a date (claims 4681, 4683). Both projects are now delayed under either date. The 5 MW Cerebras colocation agreement depends on MTL-3 and still reads as stated (claims 4708, 4684, 4642). Nothing in the record explains the shifting date, says whether MTL-3 is live, or says whether the Cerebras arrangement has started generating revenue.

The headline capacity guidance has drifted apart from its own building blocks. The interim target of 16 MW by end of 2025 (claims 4715, 4692, 4649) has lapsed and is marked delayed. The 76 MW by end of 2026 target (claims 4717, 4693, 4654) still stands as stated, with under three months left. It depends on the NC-1, MTL-2 and MTL-3 milestones above, all of which are marked delayed. The record contains no reaffirmation, reduction or reconciliation of the 76 MW figure since July 2025.

One caveat applies to the acquisition entries for MTL-2, MTL-3 and the NC-1 property (claims 4680, 4640, 4682, 4641, 4685, 4643). These also show as delayed, but their promised-by dates appear to be the historical closing dates. They were likely extracted as deadlines in error and should not be read as walked-back commitments.

Every claim on record for WhiteFiber dates from a three-week window around its July 2025 IPO, between 2025-07-10 and 2025-07-28. Nothing on file since then revises, confirms or withdraws those commitments. Each item below rests on dates that have now passed. The delayed labels were applied automatically on 2026-09-03 and are not company statements.

The nearest-term promise was 16 MW of total capacity by the end of 2025 (claims 4715, 4692, 4649). It was last repeated in the 2025-07-28 disclosure. The deadline passed more than nine months ago and the claim is now marked delayed. No year-end capacity figure, revised target or explanation appears in the record. The 16 MW appears to be the sum of MTL-1, MTL-2 and MTL-3, so whether that total was ever reached is not on record.

The Montreal expansions, MTL-2 at 5 MW and MTL-3 at 7 MW, were given completion and revenue-start dates of 2026-01-31 in the 2025-07-10 disclosure (4705, 4707). Eighteen days later, the 2025-07-28 filing put them at 2025-12-31 (4647, 4648), and nothing in the record reconciles the two. Both dates have passed and the claims are marked delayed. The 5 MW Cerebras colocation agreement at MTL-3 (4708, 4684, 4642) depends on MTL-3 being finished. It was last described on 2025-07-28 and still carries its original stated status. The record does not say whether Cerebras has taken occupancy, whether the deployment moved elsewhere, or whether the agreement still stands.

NC-1 in Greensboro was the centerpiece of the growth case, and the most milestones have lapsed there. The first 24 MW of Duke Energy power was promised by 2025-09-01 (4710, 4686, 4644). The initial 24 MW was to be completed and energized by the end of Q1 2026 (4713, 4652, 4650). Revenue was to start in May 2026, given as 2026-05-01 in 4714 and 4690 and as 2026-05-31 in 4651. All of these are now marked delayed. The second phase, with Duke's 40 MW commitment by 2026-04-01 (4711, 4687, 4645) and 40 MW at NC-1 by the end of Q2 2026 (4716), has also passed without comment. The record contains no statement on whether Duke delivered power on schedule, whether energization happened, or when NC-1 began or will begin producing revenue.

The headline target of 76 MW by the end of 2026 (4717, 4693, 4654) is still recorded as stated, and its deadline has not yet arrived. However, it was built on the 16 MW year-end 2025 base and the NC-1 phases, both of which have lapsed without public revision. Management has not restated the 76 MW figure since 2025-07-28, either to reaffirm it or to adjust it for the slippage underneath. The same applies to the unnamed cloud customer's three-year service agreement (4718, 4695). It was announced in July 2025, and its identity, size and status have not been mentioned in the record since.

Several acquisition claims are also marked delayed, including Enovum (4638), MTL-2 (4680), MTL-3 (4682, 4641) and NC-1 (4643, 4685). Their dates refer to transactions completed before the IPO. These appear to be artifacts of automated date-matching rather than missed commitments, and they are not counted above.

The NC-1 Greensboro build schedule is the most concrete set of commitments that has gone quiet. Across three July 2025 disclosures, management said the first 24 MW would be completed or energized by the end of Q1 2026 (claims 4713, 4650, 4652), cumulative capacity would reach 40 MW by the end of Q2 2026 (4716, 4653), and revenue would begin in May 2026 (4714, 4690, 4651). The last mention of these dates in the record is 2025-07-28. All of them passed without verification and were auto-transitioned to delayed on 2026-09-03. The record contains no revised energization date, no statement on whether any NC-1 capacity is live, and no explanation of what held the timeline back.

The Duke Energy power tranches that were meant to feed NC-1 have followed the same path. The 24 MW tranche was due by 2025-09-01 (4710, 4686, 4644) and the 40 MW tranche by 2026-04-01 (4711, 4687). Both are now marked delayed, with no confirmation in the record that either was delivered. The 99 MW tranche dated 2029-05-16 is still carried as stated (4712, 4688, 4646). That keeps the long-dated headline number standing, while nothing has been said about whether the nearer-term increments it depends on actually arrived.

The headline capacity guidance has not been reconciled with these slips. The 16 MW total-capacity target for year-end 2025 (4715, 4692, 4649) passed and was auto-marked delayed, and no achieved figure was disclosed in its place. The 76 MW target for the end of Q4 2026 (4717, 4693, 4654) still stands as stated, even though NC-1 was its largest building block and every dated NC-1 milestone beneath it has lapsed. No updated or reduced figure appears in the record, and nothing confirms that 76 MW is still achievable within the year.

The Montreal expansions have also left only stale dates behind. MTL-2 was due for completion and revenue by 2026-01-31 in one July 2025 filing (4705) and by 2025-12-31 in another (4647). That gap within the same month was never addressed. MTL-3 at Saint-Jérôme carried a 2026-01-31 completion and revenue date (4707). Both are now delayed. The undated operational-timeline claims (4681, 4683) and the 5 MW Cerebras colocation at MTL-3 (4708, 4684, 4642) remain listed as stated, but no subsequent disclosure says whether MTL-3 is serving Cerebras, whether the agreement has moved to revenue, or when either Montreal site came online.

Many of the delayed items are acquisition dates that were already in the past when first disclosed, such as Enovum, MTL-2, MTL-3 and the NC-1 property (4638, 4640, 4641, 4643, 4680, 4682, 4685). These more likely reflect a missing confirmation in the ledger than an actual slip, and they are not treated here as walk-backs.

WhiteFiber's year-end 2025 capacity target is the clearest lapsed commitment on the record. The company guided to 16 MW of total capacity by 31 December 2025 in three successive July 2025 disclosures (claims 4715, 4692 and 4649), the last dated 28 July 2025. The deadline passed and the ledger moved all three to delayed automatically on 3 September 2026. No later claim confirms the 16 MW was reached, restates it with a new date, or explains a shortfall.

The Montreal expansions carry the same gap, and the record was not internally consistent even before the dates passed. MTL-2 (5 MW) and MTL-3 (7 MW) were each given a completion and revenue start of 31 January 2026 in claims 4705 and 4707 on 10 July 2025. On 28 July 2025, claims 4647 and 4648 described completion by the end of 2025 instead. A third pair, claims 4681 and 4683 from 22 July, referred to an operational timeline with no date. All dated versions are now delayed. Nothing on record says whether either site is energized or revenue-generating, or which of the two timelines management considered authoritative. This also matters for the Cerebras colocation agreement, which claims 4708, 4684 and 4642 placed at MTL-3 for 5 MW. That agreement remains at stated, but the facility it depends on has no confirmed completion.

The North Carolina build-out accounts for most of the slipped dates. NC-1 in Greensboro was to complete its initial 24 MW in Q1 2026 (claims 4713, 4650 and 4652). It was to start generating revenue in May 2026, which claim 4714 placed on 1 May and claim 4651 placed on 31 May, a drift within the same month of filings that was never explained. It was to reach 40 MW by 30 June 2026 (claims 4716 and 4653). The underlying Duke Energy power commitments were described as 24 MW by 1 September 2025 (claims 4710, 4686 and 4644) and 40 MW by 1 April 2026 (claims 4711, 4687 and 4645). Every one of these is now delayed, and the last mention of any of them is 28 July 2025. The record contains no statement on whether Duke delivered either tranche, whether NC-1 has energized any capacity, or when revenue began, if it has. The 99 MW Duke tranche dated May 2029 (claims 4712, 4688 and 4646) is still at stated only because its date has not yet arrived.

The headline target of 76 MW by the end of Q4 2026 (claims 4717, 4693 and 4654) still reads as stated, with under three months left. That status reflects the calendar, not progress. The interim milestones the 76 MW was built on, at MTL-2, MTL-3 and both NC-1 phases, have all lapsed without confirmation. Management has not reaffirmed the figure, revised it, or reconciled it with the missed intermediate dates since July 2025.

One caution applies to how these entries are read. Several acquisitions were already complete when they were disclosed: MTL-2 in December 2024 (claim 4680), the MTL-3 site in April 2025 (claims 4682 and 4641), and the NC-1 property in May 2025 (claims 4685, 4709 and 4643). They appear as delayed only because the transition rule treats any past date without later verification as a slip. Those entries should not be read as walk-backs. The operational and capacity milestones above are a different matter, because each describes future delivery that has gone unaddressed since the July 2025 filings.

The NC-1 campus in Greensboro was the centerpiece of WhiteFiber's July 2025 disclosures, and its whole build schedule has passed without comment. Management said the initial 24 MW would be energized and complete by the end of Q1 2026 (claims 4713, 4650, 4652). It said revenue would begin in May 2026, given as May 1 in claims 4714 and 4690 and as May 31 in claim 4651. It said cumulative capacity would reach 40 MW by the end of Q2 2026 (claims 4716, 4653). The last mention on record is the July 28, 2025 prospectus-era filing. Every one of these dates was automatically moved to delayed on 2026-09-03 because it had passed. Nothing on record says whether the first 24 MW was energized, whether NC-1 has produced revenue, or what the new schedule is.

The Duke Energy power commitments that NC-1 depends on have gone the same way. WhiteFiber described 24 MW of utility capacity available by September 1, 2025 (claims 4710, 4686) and 40 MW by April 1, 2026 (claims 4711, 4687, 4645). Both were last mentioned in July 2025, and both were auto-flagged delayed in September 2026. The company has not said whether Duke delivered either tranche, or whether the arrangement still reads as the firm commitment it was presented as. The third phase, 99 MW by May 2029 (claims 4712, 4688, 4646), still carries stated status, but only because its date has not yet arrived.

The Montreal expansions, MTL-2 (5 MW) and MTL-3 in Saint-Jérôme (7 MW), were already inconsistent within July 2025 itself. Claims 4705 and 4707 put completion and revenue start at January 31, 2026. Claims 4647 and 4648 put completion at December 31, 2025. Claims 4681 and 4683 describe an operational timeline with no date at all. Both deadlines have lapsed and were auto-flagged delayed. This matters most for the 5 MW Cerebras colocation agreement at MTL-3 (claims 4708, 4684, 4642), the company's most prominent named customer. It was last referenced on July 28, 2025, and there is no record of the facility being delivered or of the contract beginning to bill.

The headline capacity guidance shows a gap between its parts and its whole. The end-of-2025 target of 16 MW total (claims 4715, 4692, and apparently 4649, which is truncated in this record) has lapsed and is marked delayed, with no actual year-end figure disclosed. The 76 MW target for Q4 2026 (claims 4717, 4693, 4654) remains formally stated, but nearly every milestone that was supposed to add up to it has slipped. The company has neither reaffirmed nor revised the number since July 2025, and the record does not reconcile how 76 MW is still achievable.

The unnamed cloud customer described as signing a three-year service agreement (claims 4718, 4695) also appears only in the July 2025 materials, with no later identification, contract value, or revenue attribution on record.

Two cautions apply to the record itself. First, several acquisitions flagged delayed are past-dated historical events, not missed promises: Enovum (4638), MTL-2 (4680, 4640), MTL-3 (4682, 4641) and NC-1 (4685, 4643). They are excluded from this analysis. Second, every claim in the ledger comes from three July 2025 documents, and every status change was made by automated date-passing rather than by any company statement. The silence described here may partly reflect gaps in what has been ingested. On what is on file, however, no revision announcement accompanies any of the slipped milestones.

WhiteFiber's guidance of 16 MW of total capacity by the end of 2025 appears in three disclosures from July 2025 (claims 4715, 4692 and 4649). The most recent was the July 28, 2025 prospectus. The deadline passed, and on September 3, 2026 the ledger moved all three claims to delayed. The record contains no later statement from management saying whether the 16 MW was reached, how much capacity was actually online at year-end 2025, or whether the target was formally revised.

