jain.com — data centers, connectivity and security, analyzed daily

Tag: CoreWeave

  • CoreWeave’s 15x Reload Claim Puts AI Inference on the Training Critical Path

    CoreWeave’s 15x Reload Claim Puts AI Inference on the Training Critical Path

    TL;DR · 30-second read

    The Short Version

    CoreWeave is a company that rents out the powerful computer chips used to build artificial intelligence. It has now launched Forge, a software toolkit for building and running artificial intelligence systems. Anyone can start using it for free.

    The surprising part: when these systems learn by trial and error, they keep swapping in updated versions of themselves. CoreWeave says it made that swap 15 times faster in testing, so the learning stops stalling.

    The move also puts CoreWeave in more direct competition with Amazon, Google and Microsoft, which already sell similar toolkits.

    CoreWeave has launched CoreWeave Forge, a platform that connects model serving, observability, post-training and evaluation in one environment. It was unveiled at the Fully Connected event and is free to start, with paid tiers for additional capabilities. CoreWeave’s product page describes Forge as a way to “run, observe, curate, improve, and evaluate models and agents” without locking work to one cloud, model or framework.

    SiliconANGLE reported on October 8 that Urvashi Chowdhary, CoreWeave’s vice president of product and AI services, said a preview capability called CoreWeave RL Rollouts, built on Nvidia’s Dynamo framework, improved model reload latency by 15x compared with a baseline configuration in testing.

    Executive Summary

    CoreWeave built its business renting GPU capacity, the graphics processors that train and run AI models. Forge adds managed software on top of that capacity. According to CoreWeave’s product page, it covers fine-tuning, managed reinforcement learning, per-token serverless inference, dedicated inference, agent tracing, sandboxes, a model and dataset registry, and experiment tracking through Weights & Biases Models. The free entry tier extends the platform to individual developers, not only the large labs and enterprises that sign capacity contracts.

    The most technically significant claim concerns reinforcement learning, the training method used to build agentic models that take multistep actions. In that workflow, the inference system that generates the model’s attempts has to keep reloading freshly trained versions of the model. CoreWeave says RL Rollouts cut that reload time 15-fold in testing. If that holds outside the lab, inference performance becomes a direct constraint on how fast these models can be trained, which changes how teams should size and buy GPU capacity.

    Commercially, Forge places CoreWeave alongside hyperscaler AI platforms such as Amazon Bedrock and SageMaker, Google Vertex AI and Microsoft Azure AI Foundry. Its distinguishing pitch is an open-source foundation and portability across clouds, models and frameworks.

    Inference Is Now on the Training Clock

    In conventional AI development, training and inference are separate phases. A model learns from a fixed dataset, and inference, meaning serving the model to answer requests, comes afterward. Reinforcement learning for agentic models erases that boundary. The model generates attempts at a task, called rollouts. Verifiers check the results and rewards score them, and those scores update the model. Each update produces a new checkpoint, a saved snapshot of the model’s weights, which must be loaded back into the inference servers that produce the next round of attempts. Chowdhary described inference as the bottleneck during rollouts, and the mechanics show why: every reload is time the loop spends waiting.

    That is the problem the 15x figure addresses. CoreWeave RL Rollouts is built on Nvidia’s Dynamo, an open-source framework for distributing inference across many GPUs. It loads new checkpoints into a live deployment, and Chowdhary said the approach improved model reload latency 15-fold against a baseline configuration. The practical effect, in Chowdhary’s framing, is that training keeps moving while inference scales independently. This matters to the people who plan and buy GPU capacity. When inference sits inside the training loop, the size and responsiveness of the inference fleet partly determine training throughput. Teams can no longer budget the two as unrelated line items.

    The claim deserves careful reading. The baseline configuration has not been specified. The figure comes from testing. The capability is in preview. Reload is also only one step in each loop iteration, alongside generating rollouts, scoring them and updating weights, so a 15x gain in reload time does not translate into a 15x gain in end-to-end training speed. CoreWeave has not quantified that overall effect. The direction of the argument is sound. Its magnitude in production is still unproven.