The Montreal expansions, MTL-2 and MTL-3, were presented as near-term revenue contributors. MTL-2 completion and revenue start carried a January 31, 2026 date in the July 10 disclosure (claim 4705) and a December 31, 2025 date in the July 28 prospectus (claim 4647). MTL-3 also carried a January 31, 2026 date (claim 4707). The July 22 disclosure described both timelines without dates (claims 4681 and 4683). All dated versions are now marked delayed. No announcement in the record confirms that either facility was completed, sets a new date, or explains why the company's own July 2025 documents gave two different MTL-2 dates. The Cerebras 5 MW colocation agreement tied to MTL-3 (claims 4708, 4684 and 4642) remains recorded as stated, but nothing in the record says whether service under that agreement has started.

The NC-1 Greensboro build depended on a phased power commitment from Duke Energy. That commitment was 24 MW by September 1, 2025 (claims 4710, 4686 and 4644) and 40 MW by April 1, 2026 (claims 4711, 4687 and 4645). The site milestones were 24 MW completed in Q1 2026 (claims 4713, 4652 and 4650), revenue starting May 1, 2026 (claims 4714 and 4690), and 40 MW cumulative by the end of Q2 2026 (claims 4716 and 4653). Every one of these dates has passed and every claim is now marked delayed. The last mention of each was in July 2025. The record has no update on whether Duke delivered power on schedule, whether NC-1 has been energized, or whether revenue has begun.

The headline target of 76 MW by the end of Q4 2026 (claims 4717, 4693 and 4654) and the 99 MW Duke Energy phase for 2029 (claims 4712, 4688 and 4646) are still recorded as stated. The 76 MW figure depends on the NC-1 and Montreal milestones above, all of which have lapsed without comment. Reaching it would require nearly all of that capacity to come online within the final quarter of 2026. Management has not publicly reconciled the missed intermediate dates with the unchanged year-end target.

Several acquisition claims are also flagged delayed: Enovum, the MTL-2 and MTL-3 site purchases, and the NC-1 property (claims 4638, 4640, 4641, 4643, 4680, 4682 and 4685). Their dates fall before the disclosures that describe them, so they appear to be completed transactions. The automated flag more likely reflects how the claims were extracted than any reversal by the company, and they are not treated here as walked back.

The most consequential silence concerns NC-1 in Greensboro. Across three disclosures between 10 July and 28 July 2025, management put specific dates on the site: an initial 24 MW completed and energized by the end of Q1 2026 (claims 4713, 4650, 4652), revenue beginning in May 2026 (claims 4714, 4690, 4651), and 40 MW of cumulative capacity by the end of Q2 2026 (claims 4716, 4653). The last mention on record is 28 July 2025. Every one of those dates has now passed, and each claim was moved to delayed by the ledger's automated check on 3 September 2026, not by any company statement. The record holds no revised schedule, no confirmation that the first 24 MW was energized, and no statement on whether revenue has begun. The disclosures also gave the revenue start as both 1 May and 31 May 2026, and that difference was never reconciled.

The Duke Energy power arrangement behind NC-1 has gone equally quiet. It was described as a staged commitment: 24 MW by 1 September 2025 (claims 4710, 4686, 4644), 40 MW by 1 April 2026 (claims 4711, 4687), and 99 MW by 16 May 2029 (claims 4712, 4688, 4646). The first two milestones have lapsed and were auto-transitioned to delayed on 3 September 2026. The 99 MW figure still stands as stated, but only because its date has not yet arrived. Nothing on record since July 2025 confirms that Duke delivered power at either earlier stage, describes the agreement's current terms, or says whether the later phases still hold.

In Montreal, management said MTL-2 and MTL-3 would be complete and generating revenue by 31 January 2026 (claims 4705, 4707). A disclosure from 28 July 2025 gave MTL-3 an earlier target of 31 December 2025 (claim 4648), and the record never explains which date was intended. Two other entries describe operational timelines for MTL-2 and MTL-3 without any date (claims 4681, 4683). MTL-3 matters beyond its own 7 MW (claim 4706) because it is the stated home of the 5 MW Cerebras colocation agreement (claims 4708, 4684, 4642). The completion claims were auto-flagged as delayed on 3 September 2026, and nothing on record says whether either facility is finished, whether revenue has started, or whether the Cerebras deployment has gone live.

The company's aggregate capacity guidance has drifted out of step with its own milestones. A target of 16 MW in total by the end of 2025 (claims 4715, 4692, 4649) lapsed without comment and was auto-marked delayed. The larger target of 76 MW by the end of Q4 2026 (claims 4717, 4693, 4654) still stands as stated. That figure depends on NC-1 reaching 40 MW and on the Duke milestones, both of which have slipped on the ledger. No disclosure on record reaffirms the 76 MW target, lowers it, or explains how it can be met given the earlier slippage.

Several acquisition entries also carry a delayed flag: Enovum, MTL-2, MTL-3 and the NC-1 property (claims 4638, 4640, 4680, 4682, 4641, 4685, 4643). Their dates refer to transactions that the same disclosures describe as already closed, so the flag appears to be an artifact of the automated check rather than evidence of a walked-back commitment. They are excluded here for that reason.

Every claim on record for WhiteFiber comes from three disclosures issued between July 10 and July 28, 2025, around the company's IPO. Nothing later is on file. The items below are therefore not reversals that management has acknowledged. They are dated commitments whose deadlines have passed with no revision, confirmation or explanation in the record.

The clearest case is NC-1 in Greensboro. Across all three July 2025 documents the company said the site would energize an initial 24 MW by the end of Q1 2026 (claims 4713, 4650, 4652), reach 40 MW cumulative by the end of Q2 2026 (4716, 4653) and begin generating revenue in May 2026 (4714, 4690, 4651). The last mention of any of these dates was July 28, 2025. All of them moved to delayed on September 3, 2026, when their deadlines lapsed without verification. The record contains no statement that the 24 MW came online, that revenue began, or that the schedule was reset. The property purchase itself and a phased 40 MW plan (4691, 4689) were described in present tense, but nothing on record says what the site is actually doing today.

The Duke Energy power commitments that NC-1 depends on follow the same pattern. The company described a 24 MW tranche available by September 1, 2025 (4710, 4686, 4644) and a 40 MW tranche by April 1, 2026 (4711, 4687), both now marked delayed. The long-dated 99 MW tranche for May 2029 (4712, 4688, 4646) is still carried as stated. The disclosures did not say whether either near-term tranche was delivered, deferred by the utility, or renegotiated. Because the 2029 figure keeps being repeated while the earlier milestones are not, the arrangement now reads as an ongoing relationship rather than a schedule being met.

In Montreal, the July 2025 disclosures put completion and revenue start for both the MTL-2 expansion (4705) and the 7 MW MTL-3 site in Saint-Jérôme (4707) at January 31, 2026. One of the same filings gave MTL-3 an earlier end-2025 date (4648), so the timeline was already inconsistent when it was stated. Both dates have since lapsed and are marked delayed. The 5 MW Cerebras colocation agreement at MTL-3 (4708, 4684, 4642) is still listed as stated, but nothing on record says whether that capacity has been delivered to the customer or whether its start date moved along with the facility.

The aggregate guidance shows the cumulative effect. The target of 16 MW total by year-end 2025 (4715, 4692, 4649) is marked delayed, and no year-end figure was reported in its place. The headline target of 76 MW by the end of Q4 2026 (4717, 4693, 4654) remains formally in force and was last repeated on July 28, 2025. With three months left and the NC-1, Duke and Montreal milestones beneath it unconfirmed, nothing on record either reaffirms that target or revises it.

One caution applies to the record itself. Several acquisition claims also show as delayed: Enovum (4638), MTL-2 (4640, 4680), the MTL-3 site (4641, 4682) and the NC-1 purchase (4643, 4685). Their dates fall before the July 2025 disclosures, so they almost certainly describe transactions that had already closed. Those flags reflect how the ledger treats past dates, not anything management has walked back.

Every claim visible in WhiteFiber's record dates from a three-week window around its July 2025 IPO, with the latest dated 2025-07-28. The ledger shows no later disclosures that restate, confirm or revise these commitments. The 'delayed' statuses below were assigned automatically on 2026-09-03 when promised dates passed without a verifying disclosure. They show that management has stayed silent, not that a slip has been confirmed. The record supplied here is also truncated after claim 4638, so a later mention could exist outside this view.

The Greensboro NC-1 ramp is the largest open item. Across the July 2025 filings, management set out a sequence: initial 24 MW completion and energization by the end of Q1 2026 (claims 4713, 4650, 4652), 40 MW cumulative by the end of Q2 2026 (4716, 4653), and revenue starting in May 2026 (4714, 4690, 4651). The revenue date was given as 2026-05-01 in two filings and 2026-05-31 in the third, and that difference was never reconciled. All of these were last mentioned on 2025-07-28 and now sit at delayed. No statement in the record says whether the 24 MW block was energized, whether revenue began, or whether the timetable was reset.

The Duke Energy power phases that underpin NC-1 have also gone unmentioned. The prospectus described 24 MW available by 2025-09-01 (4710, 4686, 4644) and 40 MW by 2026-04-01 (4711, 4687, 4645). Both dates have passed with no confirmation that power was delivered. Only the 99 MW phase, dated 2029 (4712, 4688, 4646), remains at stated, because its date has not yet arrived. The record does not say whether the utility met the earlier dates or whether the construction delay, if there is one, lies on the power side or the build side.

The Montreal expansions MTL-2 (5 MW) and MTL-3 (7 MW) moved even within the IPO window. The 2025-07-28 filing gave completion by 2025-12-31 (4647, 4648), while the 2025-07-10 filing tied completion and revenue start to 2026-01-31 (4705, 4707). The 2025-07-22 operational timelines (4681, 4683) carried no date at all. Both dates have passed, both claims are delayed, and no completion has been announced. The 5 MW Cerebras colocation agreement at MTL-3 (4708, 4684, 4642) was described in each filing, with 2025-07-28 the last mention. The record has nothing on whether Cerebras has taken occupancy or begun paying. The same is true of the unnamed cloud customer's three-year agreement (4718, 4695), which has not been named or described further. The past-dated acquisition entries for MTL-2, MTL-3 and NC-1 (4680, 4682, 4685, 4643, 4641) describe purchases already completed when the prospectus was written, so their delayed flags are likely a classification artifact rather than a sign of slippage.

The capacity guidance has diverged. The year-end 2025 target of 16 MW total (4715, 4692, 4649) is consistent with MTL-1's 4 MW plus the MTL-2 and MTL-3 expansions, and it lapsed with no update. Even so, the larger target of 76 MW by Q4 2026 (4717, 4693, 4654) stays at stated only because its deadline has not passed. That target depends on the same Montreal and Greensboro projects that now show missed interim dates. Management has neither reaffirmed nor reduced the 76 MW figure since 2025-07-28, and the record contains no explanation of how the earlier milestones turned out.

Every WhiteFiber claim on record comes from three filings in July 2025, dated the 10th, 22nd and 28th, around the company's IPO. The ledger holds no later disclosure that revisits any of them. On 2026-09-03, more than a dozen dated commitments moved automatically to 'delayed' because their promised dates passed without a confirming update. None of these transitions came from a company announcement. Several acquisition entries were also flagged, including MTL-2 (4640, 4680), MTL-3 (4641, 4682), NC-1 (4643, 4685) and Enovum (4638). Those appear to be historical closing dates described in the prospectus rather than forward promises, and should not be read as slippage.

The clearest gap is the NC-1 Greensboro build-out. In July 2025 management said NC-1 would reach 24 MW by the end of Q1 2026 (4713, 4650, 4652) and 40 MW by the end of Q2 2026 (4716, 4653). It said revenue would begin in May 2026, though the filings disagreed on whether that meant 1 May (4714, 4690) or 31 May (4651). The last mention of any of these dates is 2025-07-28. All are now past and marked delayed. The record holds no statement that NC-1 energized, that revenue began, or that the schedule was reset.

The Duke Energy power commitments behind NC-1 are equally quiet. The first 24 MW tranche was due by 1 September 2025 (4710, 4686, 4644) and the 40 MW tranche by 1 April 2026 (4711, 4687, 4645). Both dates lapsed without comment on whether the utility delivered. The 99 MW phase dated May 2029 (4712, 4688, 4646) is still described as stated, but nothing since July 2025 confirms that the earlier phases, on which it builds, are in place.