    From GPU Landlord to Platform Vendor

    CoreWeave’s product page lays out Forge as a five-stage loop. Teams run models and agents, observe them through traces in Agent Lens, curate failures into versioned datasets, improve the model through fine-tuning, reinforcement learning or distillation, and evaluate the new version before it ships. Pricing models span per-token serverless inference on open models and dedicated inference, where a customer’s own weights run on GPUs it chooses. Chowdhary described the strategy as building managed services in layers on top of the infrastructure, so that each service builds on the others.

    The economic logic is familiar from the hyperscalers. Raw GPU capacity is easiest to sell when it is scarce, and it is easiest to compare on price when it is not. Software layers give a provider other ways to compete. Customers who keep their experiments, traces, datasets and model registry on one platform have reasons to stay. Per-token pricing lets the provider capture value from efficiency gains on hardware it already owns. The free tier widens the funnel to individual developers who would never sign a capacity contract. Amazon, Google and Microsoft have pursued the same strategy for years by bundling compute with managed AI platforms. Forge is CoreWeave’s entry into that contest.

    CoreWeave’s chosen point of difference is openness. Chowdhary emphasized open-source tools such as vLLM and contributing back to open projects, and the product page promises work that is not locked to one cloud, model or framework. That is a credible counter to proprietary platform stacks. It also creates a tension the company will have to resolve in practice, because an integrated, connected loop is valuable precisely because it is sticky. Portability will be judged by how easily customers can move data, traces and registry entries out, not by the launch language.

    Reading the Inference Demand Signal Carefully

    The demand case rests on one data point. In a theCUBE Research survey of CoreWeave customers and prospects, one healthcare customer’s share of inference workload rose from about 10% in its first year to 40% in its second, with roughly 50% expected within 12 months. That fits the thesis that serving models, not training them, will drive the next phase of AI cloud economics. However, it is a single customer drawn from CoreWeave’s own customer base. It illustrates a trend and does not measure one.

    The infrastructure implication still holds. As more of a GPU’s working life goes to serving requests, providers compete on how much useful output each GPU produces. CoreWeave lists Nvidia’s Hopper, Blackwell and Vera Rubin generations, which competitors can also buy. When the hardware is shared, efficiency comes from the software above it. Chowdhary pointed to that layer: the vLLM serving engine, quantized models, which use lower-precision numbers to shrink a model and speed it up, and custom speculative decoders, in which a small draft model proposes output that the large model verifies. Forge packages that tuning as a product. Over time, it could matter as much as the size of the fleet.

    Background

    CoreWeave began as a GPU-based cryptocurrency operation and pivoted to become a specialized AI cloud, renting large clusters of Nvidia GPUs to AI labs and enterprises. It listed on Nasdaq under the ticker CRWV in 2025 and acquired Weights & Biases, the AI experiment-tracking company, the same year. That tooling now sits inside Forge. Its infrastructure catalog spans Nvidia’s Hopper, Blackwell and Vera Rubin GPUs, managed Kubernetes, AI object storage and fleet-management software.

    Specialized GPU clouds grew on demand for model training, which consumes large blocks of capacity for weeks at a time. As AI products reach users, inference, the continuous work of running models to answer requests, takes up a growing share of compute. That has pushed providers to compete on software efficiency and managed services, the territory where Amazon Web Services, Google Cloud and Microsoft Azure have long operated.

    Sources

    Source: AI inference gets a full stack as CoreWeave launches Forge — SiliconANGLE’s interview with CoreWeave’s Urvashi Chowdhary on Forge, inference optimization and RL Rollouts.

    Primary sources: CoreWeave Forge → Models, agents, training, and inference (CoreWeave product page).