The near-term capacity guidance has also gone unaddressed. The company guided to 16 MW of total capacity by the end of 2025 (4715, 4692, 4649). That year has closed with no disclosed figure confirming or missing the target. The larger 76 MW by end-2026 target (4717, 4693, 4654) still sits at 'stated'. However, it has not been reaffirmed since 2025-07-28, and it depends substantially on the NC-1 and Duke milestones that are now unverified.

In Montreal, the MTL-2 and MTL-3 completion and revenue dates were already inconsistent within July 2025. The 10 July filing put MTL-2 at 31 January 2026 (4705), while the 28 July filing said 31 December 2025 (4647). MTL-3 was also given 31 January 2026 (4707). All are delayed, and nothing explains the discrepancy or reports completion. The MTL-3 date matters because the 5 MW Cerebras colocation agreement is sited there (4708, 4684, 4642). The ledger has no subsequent word on whether Cerebras went live.

The unnamed three-year cloud customer agreement (4718, 4695) was presented in July 2025 as the anchor of the cloud business. It has not been mentioned since, with no disclosure of the customer's identity, ramp or continued standing.

The Greensboro, North Carolina campus known as NC-1 carried the most specific schedule in WhiteFiber's IPO-period disclosures. Filings dated 2025-07-10, 2025-07-22 and 2025-07-28 said the initial 24 MW would be complete or energized by the end of the first quarter of 2026 (claims 4713, 4650, 4652), revenue would begin in May 2026 (4714, 4690, 4651), and cumulative capacity would reach 40 MW by the end of the second quarter of 2026 (4716, 4653). The last mention on record is 2025-07-28. All three dates have passed, and the ledger moved each claim to delayed on 2026-09-03. Nothing on record confirms energization, first revenue, or a revised timetable, and nothing explains why the dates came and went.

The NC-1 schedule rested on a phased power commitment from Duke Energy. The company described 24 MW available by 2025-09-01 (4710, 4686, 4644) and 40 MW by 2026-04-01 (4711, 4687, 4645). Both phase dates have lapsed without a disclosure on record saying the utility delivered, and both are now marked delayed. The third phase, 99 MW by 2029-05-16 (4712, 4688, 4646), is still presented as stated. The record does not say whether the first two phases were energized late, renegotiated, or remain pending. That leaves the longer-dated commitment standing on the same unconfirmed basis.

In Quebec, the July 2025 disclosures said MTL-2 (5 MW) and MTL-3 (7 MW) would be completed and producing revenue by 2026-01-31 (4705, 4707). The same filings described a 5 MW Cerebras colocation agreement at MTL-3 (4708, 4684, 4642). Later operational-timeline references to both sites carry no dates at all (4681, 4683). The January 2026 target has passed and both timeline claims are marked delayed. Nothing on record says whether either site began generating revenue or whether the Cerebras deployment moved from signed agreement to live service.

The company's headline capacity guidance has diverged. The 16 MW year-end 2025 target (4715, 4692, 4649) is marked delayed, and no statement on record reports the capacity actually reached. The 76 MW target for the end of 2026 (4717, 4693, 4654) is still recorded as stated, but it depends substantially on NC-1 reaching 40 MW, which has itself lapsed. The target has not been reaffirmed or revised in any disclosure on record since July 2025.

A caveat applies to the ledger itself. Several acquisitions carry promise dates that refer to transactions already completed when the filings were made: MTL-2 on 2024-12-27, MTL-3 on 2025-04-11, NC-1 on 2025-05-20, and Enovum on 2024-10-01 (4680, 4640, 4682, 4641, 4685, 4643, 4638). Their delayed labels look like an artifact of automated date handling rather than evidence of slippage, so they are set aside here. More broadly, every claim on record dates from July 2025. The pattern above is therefore a long silence after the IPO filings, with no announced walk-backs and no explanations.

The North Carolina build-out is the clearest gap. In July 2025 disclosures, WhiteFiber said the Greensboro site, NC-1, would bring its initial 24 MW online by the end of the first quarter of 2026 (claims 4713, 4650, 4652). It said revenue would start in May 2026 (4714, 4690, 4651), and the site would reach 40 MW by the end of the second quarter of 2026 (4716, 4653). The dates were not even consistent with each other: two claims give May 1 for first revenue, and a third gives May 31. The last mention on record is 2025-07-28. All of these dates have now passed, and the ledger moved each claim to delayed on 2026-09-03 automatically, not because the company announced anything. The record holds no statement that NC-1 was energized, that revenue began, or that the schedule was revised.

The Duke Energy power commitments behind NC-1 show the same pattern. Management described a phased agreement: 24 MW by 2025-09-01 (4710, 4686, 4644), 40 MW by 2026-04-01 (4711, 4687, 4645), and 99 MW by 2029-05-16 (4712, 4688, 4646). The first two dates are well past, and both phases have been auto-marked delayed. The company has not said whether Duke delivered power on those dates, whether the phases were renegotiated, or whether the constraint sits with the utility or with WhiteFiber. The 99 MW figure is still listed as stated, but it has not been reaffirmed since July 2025.

The Montreal expansions have also gone quiet. MTL-2, a 5 MW site in Montreal (4704), and MTL-3, a 7 MW site in Saint-Jérôme (4706), were each said to be complete and producing revenue by 2026-01-31 (4705, 4707). The related timeline claims (4681, 4683) gave no dates. MTL-3 also carries the 5 MW Cerebras colocation agreement (4708, 4684, 4642), which was presented as a signed customer commitment. Neither completion date has been confirmed or revised on record. Nothing has been said about whether Cerebras is live at MTL-3, whether the agreement's timing changed, or whether it is still in place.

Those delays feed directly into the headline capacity guidance. The 16 MW target for the end of 2025 (4715, 4692, 4649) equals MTL-1 plus MTL-2 plus MTL-3 at 4, 5 and 7 MW. It has been auto-marked delayed, with no update or restatement from the company. The larger target of 76 MW by the end of 2026 (4717, 4693, 4654) is still recorded as stated. It depends on the same Montreal and Greensboro projects that have missed their interim dates, and there are three months left in the year. The company has not reaffirmed, trimmed or withdrawn it since July 2025.

Several acquisition claims are also flagged delayed: the Enovum, MTL-2, MTL-3 and NC-1 purchases (4638, 4640, 4641, 4643, 4680, 4682, 4685). Their promised dates fall before the date they were disclosed, so they appear to describe transactions that were already completed. The delayed flag on them looks like an artifact of the automated ledger, not a walk-back, and they are not treated as such here.

The most concrete commitment on record is the NC-1 Greensboro build-out, and none of its dated milestones has been confirmed. In July 2025 disclosures, management said an initial 24 MW would be complete or energized by the end of Q1 2026 (claims 4713, 4650, 4652), that revenue would start in May 2026 (claims 4714, 4690, 4651, which themselves differ on whether the date was May 1 or May 31), and that NC-1 would reach 40 MW cumulatively by the end of Q2 2026 (claims 4716, 4653). The last mention of any of these dates was 2025-07-28. All of them passed and were moved automatically to delayed on 2026-09-03. We have no record of a revised date, a statement that the 24 MW phase was energized, or a first-revenue announcement for the site.

The Duke Energy power commitments behind NC-1 show the same pattern. Across three July 2025 filings, the company described a phased schedule: 24 MW available by 2025-09-01 (claims 4710, 4686, 4644), 40 MW by 2026-04-01 (claims 4711, 4687, 4645), and 99 MW by 2029-05-16 (claims 4712, 4688, 4646). The first two dates have passed and were auto-marked delayed on 2026-09-03. Only the distant 2029 tranche remains in stated status. We have found no confirmation that Duke delivered power on either date and no explanation of whether the utility schedule, the site build, or both moved.

In Montreal, the completion and revenue-start dates for MTL-2 and MTL-3 were already inconsistent within July 2025. The 2025-07-10 disclosure put both at January 31, 2026 (claims 4705, 4707). The 2025-07-28 disclosure put both at year-end 2025 (claims 4647, 4648). The 2025-07-22 descriptions of these timelines carried no date at all (claims 4681, 4683). Every dated version has now lapsed and was auto-marked delayed. This matters most for the Cerebras colocation agreement, which the company described three times as a 5 MW deployment at MTL-3 (claims 4708, 4684, 4642). The agreement itself has never moved out of stated status, but the facility it depends on has no confirmed completion date, and nothing on record says whether Cerebras is live, deferred, or re-sited.

At the portfolio level, the company guided to 16 MW of total capacity by the end of 2025 (claims 4715, 4692, 4649). That guidance lapsed and was auto-marked delayed, with no year-end capacity figure on record to replace it. Meanwhile, the headline target of 76 MW by Q4 2026 (claims 4717, 4693, 4654) remains in stated status. It has not been reaffirmed or revised, even though the NC-1, Duke, and Montreal milestones it relies on have all lapsed. The company has not publicly reconciled that gap.

Several items flagged as delayed should not be read as walk-backs. The MTL-2, MTL-3, and NC-1 property acquisitions (claims 4680, 4682, 4685, 4640, 4641) carry dates that appear to be historical closing dates, not forward promises. Their delayed status is most likely an extraction artifact. More broadly, every claim in this record dates from July 2025, around the IPO. The delayed labels reflect dates that passed without a confirming disclosure in our record, not a company admission that anything slipped.

Every claim on record for WhiteFiber dates from a three-week window in July 2025, around its IPO (claims 4655, 4656). The ledger shows no later company disclosure revisiting any of these commitments. Each 'delayed' status below comes from an automated transition on 2026-09-03 after the promised date passed without confirmation. None reflects a company statement. What follows is what has gone unaddressed, not what is known to have failed.

The Duke Energy power ramp at NC-1 was described in three July 2025 documents as a phased commitment. That meant 24 MW by 1 September 2025 (claims 4710, 4686, 4644) and 40 MW by 1 April 2026 (claims 4711, 4687, 4645). The last mention was 28 July 2025. Both dates have since passed and both claims now sit as delayed. Nothing on record confirms energization at either step or gives a new date. The distant third phase, 99 MW by May 2029 (claims 4712, 4688, 4646), remains in stated language. So the company has kept talking about the long-dated tranche while the near-term tranches have gone quiet.

The NC-1 build-out in Greensboro depended on that power. It was promised as 24 MW complete by the end of Q1 2026 (claims 4713, 4650, 4652), revenue starting in May 2026 (claims 4714, 4690, 4651), and 40 MW cumulative by the end of Q2 2026 (claims 4716, 4653). All are now delayed, with no revision or completion notice after July 2025. Even within July the revenue start drifted from 1 May to 31 May 2026 without comment. The company-wide target of 76 MW by the end of Q4 2026 (claims 4717, 4693, 4654) is still carried as stated. NC-1 was expected to supply more than half of that figure. There has been no statement reconciling the lapsed NC-1 milestones with the unchanged year-end total, which is now just over three months away.

The Montreal expansions, MTL-2 (5 MW) and MTL-3 (7 MW in Saint-Jérôme), had completion and revenue dates that were not consistent even across the July disclosures. The 10 July document gave 31 January 2026 for both (claims 4705, 4707). The 28 July prospectus-era claims gave 31 December 2025 (claims 4647, 4648). The 22 July claims carried operational timelines with no date at all (claims 4681, 4683). Both sets of dates are now delayed, and no completion announcement appears on record. The Cerebras colocation agreement for 5 MW at MTL-3 (claims 4708, 4684, 4642) is still recorded as stated. Nothing has been said about whether service began, even though it depends on a facility whose completion has not been confirmed.

The near-term guidance of 16 MW total capacity by the end of 2025 (claims 4715, 4692, 4649) was repeated in all three July documents and not mentioned again. The date passed nine months ago. The claims auto-transitioned to delayed, and the record contains neither a year-end capacity figure nor an acknowledgment that the target was missed.

Several property acquisitions are also flagged delayed: MTL-2 (claims 4680, 4640), MTL-3 (claim 4682), and NC-1 (claims 4685, 4643). Their promised dates fall before the disclosures themselves, which suggests they describe completed transactions. Those statuses more likely reflect how the extraction handled historical dates than any walk-back. The claim list supplied here was also cut off partway through claim 4641, so later entries may exist that are not reflected above.