  • Applied Digital’s Energized 150 MW Puts Its AI Lease Backlog to a Revenue Test

    Applied Digital’s Energized 150 MW Puts Its AI Lease Backlog to a Revenue Test

    TL;DR · 30-second read

    The Short Version

    • Applied Digital just finished switching on a huge computing building for artificial intelligence. Its electricity capacity is comparable to a small city’s demand.
    • Its main customer, CoreWeave, rents out artificial intelligence computing power. CoreWeave is reported to have signed 15-year rental deals worth about $7 billion in total.
    • Signing a deal is not the same as getting paid. Rent usually starts once the space is powered and ready, and this building now is.
    • Investors remain cautious. The shares are down nearly a quarter in three months, and the company carries heavy debt.

    Applied Digital (Nasdaq: APLD) has switched on the second phase of Building 2 at Polaris Forge 1, its campus purpose-built for artificial intelligence and high-performance computing (HPC), Simply Wall St reported on October 2, 2026. The phase adds 75 megawatts (MW) of IT capacity. IT capacity is the power available to run servers and GPUs, as distinct from cooling and other building loads. The phase completes the facility’s 150 MW buildout.

    The shares rose 5.05% on the day to US$25.38. They remain down 23.23% over three months, though the three-year total shareholder return is still roughly fourfold.

    Executive Summary

    The milestone itself is simple. A second 75 MW block is now live at Building 2, and the facility’s planned 150 MW is complete. Its significance lies in what it does to the company’s revenue story. Polaris Forge 1 is tied to long-term leases with CoreWeave, a cloud provider that rents out AI computing power. Those leases are described as 15-year agreements worth $7 billion in total contracted revenue. A contract of that kind is a promise to pay. Energized, delivered capacity is generally what turns the promise into rent that appears on an income statement.

    That conversion is where the valuation debate now sits. On a price-to-sales basis, the stock trades at about 12.1 times revenue, against roughly 1.9 times for the broader US IT sector. Bullish valuation models put fair value well above the current price, but they depend on revenue that has not yet been reported. With the 150 MW buildout finished, the coming quarters should begin to show how much of the backlog is converting into revenue, and on what timetable.

    The risks are the ones an energization milestone does not remove. The company depends on a small number of hyperscale tenants, meaning very large cloud and AI compute buyers, and it carries a sizeable debt load. If contracts slip or financing tightens, the upside case is exposed quickly.

    Contracts Are Promises. Energized Megawatts Are Invoices.

    In wholesale data center leasing, rent typically starts when a tenant’s capacity is delivered and accepted, not when the contract is signed. A signed lease establishes backlog, the revenue a company is entitled to over time. Building, powering and handing over the space is what starts the billing clock. Switching on the second 75 MW phase of Building 2 completes the 150 MW facility at Polaris Forge 1. The question for Applied Digital is no longer whether it can finish this building. It is how fast the capacity turns into recognized revenue.

    The scale of that backlog gives the milestone its weight. The most-followed valuation narrative on Applied Digital describes the CoreWeave agreements as 15-year leases totalling $7 billion in contracted revenue. Spread evenly, that works out to roughly $467 million a year. That figure is simple arithmetic, not a disclosed payment schedule, and real lease revenue usually ramps as phases come online. It still shows the size of the revenue stream that delivered capacity is meant to unlock. Until the building is live, that stream exists only on paper.

    Three groups are most directly affected. Shareholders are paying today for revenue that energization begins to release. Lenders’ repayment depends on that rent arriving on schedule. CoreWeave depends on Applied Digital to house the computing capacity it in turn rents to its own customers.

    Why a 12.1x Sales Multiple Rides on Delivery

    Simply Wall St’s own price-to-sales comparison puts the stock at about 12.1 times sales. That compares with roughly 1.9 times for the wider US IT group and a 5.5 times multiple the publication treats as a fair ratio. A high price-to-sales ratio means investors are paying for revenue they expect rather than revenue already booked. As the publication notes, a gap like this can close through stronger fundamentals, a lower share price, or both. Delivered megawatts are the most direct route to stronger fundamentals. Each phase that starts paying rent raises the sales line without the share price having to fall.