WhiteFiber's North Carolina build-out is the most prominent commitment that has gone quiet. Across three disclosures dated 2025-07-10, 2025-07-22 and 2025-07-28, management described NC-1 in Greensboro as reaching an initial 24 MW by the end of the first quarter of 2026 (claims 4713, 4652, 4689) and expanding to 40 MW by the end of the second quarter of 2026 (claims 4716, 4691). Revenue was to begin in May 2026 (claims 4714, 4690, 4651). The filings did not agree on the exact date: two gave 1 May and the last gave 31 May. The last mention on record is 2025-07-28. Every one of those dates has now passed, and the ledger moved the dated claims to delayed on 2026-09-03 by automated check, not in response to anything the company said. Nothing on record says whether NC-1 has been energized, whether any revenue has started, or what the new schedule is.

The Montreal expansions follow the same pattern, and their target moved even within July 2025. The 2025-07-10 and 2025-07-22 disclosures placed completion and revenue start for MTL-2 and MTL-3 at the end of January 2026 (claims 4705, 4707). The 2025-07-28 disclosure gave the end of 2025 instead (claims 4647, 4648). Other entries in that same set describe the operational timelines without any date (claims 4681, 4683). Both facilities are now marked delayed, again by automated transition on 2026-09-03. The Cerebras colocation agreement for 5 MW at MTL-3 (claims 4708, 4684, 4642) depends on that facility being ready. The record contains nothing on whether Cerebras is being served, or on the revised completion dates for either site.

The company's aggregate capacity guidance has been neither confirmed nor withdrawn. Management guided to 16 MW of total capacity by the end of 2025 (claims 4715, 4692, 4649). That deadline passed roughly nine months ago, and the record contains no statement on whether the figure was reached. The larger target of 76 MW by the end of 2026 (claims 4717, 4693, 4654) is still recorded as stated, but it was built on NC-1 contributing 40 MW by mid-2026, and those near-term milestones have lapsed. There has been no reaffirmation of the 76 MW figure and no revision of it since July 2025.

The Duke Energy power arrangement was presented as a phased schedule: 24 MW by 1 September 2025 (claims 4710, 4686, 4644), 40 MW by 1 April 2026 (claims 4711, 4687, 4645) and 99 MW by May 2029 (claims 4712, 4688, 4646). The first two dates have passed without any disclosure confirming delivery, and both are now marked delayed. Because NC-1's timeline depends on these deliveries, the absence of any update leaves open whether the utility schedule or the facility build is the constraint. Only the 2029 phase remains on record as simply stated.

One caution applies to all of the above. Every claim in this ledger comes from disclosures issued in July 2025, around the company's IPO, and no later disclosure has been captured. The delayed statuses therefore reflect elapsed deadlines, not any acknowledgment by management. Several past-dated acquisition entries, such as MTL-2, MTL-3, NC-1 and Enovum (claims 4680, 4682, 4685, 4638), were flagged delayed mechanically even though they describe transactions that had already closed. They are not treated here as slippage.

Every claim on record for WhiteFiber comes from three disclosures issued between 2025-07-10 and 2025-07-28, around the IPO. Nothing later appears in the file. On 2026-09-03 the ledger automatically moved a cluster of dated commitments to delayed because their deadlines had passed. None of these moves reflects a company statement. The pattern that matters is what the company set out in detail in July 2025 and has not addressed on the record since.

The NC-1 Greensboro schedule is the most specific of these. All three July 2025 filings repeated the same plan: an initial 24 MW energized by the end of Q1 2026 (claims 4713, 4650, 4652), revenue starting around May 2026 (4714, 4690, 4651), and 40 MW cumulative by the end of Q2 2026 (4716, 4653). The revenue date is given as May 1 in two filings and May 31 in the third, a small inconsistency that was never reconciled. All of these dates have now passed, and each claim is marked delayed. There is no disclosure on record confirming energization, first revenue or a revised timetable, and no explanation for why these milestones are no longer discussed.

The NC-1 plan rested on a staged Duke Energy power commitment: 24 MW by 2025-09-01 (4710, 4686, 4644), 40 MW by 2026-04-01 (4711, 4687, 4645), and 99 MW by 2029-05-16 (4712, 4688, 4646). The first two dates have lapsed and are marked delayed. The 99 MW figure remains stated and unchanged in every mention. The record contains no statement on whether Duke delivered any of the earlier tranches, whether the agreement was amended, or whether the utility schedule, rather than construction, explains the NC-1 slip.

In Montreal, the 2025-07-10 disclosure gave MTL-2 and MTL-3 a specific completion and revenue-start date of 2026-01-31 (4705, 4707). Twelve days later the equivalent claims described only an "operational timeline" with no date attached (4681, 4683). No statement accompanied the change. The Cerebras 5 MW colocation agreement at MTL-3 (4708, 4684, 4642) is still recorded as stated. However, nothing on record says whether that customer went live on the original schedule.

The capacity guidance has diverged from these building blocks. The company guided to 16 MW by the end of 2025 (4715, 4692, 4649), which matches exactly the sum of MTL-1 (4 MW), MTL-2 (5 MW) and MTL-3 (7 MW). That target has passed without a reported result. The 76 MW target for the end of 2026 (4717, 4693, 4654) is still recorded as stated even though its largest component, NC-1's 40 MW, is past due. Management has not said on the record whether 76 MW is still the target or how the missed interim milestones affect it.

Several acquisition claims are also marked delayed, including Enovum, MTL-2, MTL-3 and the NC-1 property (4680, 4682, 4685, 4638, 4640, 4641, 4643). Those dates refer to transactions the filings described as already completed. The delayed label on them looks like a classification artifact rather than evidence of a slip.

Every WhiteFiber disclosure on record dates from a three-week window in July 2025, around the IPO (July 10, July 22 and July 28). Nothing in the ledger records management returning to these commitments since then. The "delayed" flags below were applied automatically on 2026-09-03 because the promised dates passed without any confirming disclosure. They mark an absence of confirmation, not a documented failure.

The NC-1 Greensboro build-out is the most concrete unconfirmed timeline. The July 2025 filings promised an initial 24 MW energized by the end of Q1 2026 (claims 4713, 4652), revenue starting in May 2026 (4714, 4690, 4651) and 40 MW cumulative by the end of Q2 2026 (4716, 4653, 4691). The documents did not agree on the revenue date: 4651 gives May 31, while 4714 and 4690 give May 1. The last mention was 2025-07-28. All three dates have passed, and the record holds no statement that the site was energized, that revenue began, or that the schedule moved.

The Duke Energy power commitments behind NC-1 are similarly unconfirmed. Management described a phased agreement: 24 MW by 2025-09-01 (4710, 4686, 4644), 40 MW by 2026-04-01 (4711, 4687, 4645) and 99 MW by 2029-05-16 (4712, 4688, 4646). The two near-term phases are now flagged delayed. The 99 MW phase remains "stated" only because its date has not arrived. Nothing on record says whether the first phases were delivered, renegotiated or deferred, or whether they still support the later tranche. The last mention was 2025-07-28.

The Montreal expansions, MTL-2 (5 MW, claim 4704) and MTL-3 in Saint-Jérôme (7 MW, claim 4706), were due to complete and begin generating revenue by January 2026 (4705, 4707). The July 28 filing gave an earlier MTL-3 date of year-end 2025 (4648), while the July 22 timelines (4681, 4683) carry no date at all. MTL-3 is where the 5 MW Cerebras colocation agreement sits (4708, 4684, 4642). No subsequent disclosure on record confirms completion of either facility or the start of Cerebras revenue.

The headline capacity guidance has not been revisited. The target of 16 MW total by the end of 2025 (4715, 4692, 4649) passed without confirmation. The target of 76 MW by the end of Q4 2026 (4717, 4693, 4654) is still carried as "stated." Its main building blocks, the NC-1 phases and the Montreal expansions, are themselves unconfirmed, and no revised figure or reaffirmation appears on record. The three-year cloud customer agreement (4718, 4695) has likewise gone unmentioned since July 2025.

Several acquisition claims also carry the delayed flag: Enovum (4638), MTL-2 (4640, 4680), MTL-3 (4641, 4682) and NC-1 (4643, 4685). Their dates fall before the disclosures that described them, so these appear to be completed transactions caught by the automated rule rather than walked-back promises.

Every claim on file for WhiteFiber comes from three disclosures dated 10, 22 and 28 July 2025, around the IPO. Nothing later appears in this record. The status changes logged on 2 and 3 September 2026 were made automatically when promised dates passed. None of them follows a company statement. The silences below should be read with that in mind: the company may have addressed some of these points in materials not captured here, but none of those statements appears in the record.

The nearest-term capacity target was 16 MW of total capacity by the end of 2025 (claims 4715, 4692, 4649). It was last repeated on 28 July 2025 and is now marked delayed, the deadline having passed nine months ago. The record contains no confirmation that the figure was reached, no revised date and no restated number.

The Montreal expansions, MTL-2 and MTL-3, were given two different completion dates within the same month. The 10 July filing said January 2026 for completion and first revenue (claims 4705, 4707). The 28 July filing said the end of 2025 (claims 4647, 4648). The 22 July filing described an operational timeline with no date at all (claims 4681, 4683). The date moved in both directions and was then dropped, and it was never reconciled. Both facilities are now marked delayed, with no statement on whether either is energized or earning revenue.

NC-1 in Greensboro was the centerpiece of the growth story. Management promised 24 MW by the end of the first quarter of 2026 (claims 4713, 4650, 4652), first revenue in May 2026 (claims 4714 and 4690 say 1 May; claim 4651 says 31 May), and 40 MW cumulative by the end of the second quarter (claims 4716, 4653). All three milestones are now marked delayed. The last mention of each was 28 July 2025. The record contains no energization announcement, no revenue-start confirmation and no revised schedule.

The Duke Energy power arrangement was described as a phased commitment. The phases were 24 MW by 1 September 2025 (claims 4710, 4686, 4644), 40 MW by 1 April 2026 (claims 4711, 4687, 4645) and 99 MW by May 2029 (claims 4712, 4688, 4646). The first two phases are overdue, and no disclosure says whether that power was delivered. Meanwhile the headline target of 76 MW by the end of 2026 is still carried as stated (claims 4717, 4693, 4654). The top of the plan remains standing while every intermediate step toward it has lapsed without comment. Nothing on file explains how 76 MW stays achievable if the 24 MW and 40 MW steps have not happened.

One caution about the ledger itself. Several acquisitions are also marked delayed: Enovum (claim 4638), the MTL-2 property (claims 4680, 4640), the MTL-3 site (claims 4682, 4641) and the NC-1 property (claims 4685, 4643). Their promised-by dates are earlier than the July 2025 disclosures, so these most likely describe transactions already completed when they were reported. They are probably extraction artifacts, not walked-back commitments.

The 16 MW year-end 2025 capacity target is the cleanest case. It was stated three times in eighteen days — claims 4715 (July 10), 4692 (July 22) and 4649 (July 28) — each time as a firm figure attached to a firm date of December 31, 2025. All three moved to 'delayed' on September 3, 2026, and the transition ledger attributes that move to automated extraction rather than to any company revision. Nine months past the date, the record contains no restated number, no partial-completion figure, and no explanation of which facility fell short. The last time management put a capacity figure on 2025 was the July 28, 2025 prospectus-window disclosure; nothing since has amended or withdrawn it.

The Duke Energy arrangement has quietly bifurcated. Phase 1, 24 MW by September 1, 2025 (claims 4710, 4686, 4644), and Phase 2, 40 MW by April 1, 2026 (claims 4711, 4687, 4645), are both now carried as delayed on date-lapse alone. Phase 3, 99 MW by May 16, 2029 (claims 4712, 4688, 4646), remains 'stated' — untouched only because its date is four years out. The language also softened across the three filings: what was described as a 'capacity commitment' on July 10 became 'Duke Energy — 24 MW' on July 22 and a 'capacity agreement — Phase 1' on July 28. What has not been said is whether the September 2025 energization actually occurred, whether the Duke interconnection queue position was preserved when it did not, and whether Phase 2 and Phase 3 dates shift with it.

NC-1 in Greensboro shows a revision made without an announcement. The revenue start was given as May 1, 2026 in claims 4714 and 4690; eighteen days later, in claim 4651 of July 28, the same milestone is dated May 31, 2026. A one-month slip was absorbed into a subsequent document rather than disclosed as a change. The underlying construction milestones — 24 MW complete by March 31, 2026 (claims 4713, 4650, 4652) and 40 MW cumulative by June 30, 2026 (claims 4716, 4653) — have both since lapsed into 'delayed'. The last affirmative statement on NC-1 remains July 28, 2025. Nothing on record addresses the site's current energized capacity or whether it has begun generating revenue at all.