    The bullish narrative assumes much more than one campus. Its US$68.35 fair value, about 63% above the US$25.38 close, rests on an 11.07% discount rate. It also assumes Applied Digital reaches roughly US$3.0 billion in revenue and US$295.0 million in earnings by around 2029, which still implies a forward price-to-earnings multiple near 104.7 times. Against that US$3.0 billion target, the CoreWeave leases at their simple annual average would cover only about a sixth. The remainder of the bull case therefore depends on capacity beyond this 150 MW facility, along with the financing and tenants needed to fill it.

    A fair reading cuts both ways. Completing the facility is real execution, and it narrows the gap between contracted and earned revenue. It does not, by itself, validate a valuation built on several times more revenue than this campus is reported to be contracted for.

    The Risks Energization Does Not Retire

    Bringing capacity online moves risk without eliminating it. Construction and delivery risk on this building falls away. Counterparty risk, the chance a tenant pays late, renegotiates or fails to take full occupancy, becomes more prominent as revenue concentrates on the leases now running. The valuation commentary flags dependence on a few hyperscaler tenants as a central risk. With CoreWeave named as the anchor customer at Polaris Forge 1, much of the campus’s near-term revenue rests on one counterparty.

    The balance sheet is the other pressure point. Data center campuses are capital-intensive, and the sizeable debt load cited in the valuation commentary must be serviced whether or not rent ramps on schedule. The recent three-month decline of 23.23%, after a roughly fourfold three-year return, suggests the market is giving more weight to these execution and funding questions than it did earlier in the cycle. A completed building is a necessary condition for the bull case, but not a sufficient one.

    Background

    Applied Digital designs, develops and operates digital infrastructure for high-performance computing and artificial intelligence customers in North America, and trades on the Nasdaq Global Select Market under the ticker APLD. Its fiscal year runs from June 1 to May 31, according to the definitive proxy statement it filed with the SEC on September 25, 2026.

    The company has positioned Polaris Forge 1 as a purpose-built AI campus anchored by long-term leases with CoreWeave. CoreWeave is one of the specialist cloud providers, often grouped with hyperscalers, the largest buyers of data center capacity, that rent GPU computing to AI developers. Investors have rewarded that strategy with a roughly fourfold three-year shareholder return, though the shares have cooled sharply in recent months as attention has shifted to execution and funding.

    Sources

    Source: Applied Digital (APLD) Completes Polaris Forge 1 Expansion, Is The Valuation Gap Too Wide? (Simply Wall St). Report on the second 75 MW phase of Building 2 completing Polaris Forge 1’s 150 MW buildout, with valuation commentary.

    Primary sources: Applied Digital, Form DEF 14A (definitive proxy statement), filed with the SEC on September 25, 2026.

  • CoreWeave Brings Up Multi-Rack Vera Rubin NVL72: What It Asks of Operators

    CoreWeave Brings Up Multi-Rack Vera Rubin NVL72: What It Asks of Operators

    TL;DR · 30-second read

    The Short Version

    CoreWeave, a company that rents out computing power for artificial intelligence, says it has switched on several of Nvidia’s newest server cabinets and wired them together to work as one giant machine.

    Each cabinet packs 72 of Nvidia’s top chips. Linking hundreds of those chips lets companies build and run bigger, more capable digital assistants.

    Why it matters: machines this crowded need a lot of electricity and liquid cooling, so the buildings that house them have to change quickly. CoreWeave has not said who is using the system or when customers can rent it.

    CoreWeave (Nasdaq: CRWV) has brought up a multi-rack NVIDIA Vera Rubin NVL72 cluster on CoreWeave Cloud, joining hundreds of NVIDIA Rubin GPUs into a single scale-out cluster aimed at agentic AI, Light Reading reported on September 16, 2026, publishing the company’s announcement. “Bring up” is the industry term for powering on, configuring and validating new hardware so it can run real workloads.