The Montreal timelines lost their dates rather than missing them. On July 10, MTL-2 and MTL-3 completion and revenue start were both pinned to January 31, 2026 (claims 4705 and 4707, now delayed). By July 22, the corresponding 'operational timeline' claims, 4681 and 4683, carry no promised-by date at all and consequently still sit at 'stated' — a milestone that becomes unfalsifiable by having its deadline removed. Cerebras sits inside that ambiguity: its 5 MW colocation agreement was a 'partnership' in claim 4708 and a 'customer' in claims 4684 and 4642, tied specifically to MTL-3, a facility whose own completion date has passed without comment. Whether that contract is in force, repriced, or relocated has not been addressed.

Two smaller items have simply stopped appearing. The three-year cloud service agreement described in claim 4718 and the 'first cloud customer' of claim 4695 were named on July 10 and July 22 and do not recur in the July 28 disclosures; no contract value, term status, or renewal has been mentioned since. Bit Digital's 80.0 percent post-IPO ownership, specified in claim 4719 on July 10, appears on July 22 only as an unquantified 'Bit Digital contribution' (claim 4694). Standing above all of it, the 76 MW end-2026 target (claims 4717, 4693, 4654) is still carried as 'stated' despite resting on the 16 MW and 40 MW rungs that have already slipped. It has not been reaffirmed and it has not been withdrawn; with roughly a quarter left to run, the silence is the disclosure.

The disclosure record for WhiteFiber ends on 2025-07-28. Everything on file was extracted from a three-week window around the IPO, and every status change since — the batch to 'stated' on 2026-09-02 and the batch to 'delayed' on 2026-09-03 — was generated automatically by date arithmetic, not by anything management said. Roughly fourteen months have passed since the last dated claim. The items below are not contradicted by later statements; they are uncontradicted because there have been none.

The Duke Energy phasing is the clearest case. Phase 1, 24 MW, was promised by 2025-09-01 (claims 4710, 4686, 4644). That date is now more than a year past and the claims auto-transitioned to delayed on 2026-09-03. Phase 2, 40 MW by 2026-04-01 (claims 4711, 4687, 4645), has likewise lapsed. Phase 3, the 99 MW commitment carrying a 2029-05-16 horizon (claims 4712, 4688, 4646), remains at 'stated' purely because its date has not arrived — it sits on top of two phases that have. Nothing on record explains interconnection timing, revised energization dates, or whether the phased structure with Duke still holds in its original form. Last mention of any of it: 2025-07-28.

The capacity guidance has not been re-set. Sixteen MW by 2025-12-31 was stated three separate times (claims 4715, 4692, 4649) and lapsed without comment. The 76 MW figure for end-2026 (claims 4717, 4693, 4654) still stands at 'stated' and is due in roughly three months, but it was built on the 16 MW base and on NC-1 reaching 24 MW by 2026-03-31 (claims 4713, 4650, 4652) and 40 MW by 2026-06-30 (claims 4716, 4653) — all four of which are now flagged delayed. A downward revision to the 76 MW target has not been announced; neither has a reaffirmation.

Two dates moved quietly inside the July 2025 window itself. NC-1 revenue start was given as 2026-05-01 on 2025-07-10 and 2025-07-22 (claims 4714, 4690), then as 2026-05-31 on 2025-07-28 (claim 4651) — a month added eighteen days later with no note that anything had changed. MTL-3 completion moved the other way, from 2026-01-31 (claim 4707) to 2025-12-31 (claim 4648), without reconciliation. Neither adjustment was presented as an adjustment.

The Bit Digital relationship was quantified once and then was not. Claim 4719, dated 2025-07-10, states 80 percent post-IPO ownership. It does not recur. The 2025-07-22 set describes a 'Bit Digital contribution' with no figure attached (claim 4694), and the 2025-07-28 set gives only the offering mechanics — 7,812,500 shares and a 1,171,875-share over-allotment (claims 4655, 4656). The ownership percentage after the over-allotment, and whether the parent intends to retain that stake, has not been restated since 2025-07-10.

Cerebras drifted in category. On 2025-07-10 the 5 MW arrangement was recorded as a partnership (claim 4708); on 2025-07-22 and 2025-07-28 the same 5 MW at MTL-3 appears as a customer colocation agreement (claims 4684, 4642). The economics may be identical and the reframing may be nothing more than precision, but it was never described as a change. The same applies to the three-year cloud services agreement (claims 4718, 4695), which has had no follow-up — no confirmation that it commenced, no revenue attached to it — since 2025-07-28.

One qualification, for fairness. Several acquisition claims now carrying a delayed flag — Enovum at 2024-10-01 (claim 4638), MTL-2 at 2024-12-27 (claims 4640, 4680), the MTL-3 site at 2025-04-11 (claim 4682), NC-1 Greensboro at 2025-05-20 (claims 4685, 4643) — are dated retroactively to transactions described in the same filings as already completed. Their status is an artifact of how the dates were captured, not evidence that the closings slipped. They should not be read alongside the guidance and Duke Energy items above.

The Duke Energy capacity agreement is the clearest case. Its first phase, 24 MW, was presented across all three July 2025 disclosures with a delivery date of 2025-09-01 (claims 4710, 4686, 4644). That date is now more than a year in the past. The only thing that has happened to the claim is a system-generated move to delayed on 2026-09-03, applied in the same batch that touched every other dated promise on the record. There is no company statement anywhere in the ledger revising the energization date, explaining an interconnection or permitting cause, or restating what the 24 MW now depends on. The same pattern covers Phase 2, 40 MW by 2026-04-01 (claims 4711, 4687, 4645), which also lapsed without comment. Phase 3, the 99 MW commitment dated 2029-05-16 (claims 4712, 4688, 4646), remains in stated status and is the only Duke figure management can still repeat without immediately contradicting itself.

The capacity guidance ladder has come apart from the bottom without being rebuilt at the top. The 16 MW by end-2025 target (claims 4715, 4692) is delayed. The 40 MW at NC-1 by end of Q2 2026 (claims 4716, 4653) is delayed. NC-1 initial 24 MW completion by 2026-03-31 (claims 4713, 4650, 4652) is delayed, and NC-1 revenue start (claims 4714, 4690, 4651) is delayed. Yet the 76 MW total capacity target for Q4 2026 (claims 4717, 4693, 4654) still sits in stated status, untouched, with a deadline roughly three months out. Every input to that number has slipped and the output has not been revised, withdrawn, or reaffirmed with a new bridge. What has not been said is how 76 MW survives the failure of the 16 MW and 40 MW steps beneath it.

There is unexplained drift in the NC-1 and Montreal dates inside the July 2025 window itself. NC-1 revenue generation was given as 2026-05-01 on 2025-07-10 and 2025-07-22 (claims 4714, 4690) and as 2026-05-31 on 2025-07-28 (claim 4651), a quiet one-month push within eighteen days and without a revision note. MTL-2 and MTL-3 move the other way: completion and revenue start were dated 2026-01-31 on 2025-07-10 (claims 4705, 4707), then 2025-12-31 in the 2025-07-28 filing (claims 4647, 4648), while the 2025-07-22 disclosure carried MTL-2 and MTL-3 operational timelines with no date attached at all (claims 4681, 4683). Dropping the date in the middle disclosure and changing it on either side was never characterized as a change. All of these are now delayed.

The Cerebras arrangement changed shape and then disappeared. On 2025-07-10 it was disclosed as a partnership, a 5 MW colocation agreement (claim 4708). On 2025-07-22 the same 5 MW arrangement was recorded as a customer relationship (claim 4684). It does not appear in the 2025-07-28 set. The first cloud customer three-year service agreement follows the same arc, present on 2025-07-10 and 2025-07-22 (claims 4718, 4695) and absent afterward. Neither has been mentioned since 2025-07-22. Nothing has been said about whether the Cerebras megawatts are contracted, where they sit in the 76 MW total, or whether the three-year cloud agreement is still running. Similarly, the Bit Digital post-IPO ownership figure of 80 percent (claim 4719), stated on 2025-07-10 and echoed as a contribution on 2025-07-22 (claim 4694), was not restated on 2025-07-28, where the financing disclosures reverted to raw share counts (claims 4655, 4656).

The broader omission is the silence itself. Every claim on the record originates in the three-week IPO window of 2025-07-10 to 2025-07-28. The last mention date for the entire disclosed history is 2025-07-28. In the roughly fourteen months since, the acquisition claims backdated to 2024-12-27 for MTL-2 (claims 4680, 4640), 2025-04-11 for the MTL-3 site (claims 4682, 4641) and 2025-05-20 for the NC-1 property (claims 4685, 4643) all moved to delayed, as did the Enovum acquisition dated 2024-10-01 (claim 4638). Not one of those transitions was triggered by a company update; all carry the same automated extraction source and the same 2026-09-03 timestamp. One smaller change is worth noting for the record: the CEO was named as Sam Tabar on 2025-07-10 (claim 4701) and identified only by title in both subsequent disclosures (claims 4696, 4657). The available claim set is truncated after claim 4649 in the 2025-07-28 group, so this reading covers the record as provided rather than the complete filing history.

Every claim on record for WhiteFiber dates from a three-week window in July 2025 (2025-07-10, 2025-07-22 and 2025-07-28), which corresponds to the IPO filing period. None of the subsequent quarters has added a disclosure that confirms, revises or withdraws those commitments. The 'delayed' statuses below were assigned automatically on 2026-09-03, when the promised dates passed without a verifying update. They record silence, not a confirmed miss.

The NC-1 Greensboro build-out is the largest open item. Management described an initial 24 MW completion and energization by the end of Q1 2026 (claims 4713, 4650, 4652), 40 MW of cumulative capacity by the end of Q2 2026 (4716, 4653), and revenue starting in May 2026 (4714, 4690, with 4651 giving 2026-05-31). The last mention was 2025-07-28. All of these dates have now passed, and each claim sits at delayed. Nothing on record says whether the 24 MW is energized, whether revenue has begun, or whether the phasing was reset. The 24 MW (4689) and 40 MW (4691) capacity figures themselves remain on record as unrevised.

The Duke Energy power commitments that underpin NC-1 have gone the same way. The first tranche of 24 MW by 2025-09-01 (4710, 4686, 4644) and the second of 40 MW by 2026-04-01 (4711, 4687, 4645) are both past due and auto-marked delayed. The long-dated 99 MW tranche for 2029-05-16 (4712, 4688, 4646) is still carried as stated. The July 2025 filings described these as agreements with specific delivery dates. There has been no subsequent statement on whether Duke delivered power on those dates, whether the schedule was renegotiated, or whether the 99 MW figure still stands.

In Montreal, the MTL-2 and MTL-3 completion and revenue dates have lapsed without comment. MTL-2 was given as complete by 2025-12-31 in one filing (4647) and by 2026-01-31 in another (4705), within weeks of each other. MTL-3 was given as 2026-01-31 (4707). Both are now delayed. The undated operational-timeline claims (4681, 4683) were never given a firmer date. The Cerebras 5 MW colocation agreement at MTL-3 (4708, 4684, 4642) remains carried as stated, but it depends on a facility whose completion has not been confirmed. Nothing has been said about whether Cerebras is live or whether the agreement's timing has changed.

The headline capacity guidance has drifted in a way that has not been addressed. The 16 MW year-end 2025 target (4715, 4692, 4649) passed without any confirmation that it was reached, and it is now delayed. The 76 MW target for the end of Q4 2026 (4717, 4693, 4654) remains stated and has never been revised. Yet the NC-1 and Montreal milestones it was built on have each slipped without comment. No statement reconciles the two.

One caution applies to the ledger itself. Several acquisition claims carry promise dates that precede the filings they appear in: Enovum (4638), MTL-2 (4680, 4640), MTL-3 (4682, 4641) and the NC-1 property (4685, 4643). These most likely describe transactions that had already closed, which the automated extraction read as forward-looking commitments. Their 'delayed' status should not be read as a walk-back.

Every claim on record for WhiteFiber dates from a three-week window in July 2025, spanning 10 July to 28 July around the IPO. Nothing in the ledger has been added since. The statuses below changed through automated transitions on 3 September 2026, when promised dates passed without any verifying disclosure. They do not reflect any statement from the company. The pattern that matters is the silence. A detailed build schedule was set out once and has not been revisited on the record.

The NC-1 Greensboro schedule is the largest gap. The July 2025 materials committed to 24 MW of initial capacity by the end of Q1 2026 (claims 4713, 4650, 4652) and 40 MW cumulative by the end of Q2 2026 (4716, 4653, with the 40 MW phasing restated in 4691). Revenue was to begin in May 2026. Claims 4714 and 4690 give 1 May, while 4651 gives 31 May, so the date had already shifted slightly between filings less than three weeks apart. The last mention of any of these was 28 July 2025. All of them are now marked delayed. The record has no announcement of energization, first revenue, a revised date, or a reason for the slip.

The Duke Energy power commitments behind NC-1 have also gone quiet. The company described 24 MW of utility capacity available from 1 September 2025 (4710, 4686, 4644) and 40 MW from 1 April 2026 (4711, 4687, 4645). Both dates have passed with no confirmation that power was delivered, and both claims are now delayed. Only the long-dated 99 MW commitment for May 2029 (4712, 4688) remains in stated status, and the record says nothing further about it. It is unclear whether the utility timeline slipped, the site build slipped, or both, because neither has been addressed.

The Montreal expansions, MTL-2 at 5 MW and MTL-3 at 7 MW, were given completion and revenue-start dates of 31 January 2026 on 10 July 2025 (4705, 4707). On 28 July 2025 those dates were restated as 31 December 2025 (4647, 4648). The generic operational-timeline claims from 22 July (4681, 4683) gave no date at all. Both dated versions are now delayed. The 5 MW Cerebras colocation agreement depends on MTL-3 (4708, 4684, 4642). Its language actually firmed between filings, moving from partnership to customer. Even so, nothing on record confirms that Cerebras went live or that its start moved with the facility.

The aggregate capacity guidance has not been reconciled with these slips. The target of 16 MW by the end of 2025 (4715, 4692, 4649) is now delayed, and no restated figure has been given. The target of 76 MW by the end of Q4 2026 (4717, 4693, 4654) still stands as stated. It was built on NC-1 reaching 40 MW and the Montreal sites coming online, and none of those components is verified. With roughly three and a half months left, the target has been neither reaffirmed nor withdrawn. The three-year agreement with an unnamed first cloud customer (4718, 4695) has likewise not been followed up with a name, a revenue contribution, or any later reference.

Every claim on record for WhiteFiber dates from a three-week window in July 2025, around the IPO, and the ledger holds no later disclosure against which those statements were confirmed, revised or withdrawn. Each 'delayed' status below was assigned automatically on 2026-09-03 because a promised date passed without verification, not because the company announced a slip. Several 'delayed' flags are on acquisitions whose dates had already passed when they were described: Enovum (4638), MTL-2 (4680), MTL-3 (4682, 4641) and the NC-1 property (4685, 4643). These read as completed transactions rather than missed promises and are set aside here. The ledger as supplied also appears truncated after claim 4646.

The clearest silence concerns the NC-1 Greensboro build-out. Filings dated 2025-07-10, 2025-07-22 and 2025-07-28 committed to 24 MW at NC-1 by the end of Q1 2026 (4713, 4650, 4652), revenue beginning in May 2026 (4714, 4690, 4651) and 40 MW cumulative by the end of Q2 2026 (4716, 4653). The last mention was 2025-07-28, and the filings did not agree on the revenue date: May 1 in two, May 31 in the third. All three milestones are now past and marked delayed. Nothing on record says whether the first 24 MW was energized, whether revenue has begun, or whether the schedule was reset.

The Duke Energy power arrangements that support NC-1 follow the same pattern. Management described a phased commitment of 24 MW by 2025-09-01 (4710, 4686, 4644), 40 MW by 2026-04-01 (4711, 4687, 4645) and 99 MW by 2029-05-16 (4712, 4688). The last mention was 2025-07-28. The first two phases are marked delayed, and the 99 MW figure is still carried as stated without reaffirmation. The company has not said whether Duke delivered power on either date, whether the phases were renegotiated, or whether the utility commitment remains in the form first described.

In Montreal, the MTL-2 (5 MW) and MTL-3 Saint-Jérôme (7 MW) completion dates were inconsistent from the start. The 2025-07-10 disclosure gave completion and revenue start by January 31, 2026 (4705, 4707). The 2025-07-28 disclosure gave December 31, 2025 (4647, 4648). The 2025-07-22 claims described operational timelines with no date at all (4681, 4683). Both sets of dates have passed and are marked delayed, and there has been no subsequent statement. This matters for the 5 MW Cerebras colocation agreement at MTL-3 (4708, 4684, 4642), which the IPO-period filings presented as signed. The record contains nothing on whether Cerebras is now live at that site or whether the arrangement is still pending.

The aggregate capacity guidance has gone unaddressed as well. The target of 16 MW by the end of 2025 (4715, 4692, 4649) roughly matches MTL-1, MTL-2 and MTL-3 combined, so it depended on the Montreal completions above. It is now marked delayed, and no year-end figure was reported. The 76 MW target for Q4 2026 (4717, 4693, 4654) is still carried as stated, but it depended on the delayed NC-1 phases. Management has not reaffirmed or revised it since July 2025, and it has not identified what capacity beyond NC-1 and Montreal would make up the full 76 MW. Similarly, the unnamed cloud customer on a three-year service agreement (4718, 4695) was cited twice in July 2025 and has not appeared in any later disclosure on record.

The Greensboro, North Carolina campus (NC-1) carries the most detailed schedule that has gone quiet. In July 2025, management said the site's initial 24 MW would be complete and energized by the end of Q1 2026 (claims 4713, 4650, 4652). It said capacity would reach 40 MW cumulatively by the end of Q2 2026 (4716, 4653), and that revenue would begin in May 2026 (4714, 4690, 4651). Even at the time, the filings disagreed on the revenue date: two claims say 1 May and one says 31 May. The last mention on record for any of these milestones is 28 July 2025. All of the dates have now passed, and each was moved to delayed by automated review on 3 September 2026, not by any company revision. Nothing on record says whether the 24 MW was energized, whether revenue has started, or whether the 40 MW target still stands.

The NC-1 dates rested on a set of Duke Energy power commitments, and those have gone unmentioned too. The company said Duke would deliver 24 MW by 1 September 2025 (4710, 4686, 4644) and 40 MW by 1 April 2026 (4711, 4687). Both dates have passed and both claims were automatically marked delayed. Meanwhile the long-dated 99 MW Phase 3 commitment for May 2029 (4712, 4688, 4646) is still carried as stated. The record does not show whether the first two tranches arrived late, arrived on time without comment, or were renegotiated. That leaves the 2029 figure without any interim confirmation.

The Montreal expansions show a smaller version of the same pattern. On 10 July 2025, MTL-2 (5 MW) and MTL-3 (7 MW) were each given completion and revenue-start dates of 31 January 2026 (4705, 4707). Eighteen days later, a filing described MTL-3's completion as due by the end of 2025 (4648). The record contains no explanation for the difference and no later update. Both facilities are now marked delayed. The 5 MW Cerebras colocation agreement tied to MTL-3 (4708, 4684, 4642) is still described only in its original signed form. Nothing has been said about whether that customer's deployment began on the schedule implied by the site's completion date.

The aggregate capacity path has two parts that no longer fit together. The near-term goal of 16 MW by the end of 2025 (4715, 4692, 4649) has passed without a restatement and is marked delayed. The goal of 76 MW by Q4 2026 (4717, 4693, 4654) is still carried unchanged as stated, even though the NC-1 and Montreal milestones it depended on have slipped. About three and a half months remain before that deadline. Management has not reaffirmed or revised the 76 MW figure since 28 July 2025, and has not said how the missed intermediate steps affect it.

For completeness, several acquisition claims are also marked delayed only because their reference dates are in the past: MTL-2, MTL-3, NC-1 and Enovum (4680, 4682, 4685, 4638, 4640, 4641, 4643). These appear to describe transactions that were already completed when disclosed. They should not be read as walked-back commitments. More broadly, every claim available here comes from disclosures dated 10 to 28 July 2025. The silence described above therefore covers roughly fourteen months, measured against the ledger as provided.

Every claim on record for WhiteFiber dates from a three-week window around its IPO (2025-07-10, 2025-07-22 and 2025-07-28), and nothing later appears in the ledger. All status changes were automated on 2026-09-02 and 2026-09-03, when promised dates passed without a recorded update. The silence described below is therefore the absence of any follow-up disclosure in our record. It may partly reflect gaps in what has been ingested, and the ledger supplied here was cut off at claim 4645.

The clearest case is the NC-1 build in Greensboro. Management said the initial 24 MW would be complete and energized by 2026-03-31 (claims 4713, 4650, 4652), that revenue would start at the site (claims 4714, 4690, 4651), and that cumulative capacity would reach 40 MW by 2026-06-30 (claims 4716, 4653). These were last mentioned on 2025-07-28 and are now marked delayed. Even within July 2025 the revenue date was inconsistent: 2026-05-01 in two filings and 2026-05-31 in the third. That discrepancy was never reconciled. No statement on record confirms energization, revises the schedule, or explains what held it back.

The Duke Energy power commitments behind NC-1 have gone quiet in the same way. The 24 MW tranche was due by 2025-09-01 (claims 4710, 4686, 4644) and the 40 MW tranche by 2026-04-01 (claims 4711, 4687). Both were last referenced on 2025-07-28 and both are now delayed. The 99 MW commitment dated 2029-05-16 (claims 4712, 4688, 4646) is still carried as stated, but it rests on the same relationship. Nothing on record says whether the earlier tranches were delivered late, renegotiated, or remain outstanding.

In Quebec, management put completion and revenue start for MTL-2 and MTL-3 at 2026-01-31 (claims 4705, 4707). Weeks later it gave 2025-12-31 (claims 4647, 4648), and one filing gave no date at all (claims 4681, 4683). Those two sites were the building blocks of the 16 MW year-end 2025 target, which was repeated three times (claims 4715, 4692, 4649) and is now delayed. The Cerebras 5 MW colocation agreement was tied specifically to MTL-3 (claims 4708, 4684, 4642) and was last mentioned on 2025-07-28. No later disclosure addresses whether that customer went live on the planned schedule.

The 76 MW end-of-2026 capacity target (claims 4717, 4693, 4654) remains formally stated because its deadline has not arrived. However, nearly every interim milestone it depends on is now past due without comment, and there is no record of the target being reaffirmed or revised since July 2025. Several acquisition items (claims 4680, 4682, 4685, 4638) are also flagged delayed, but their dates precede the disclosures that cite them. These appear to describe completed transactions rather than missed promises and should not be read as walk-backs.

WhiteFiber's 2025 capacity guidance target has silently missed. Claims 4715 and 4692 promised 16 MW total capacity by December 31, 2025. The deadline passed eight months ago; both claims auto-transitioned to 'delayed' status on September 3, 2026, with no revision statement or delay announcement from management. The company made the same 16 MW year-end 2025 commitment repeatedly (July 10, 22, and 28, 2025) across three separate filing dates but never publicly adjusted or explained when that target was not met. The NC-1 Greensboro facility—central to WhiteFiber's growth narrative—promised initial 24 MW completion by Q1 2026 (claims 4713, 4650, 4652: promised by March 31, 2026) and revenue generation by May 2026 (claims 4714, 4690: May 1; claim 4651: May 31). All six timelines have auto-delayed after the deadlines passed; as of September 2026, no management communication has been recorded explaining why the facility missed Q1, Q2, or mid-year 2026 targets. Similarly, both MTL-2 Montreal (5 MW expansion) and MTL-3 Saint-Jérôme (7 MW) facilities promised operational and revenue-start dates of January 31, 2026 (claims 4705, 4707). Both have been marked delayed since September 3; neither facility was discussed in subsequent filings to clarify revised timelines. The Duke Energy partnership framework, repeatedly announced with phased commitments, shows Phase 1 (24 MW) promised September 1, 2025 and Phase 2 (40 MW) promised April 1, 2026 (claims 4710, 4711, 4687, 4686, 4644). Both dates have passed; status changed to delayed without revision guidance. The Enovum acquisition (claims 4638, 4703, 4678) was promised by October 1, 2024, and remains delayed as of September 2026—now overdue by nearly two years with no update statement on close timing or abandonment.

WhiteFiber's North Carolina expansion has vanished from active discussion despite being positioned as the company's major 2026 growth initiative. In July 2025, management promised the NC-1 Greensboro property acquisition by May 2025 (claim 4685), initial 24 MW energization by March 2026 (claims 4713, 4650), 40 MW cumulative capacity by June 2026 (claims 4716, 4653), and revenue generation starting May 2026 (claims 4714, 4690, 4651). All four milestones have now transitioned to delayed status. No revised timeline has been disclosed, and the facility is not mentioned in current investor materials, suggesting the project has effectively stalled or been de-prioritized without formal acknowledgment.