    Alongside the compute news, CoreWeave introduced two additions to CoreWeave AI Object Storage: cross-region write acceleration, which lets a job write data locally while it is copied to another region in the background, and a new Archive tier for colder data.

    Executive Summary

    Each Vera Rubin NVL72 rack combines 72 Rubin GPUs, 36 Vera CPUs, NVIDIA’s NVLink 6 chip-to-chip interconnect, ConnectX-9 SuperNICs (high-speed network cards) and BlueField-4 DPUs (processors that offload networking and security tasks). CoreWeave says it has joined several of these racks over NVIDIA Spectrum-X Ethernet so they behave as one cluster for training, inference and reinforcement learning. The company attributes the result to automation in its Mission Control software, including a Rack LifeCycle Controller that handles hardware detection, firmware, validation, power and cooling.

    The significance is less about any single rack than about what comes with it. Rack-scale systems shift the unit of deployment from the server to the entire cabinet, and that concentrates power, heat and networking demands in a way that tests data center design. CoreWeave is positioning itself as an early operator of NVIDIA’s next generation, and signaling that speed of bring-up is now a competitive metric in its own right.

    The announcement is detailed on architecture but light on the facts that determine commercial impact: how many racks are running, where, for whom, and when customers can buy the capacity.

    The Rack Is the New Computer

    For most of cloud computing’s history, the basic building block was a server: a box with a few processors that could be installed, swapped and scaled one at a time. NVIDIA’s NVL72 designs change that. Inside one rack, 72 GPUs are tied together by NVLink so they can share work almost as if they were a single, very large chip. That is known as scale-up. Joining multiple racks over a network, as CoreWeave describes with Spectrum-X Ethernet, is scale-out.

    The combination matters because the largest AI models and the most demanding inference jobs no longer fit comfortably inside one rack. CoreWeave frames the multi-rack cluster around agentic AI, meaning systems that chain many model calls and tool uses together to complete a task. Its argument is that delays compound across those steps, so tightly coupled compute and fast data access pay off. That reasoning is plausible, though the company did not publish benchmarks showing how much faster agentic workloads run on the new cluster.

    The networking claims are ambitious. CoreWeave says each Rubin GPU gets two ConnectX-9 SuperNICs, providing 1.6 terabits per second of scale-out connectivity, and that the fabric supports roughly 128,000 GPUs per rail in a non-blocking design, meaning traffic should not queue for lack of capacity. That figure describes what the network architecture can accommodate, not what has been deployed; the cluster announced here is described as hundreds of GPUs.

    Deployment Speed Becomes a Competitive Weapon

    The most revealing part of the announcement is how much of it is about process rather than silicon. CoreWeave describes racks arriving as hardware that still has to be connected and validated, and says its Rack LifeCycle Controller automates that work, with software components it calls Racky handling rack control and Valvey executing cooling actions. Racks then go through NVIDIA field diagnostics plus full-rack workload testing, and only those that pass as a system move into production.

    This matters economically. Every week a costly rack sits installed but unvalidated is capacity that earns nothing while its financing costs keep accruing. For GPU cloud providers competing for the same scarce next-generation hardware, the ability to turn delivered racks into billable capacity quickly can be as important as securing the allocation in the first place. Automating firmware updates, power sequencing and cooling control is how an operator compresses that interval.

    The claim that “every GPU performs at its best” is marketing language and should be read as such. What the release does substantiate is a described validation workflow; it does not provide failure rates, burn-in durations or time-to-production figures that would let buyers compare CoreWeave’s process with competitors’.

    What Next-Generation Racks Demand of Facilities

    CoreWeave explicitly lists cooling and power among the layers that had to be engineered for the cluster to perform as one system, and it built a dedicated software component to execute cooling actions. That is a clear signal of where the operational burden now sits. Dense rack-scale AI systems of this class have relied on liquid cooling, piping coolant directly to the chips because air alone cannot remove heat fast enough.