The Montreal capacity buildout (MTL-2 and MTL-3) promised a sequential acquisition and operational ramp that has slipped by up to two years. The Enovum acquisition was promised by October 2024 (claims 4703, 4638)—now nearly two years overdue and transitioned to delayed. MTL-2 was promised by December 2024 (claims 4680, 4640) and MTL-3 by April 2025 (claims 4682, 4641), both now severely delayed. Operational timelines for both facilities were promised by January 2026 and December 2025 respectively (claims 4705, 4707, 4681, 4683, 4648) and have also transitioned to delayed. The entire Montreal expansion strategy has quietly stalled without management providing an updated master schedule or revised guidance.

The Duke Energy partnership phases have not materialized as promised, yet the relationship remains marked as ongoing without revision. Phase 1's 24 MW capacity commitment was due September 2025 (claims 4710, 4644)—now over a year overdue and transitioned to delayed. Phase 2's 40 MW was due April 2026 (claims 4711, 4645)—now five months past deadline and delayed. Phase 3's 99 MW framework remains marked as stated through 2029 (claims 4712, 4646), but given the slippage in earlier phases, the credibility of the long-dated commitment is questionable. Management has not disclosed revised rollout dates or acknowledged customer pressure.

The company's baseline 2025 capacity guidance of 16 MW (claims 4715, 4692) was promised by December 2025 and is now marked delayed nearly nine months past deadline in September 2026. This was supposed to demonstrate operational readiness and scaling capability. The absence of any revision to or discussion of this metric suggests it was never achieved and has been quietly abandoned rather than replaced with realistic 2026 or 2027 targets.

WhiteFiber's 2025 capacity target of 16 MW (claims 4715, 4692, 4649) was promised by end-2025 in July 2025 disclosures and has now silently transitioned to delayed status as of September 2026, more than eight months past the deadline. No revision announcement appears in the record to explain the miss or provide a new expected completion date. The MTL-2 and MTL-3 completion and revenue-start milestones (claims 4705, 4707), both pegged for 2026-01-31 in July 2025 announcements, have similarly shifted to delayed without a public walkback or updated guidance. The NC-1 Greensboro facility, described in July 2025 as having a 24 MW completion target for Q1 2026 (claims 4713, 4650) and revenue generation by May 2026 (claims 4714, 4690), has also missed both timelines and now carries delayed status; the related 40 MW cumulative capacity target by Q2 2026 (claims 4716, 4653) followed suit. Duke Energy partnership execution has fractured along multiple tranches: the 24 MW capacity commitment promised by September 2025 (claims 4710, 4686, 4644) and the 40 MW phase promised for April 2026 (claims 4711, 4687, 4645) have both transitioned to delayed status without disclosed revisions to completion dates. Underlying these delays is a chain of property and facility acquisitions that themselves missed earlier deadlines—the Enovum acquisition (promised October 2024, now delayed), MTL-2 acquisition (December 2024, now delayed), and NC-1 property purchase (May 2025, now delayed)—suggesting cascading facility-readiness issues that management has not explicitly addressed in revisions to the record.

WhiteFiber has gone silent on several major capacity and construction milestones without publicly revising timelines. The 2025 year-end capacity target of 16 MW (claims 4715, 4692, 4649) was promised by December 31, 2025, and last mentioned on July 28, 2025; the company has provided no update despite being nine months past deadline as of September 2026. The MTL-2 data center acquisition (claims 4680, 4640) was committed for December 27, 2024, and remains in delayed status with no revised completion date disclosed. The MTL-3 facility (claims 4707, 4648) was promised operational by end-2025 or early January 2026 and has fallen silent with no explanation for the now-eight-month slip. The NC-1 Greensboro facility, perhaps the most ambitious project, has accumulated multiple missed milestones: property acquisition promised for May 20, 2025 (claims 4709, 4685, 4643), initial 24 MW completion promised for March 31, 2026 (claims 4713, 4650, 4652), revenue generation promised for May 1–31, 2026 (claims 4714, 4690, 4651), and 40 MW cumulative capacity promised by June 30, 2026 (claims 4716, 4653). All remain in delayed status with no update on revised timelines. Duke Energy's two capacity phases—24 MW committed by September 1, 2025 (claims 4710, 4644, 4686) and 40 MW by April 1, 2026 (claims 4711, 4645, 4687)—have both slipped without explanation. The Enovum acquisition (claims 4703, 4678, 4638) was originally promised for October 1, 2024, and is now eleven months overdue with no progress update in public disclosures. Across all these items, there is no announcement explaining the delays, no revised target dates, and no material updates to shareholders despite substantial slippage.

WhiteFiber's end-of-2025 capacity target of 16 MW (claims 4715, 4692, 4649) was stated repeatedly across July 2025 disclosures and promised by 2025-12-31, but has now auto-transitioned to delayed status as of September 2026. The company has not issued a revised 2025 guidance target or explained the nine-month miss; the last public mention was July 2025. The Enovum acquisition (claim 4638), promised by 2024-10-01, is now nearly two years overdue and similarly marked delayed since September 2026, yet has not been revisited in any disclosed commentary since its July 2025 mention—no timeline push, no contingency narrative, nothing. The NC-1 facility rollout was promised across multiple tranches: initial 24 MW energization by Q1 2026 (claims 4713, 4650, 4652), revenue start by May 2026 (claims 4714, 4690), and cumulative 40 MW by Q2 2026 (claims 4716, 4653). All are now delayed by three to five months as of September 2026, yet no revised completion guidance has been disclosed to supersede the original promised dates. MTL-2 acquisition was promised by 2024-12-27 (claims 4680, 4640) and facility completion by 2026-01-31 (claim 4705); both are now marked delayed more than nine months past their deadlines, with no public acknowledgment that these anchor projects have stalled. The Duke Energy Phase 1 commitment of 24 MW, promised by 2025-09-01 (claims 4710, 4644, 4686), is now twelve months overdue and in delayed status, yet the company has not announced a revised milestone or explained why the largest North American utility capacity deal has quietly slipped without revision announcement.

WhiteFiber's NC-1 Greensboro facility, extensively promoted across three separate July 2025 disclosures, was promised to reach 24 MW operational capacity by March 31, 2026 and commence revenue generation by May or May 31, 2026 (claims 4713, 4714, 4650, 4651). As of September 6, 2026, this facility remains delayed—over four months past the most recent revenue target—with no management statement explaining the impediment or providing a revised timeline. Similarly, both Montreal facilities (MTL-2 and MTL-3) were promised completion and revenue start by December 31, 2025 (claims 4705, 4707, 4681, 4683, 4647, 4648). These are now nine months overdue with no public announcement revising expectations or disclosing what caused the slip. The Duke Energy Phase 1 partnership for 24 MW capacity carries the starkest absence of commentary: this commitment was promised by September 1, 2025 (claims 4710, 4686, 4644) and is now delayed by a full year without any progress update, revised target, or acknowledgment from management. Finally, WhiteFiber issued 2025 full-year capacity guidance of 16 MW (claims 4715, 4692, 4649), which the company missed by nine months; no revised guidance or explanation of why the baseline capacity target went unmet has been issued. All of these milestones were extracted from July 2025 disclosures, auto-transitioned to delayed status only on September 3, 2026, and remain unaccompanied by any management statement offering revised dates or rationale for the reversals.

WhiteFiber's guidance for 16 MW total operational capacity by end of 2025 (claims 4715, 4692, 4649) has been systematically repeated across three disclosure cycles from July 10 through July 28, 2025, each time with a promised delivery date of December 31, 2025. As of September 6, 2026, this target is now nine months overdue and marked 'delayed' with no revised timeline issued or explanatory statement from management. The Duke Energy Phase 1 capacity commitment of 24 MW (claims 4710, 4686, 4644) was promised for September 1, 2025, and has now slipped twelve months into a 'delayed' status, similarly without a public revision to the commitment or updated delivery expectations. The NC-1 facility's initial 24 MW energization was promised for March 31, 2026 (claims 4713, 4652), placing it six months overdue as of the current date; despite multiple redisclosures of this target between July 10 and July 28, 2025, management has not announced a revised schedule or addressed the delay. The Enovum data center acquisition first appeared on record as an accomplished acquisition without timeline detail in the July 10 and July 22 disclosures (claims 4703, 4678, status 'stated'), but reappeared on July 28 (claim 4638) with a retroactive promised date of October 1, 2024—already seven months past when disclosed—and marked 'delayed,' suggesting the acquisition encountered problems that were not communicated in prior statements. Similarly, MTL-3 site acquisition began as an announced project with no stated deadline (claim 4706, July 10, status 'stated'), then appeared on July 22 and July 28 (claims 4682, 4641) with a promised date of April 11, 2025, which was already three months past when first assigned to this facility in the July 22 disclosure, indicating a significant missed milestone that management did not explicitly call out.

WhiteFiber's 2025 year-end capacity target of 16 MW (claims 4715, 4692, 4649) was promised by December 31, 2025 and last mentioned across three separate July 2025 disclosures. As of September 2026, this commitment is now nine months overdue and marked delayed, yet no revision announcement or updated guidance has been published to acknowledge the miss or provide a new target date. Similarly, the NC-1 facility represents a cascade of quietly slipping milestones: the 24 MW initial capacity completion (promised by March 31, 2026 per claims 4713, 4650, 4652), the revenue generation start (promised by May 1 or May 31, 2026 per claims 4714, 4690, 4651), and the full 40 MW build-out (promised by June 30, 2026 per claims 4716, 4653) are all now in delayed status with no corresponding investor update or revised timeline. The Duke Energy Phase 1 partnership commitment of 24 MW (claims 4710, 4686, 4644) was promised by September 1, 2025 and is now over one year overdue without public explanation for what went wrong or when delivery is expected. Foundational acquisitions—MTL-2 (promised by December 27, 2024), MTL-3 (promised by April 11, 2025), and Enovum plus the NC-1 property (promised by dates ranging from October 2024 to May 2025)—all appear in July 2025 disclosures and are now substantially delayed. None of these slippages have been accompanied by revised guidance or explanations in the public record; instead, the claims remain frozen in their original promised states as of the last disclosure, having transitioned to delayed status only through automated flagging.

WhiteFiber's Montreal expansion projects (MTL-2 and MTL-3) were disclosed with end-of-2025 completion targets, explicitly appearing in claims 4647 and 4648 (promised by 2025-12-31), then quietly rephased to January 2026 in claims 4705 and 4707 (promised by 2026-01-31). As of today, both remain in delayed status with no disclosed explanation for the slip or revised timeline communicated; revenue generation from these facilities, also promised for Q4 2025 and pushed to 2026-01-31, has not been publicly updated. Duke Energy's Phase 1 capacity commitment of 24 MW (claims 4710, 4686, 4644) was promised by 2025-09-01 and is now one year overdue with delayed status persisting since 2026-09-03, yet no investor disclosure has addressed why this partnership milestone was missed or when it is now expected. The NC-1 Greensboro property acquisition (claims 4685, 4643) was promised by 2025-05-20 and remains in delayed status as of 2026-09-05—a 16-month slippage with no revision statement; similarly, NC-1 revenue generation (claims 4714, 4651) was promised for 2026-05-01 and 2026-05-31 respectively and is now delayed four months beyond those dates without public update. More strikingly, two foundational acquisitions—Enovum (claim 4638, promised by 2024-10-01) and MTL-2 (claims 4680, 4640, promised by 2024-12-27)—are nearly two years late and remain marked delayed, suggesting material integration or regulatory obstacles that management has not disclosed. The 2025 capacity guidance of 16 MW (claims 4715, 4692, promised by 2025-12-31) was missed entirely; while guidance was rephased upward to 76 MW by end-2026 (claims 4717, 4654, 4693), no statement has addressed why the 2025 target failed or what went wrong in the year-earlier deployment cycle.

WhiteFiber disclosed a 16 MW capacity target by end of 2025 (claims 4715, 4692, 4649) in July 2025. That deadline passed nine months ago; the status is now 'delayed' with no revision announcement or explanation of the slip. The North Carolina facility (NC-1) was promised in three distinct milestones: 24 MW operational by March 31, 2026 (claims 4713, 4650, 4652), 40 MW cumulative by June 30, 2026 (claims 4716, 4653), and revenue start by May 1, 2026 (claims 4714, 4690). All three are now 'delayed' as of September 2026 with no public update reframing timelines or explaining the setback. Montreal's two facilities underwent parallel timeline misses: MTL-2 and MTL-3 completions were promised for end of 2025 (claims 4647, 4648) sliding to January 31, 2026 (claims 4705, 4707), while underlying property acquisitions for MTL-3 and MTL-2 were promised for April 2025 and December 2024 respectively (claims 4682, 4680)—all now flagged 'delayed' without management commentary on causes or revised dates. WhiteFiber's Duke Energy partnership carried near-term implementation targets: Phase 1 (24 MW) promised by September 1, 2025 (claims 4710, 4686, 4644) and Phase 2 (40 MW) promised by April 1, 2026 (claims 4711, 4687, 4645) both now 'delayed' as of September 2026, while only the long-dated Phase 3 (99 MW for May 2029) remains 'stated'. The Enovum data center acquisition was originally disclosed as promised by October 1, 2024 (claim 4638), appearing again in July 2025 disclosures as in-progress, yet is now 'delayed' with no explanation of the 11-month slip or revised close timeline.

WhiteFiber's Montreal expansion projects—MTL-2 and MTL-3—were promised to reach operational status by end of 2025 (claims 4647, 4648) with MTL-3's property acquisition explicitly tied to April 11, 2025 (claim 4641). Both now carry 'delayed' status as of September 3, 2026, yet the last disclosure mentioning them dates to July 28, 2025. Management has not announced revised timelines, capacity impacts, or reasons for the slip. The 2025 capacity guidance of 16 MW (claims 4649, 4715, 4692) was promised by December 31, 2025; it is now September 2026 and the company has not publicly revised or withdrawn this target, nor explained why operational capacity did not meet the stated 2025 threshold. Three Duke Energy partnership milestones were woven into July 2025 filings: a 24 MW phase promised by September 1, 2025 (claim 4644), a 40 MW phase by April 1, 2026 (claim 4645), and a 99 MW phase by May 2029 (claim 4646). The first two are now overdue by months and carry 'delayed' status, yet no announcement has revised capacity commitments, delayed the obligations, or explained execution gaps. The NC-1 Greensboro property acquisition, promised by May 20, 2025 (claim 4685), is now 16 months past deadline and marked delayed; initial 24 MW capacity was promised by March 31, 2026 (claim 4650) and is now 5 months overdue; cumulative 40 MW capacity was promised by June 30, 2026 (claim 4653) and is now 2+ months behind. No revised NC-1 construction schedule has been disclosed. The Enovum data center acquisition was promised by October 1, 2024 (claim 4638)—nearly two years before today—and carries 'delayed' status; there is no public statement clarifying whether the deal was abandoned, renegotiated, or remains pending. Across all these items, the last management disclosures date to July 28, 2025; the automatic 'delayed' transition occurred September 3, 2026, suggesting no proactive revision communication preceded the status change.

WhiteFiber's 2025 capacity guidance—16 MW promised by December 31, 2025 (claims 4715, 4692)—slipped silently into the ledger as 'delayed' on September 3, 2026, nearly nine months overdue. The guidance was last discussed in the July 22 disclosure, but no revision announcement or rescheduled target appears in the record; management has simply stopped mentioning it. Similarly, the Duke Energy Phase 1 partnership commitment of 24 MW, promised by September 1, 2025 (claims 4710, 4686, 4644), shifted to delayed status without explanation and is no longer discussed in subsequent filings. The NC-1 facility in Greensboro underwent a cascade of promised milestones—initial 24 MW completion by March 31, 2026 (claims 4713, 4650, 4652), revenue start by May 1, 2026 (claims 4714, 4690, 4651), and 40 MW cumulative by June 30, 2026 (claims 4716, 4653)—all now marked delayed as of September 2026. Management disclosed these targets through July 28, 2025, but no updated timeline or revised guidance for NC-1 appears in the record; the promises simply transitioned to 'delayed' without accompanying disclosure revision. Property acquisitions for MTL-2 (promised December 27, 2024), MTL-3 (promised April 11, 2025), and the Enovum data center (promised October 1, 2024) all similarly auto-transitioned to delayed status, with no statement clarifying their current status, revised completion dates, or reasons for delay. The pattern suggests management has allowed promised facility timelines and capacity targets to lapse into the automated 'delayed' status without issuing formal revised guidance or explanatory updates.

WhiteFiber has experienced systematic slippage across its facility buildout roadmap without public correction or revised guidance. The MTL-2 acquisition (claimed 4704, 4640) promised by 2024-12-27 is now twenty months overdue; the MTL-3 site acquisition (claimed 4706, 4641) promised by 2025-04-11 is sixteen months behind schedule; the NC-1 Greensboro property acquisition (claimed 4709, 4643, 4685) promised by 2025-05-20 is fifteen months late. All three facility acquisition milestones were stated repeatedly across July 2025 press disclosures (July 10, 22, 28) but have never been reported complete. The company's 2025 capacity guidance of 16 MW (claimed 4715, 4692, 4649) promised by 2025-12-31 has missed by nine months with no updated 2025 outcome disclosed. The Duke Energy Phase 1 capacity commitment of 24 MW (claimed 4710, 4686, 4644) promised by 2025-09-01 is twelve months overdue and represents the most material partnership milestone to be silently abandoned. Downstream deliverables—NC-1 initial 24 MW completion by 2026-03-31 (claimed 4713, 4650), MTL-2 revenue start by 2026-01-31 (claimed 4705), MTL-3 revenue start by 2026-01-31 (claimed 4707), and NC-1 revenue start by 2026-05-01 (claimed 4714, 4690)—are all marked delayed as of 2026-09-03 with no management revision statements in the record. The 76 MW total capacity target by 2026-12-31 (claimed 4717, 4654) remains stated, despite every component facility being behind plan. What has not been said: no press release or investor update explaining the root causes of cascade delays, no revised completion timelines, no disclosure of the impact on Bit Digital's post-IPO ownership stake or the company's cash runway.

The 2025 total capacity target represents WhiteFiber's first major promise to slip without revision. Management promised 16 MW of operational capacity by December 31, 2025 (claims 4715, 4692, 4649), last mentioned in July 2025 disclosures. That deadline has now passed without any public guidance update or delay announcement from management. The NC-1 North Carolina facility expansion comprises three cascading milestones: 24 MW energization by March 31, 2026 (claims 4713, 4650), revenue generation by May 31, 2026 (claims 4714, 4690, 4651), and 40 MW cumulative by June 30, 2026 (claims 4716, 4653), all appearing in July 28, 2025 disclosures. All three promised dates have passed and all three are marked delayed in current records, yet no revised timelines or public delay explanation has been disclosed. Montreal's MTL-2 and MTL-3 facilities were promised parallel completion to operational and revenue-generating status by January 31, 2026 (claims 4705, 4707, 4647, 4648), stated in July 2025 disclosures. That deadline passed seven months ago; both remain marked delayed without updated completion dates or management explanation of the slippage. The Duke Energy partnership commitment phases similarly slipped without revision. Phase 1 promised 24 MW capacity delivery by September 1, 2025 (claims 4710, 4686, 4644), and Phase 2 promised 40 MW by April 1, 2026 (claims 4711, 4687, 4645), both disclosed in July 2025. Both promised dates have passed and both remain marked delayed today, but management has not provided revised commitment schedules or public explanation for the delays.

WhiteFiber's Enovum data center acquisition stands as the most conspicuous gap. Disclosed across three separate disclosure dates (July 10, 22, and 28 of 2025), the company promised acquisition completion by October 1, 2024—yet as of today, September 5, 2026, this milestone remains in 'delayed' status with no public announcement explaining a two-year slip or revision to the promised date. Similarly, the MTL-2 Montreal property acquisition, originally promised by December 27, 2024, has drifted into 2026 without management addressing the delay through a formal update.

The Duke Energy partnership capacity commitments reveal a pattern of silent erosion. Three tranches—24 MW (promised September 1, 2025), 40 MW (promised April 1, 2026), and 99 MW (promised May 2029)—were prominently featured in July 2025 investor disclosures. The first two phases are now delayed as of today, having missed their promised windows by months and a year respectively, yet no press release or investor update has recast the timing or explained what halted progress against Duke's stated commitments.

The NC-1 Greensboro facility represents a broader construction and commercial collapse from the 2025 projections. WhiteFiber promised property acquisition by May 20, 2025; initial 24 MW energization by March 31, 2026; 40 MW cumulative capacity by June 30, 2026; and revenue generation start by May 31, 2026. All four of these interdependent milestones are now marked delayed as of today. No announcement has revised the phased expansion roadmap or explained what halted ground-level progress at the North Carolina site.

Capacity guidance targets disclosed in July 2025—16 MW total by year-end 2025 and 76 MW by year-end 2026—have similarly transitioned to delayed status without a revised guidance statement. The 16 MW target missed by over eight months; the 76 MW target is at risk as of today with no management commentary on attainment probability or revised 2027 expectations.

WhiteFiber's Enovum data center acquisition has drifted into indefinite delay without public revision. Claimed as closed in July 2025 (claims 4638, 4703), it carried a completion date of October 2024—already nine months past—when disclosed in July 2025 itself. The company has not announced a revised timeline, a post-deal integration update, or any explanation for why an acquisition supposedly already consummated carries a missed deadline. The last substantive mention appears in the July 2025 filings; no investor update since clarifies whether the deal is still active or has quietly stalled.

Duke Energy's Phase 1 commitment of 24 MW has transitioned from 'signed' status to 'delayed' without accompanying announcement. Initially promised for September 2025 (claims 4710, 4644, 4686), this milestone now sits twelve months overdue as of today. Management disclosed the partnership multiple times across three separate July 2025 disclosures but has not publicly revised the delivery date, walked back the capacity commitment, or explained why the first phase of a marquee partnership with a major utility has not materialized. The three July filings treat it as firm; the system now marks it as delayed.

The MTL-2 Montreal expansion facility was promised complete and revenue-generating by end of 2025 (claims 4647, 4705, 4681). No revision announcement appears in public record. Today is September 2026—nine months past deadline—and the company has not issued a new completion guidance, reduced the promised 5 MW capacity allocation, or disclosed external factors (permitting delays, supplier issues, capital constraints) that would explain the miss. The facility appears in multiple July disclosures as an active expansion vector; its current non-operational status represents unannounced downward revision of near-term revenue expectations.

NC-1 Greensboro was presented as a phased deployment with three distinct deadlines: 24 MW initial capacity by March 2026 (claims 4713, 4652, 4650), revenue start by May 2026 (claims 4714, 4690), and 40 MW by Q2 2026 (claims 4716, 4653). All three are now delayed. The company provided specific quarterly guidance tying these to Duke Energy's Phase 2 (40 MW by April 2026, also delayed). No revision has been announced; no explanation for three slipped facility milestones appears in the record. The project remains described as planned but with no updated construction timelines or reduced capacity claims.

The 2025 year-end capacity guidance of 16 MW (claims 4715, 4649, 4692) was missed entirely. Management committed to this target in three separate July 2025 disclosures; December 2025 came and went without either achievement or revision. The company has not subsequently disclosed revised full-year 2025 results, restated capacity, or acknowledged the gap. This silence is particularly notable because capacity deployment is WhiteFiber's core metric; missing annual guidance by this margin typically warrants an investor update.

The convertible notes exchange and reduction (claims 22 and 23) carried promised dates of 2026-08-21, now nine days past, yet remain marked as 'stated' status with no announcement of completion or delay acknowledgment. Management disclosed these transactions prominently on August 20 and 21 but has not updated investors on whether the $198.15M-to-$118.5M exchange actually settled by the promised date or slipped. The silence is notable given that the transaction involves material restructuring of $310M in debt. Separately, facility capacity language lacks consistency: claim 31 (August 16) commits to a minimum 60 MW by October 30, 2026, while claim 9 (August 21) targets '100+ MW' for 2027, but management has not clarified whether the earlier commitment was revised upward, superseded, or applies to different properties. Claims 24 and 25, announced August 20, reference 'Data Center Expansion' and 'Energy Service Agreements' without promised dates, specific megawatt targets, location detail, or supplier names, contrasting with the quantified precision of the earlier Yadkin County acquisition claim and suggesting these initiatives remain unbounded or contingent. Claim 27 uses the phrase 'Potential Acquisitions, Partnerships and Joint Ventures' without distinguishing between signed, discussed, or exploratory stages, which is weaker disclosure language than the definite 'purchase agreement' language in claim 19 announced five days later. The record shows no clarification on whether these represent new efforts or rephrasing of previously disclosed intentions.