    For data center developers and colocation providers, the implication is that readiness for the rack era is not just about floor space. It requires facility cooling loops, electrical distribution designed for highly concentrated loads, and structural and networking plans that allow new racks to be added without rework. CoreWeave’s point that its modular network topology lets racks be added without redesigning the fabric at each expansion speaks to that last requirement.

    Operators whose buildings were designed for traditional air-cooled servers face costly retrofits or the prospect of being passed over for this generation of hardware. Suppliers of liquid-cooling equipment, power distribution gear and high-capacity optical networking stand to benefit, while the release gives no per-rack power figures that would let facility planners size requirements precisely.

    Storage Moves Closer to the Loop

    The storage announcements address a less glamorous bottleneck: GPUs sitting idle while they wait for data. With cross-region write acceleration, CoreWeave says a job writes locally while replication to another region happens in the background, so intermediate results, retrieved context and outputs are not held up by long-distance transfers. The new Archive tier gives customers a place to keep colder data within the same storage service.

    Background replication is a well-established technique, and its value depends on details the company has not provided, such as how far replicated copies can lag, how conflicts are handled and what happens if a region fails before replication completes. For buyers, the practical question is whether keeping training data, checkpoints and archives within one provider’s storage stack is worth the added dependence on that provider.

    Background

    CoreWeave is a Nasdaq-listed cloud provider (ticker CRWV) that specializes in renting GPU computing capacity for artificial intelligence workloads, marketing itself as “The Essential Cloud for AI.” Unlike general-purpose clouds, its business is built around large clusters of NVIDIA accelerators plus the networking, storage and orchestration software needed to run them for AI training and inference.

    NVIDIA’s Vera Rubin platform is the successor to its Grace Blackwell generation, pairing Rubin GPUs with Vera CPUs. The NVL72 designation refers to rack-scale systems in which 72 GPUs are interconnected with NVLink, a design approach that has moved AI infrastructure from individual servers toward entire racks engineered as single units, with corresponding increases in the power, cooling and networking demands placed on data centers.

    Sources

    Source: CoreWeave brings up multi-rack NVIDIA Vera Rubin NVL72 cluster — Light Reading, September 16, 2026: CoreWeave announces a multi-rack Vera Rubin NVL72 cluster and new AI Object Storage capabilities.

  • CoreWeave’s $2.6B AI Loan Prices Risk Far Above Nebius’s Debt

    CoreWeave’s $2.6B AI Loan Prices Risk Far Above Nebius’s Debt

    TL;DR · 30-second read

    The Short Version

    Two companies that rent out powerful computers for artificial intelligence recently borrowed money, and one is paying far more for it.

    CoreWeave took a $2.6 billion loan that costs 5.5 percentage points above a standard benchmark interest rate. Its rival Nebius borrowed $775 million at just 2.5 points above that same benchmark.

    The difference comes down to what backs each loan. Nebius tied its loan to one financially strong customer’s contract, while CoreWeave’s covers many customers. If demand for artificial intelligence cools, cheaper borrowing could become a real advantage.

    Dealroom reported that CoreWeave closed a $2.6 billion delayed-draw term loan in August, priced at Term SOFR plus 5.50% and maturing in September 2031, with $1.2 billion drawn so far. That margin is three percentage points above the SOFR plus 2.50% Nebius Group secured on a $775 million facility in July. Nebius has described that facility as backed by deployed GPU infrastructure and cash flows from an investment-grade customer.

    Separately, Nebius closed $5.75 billion of convertible notes in August. In a Form 6-K filed September 8, it announced that Palantir had named it a preferred sovereign AI infrastructure partner.

    Executive Summary

    Two of the best-known neoclouds, the specialist providers that rent GPU computing for artificial intelligence, have borrowed against their buildouts within weeks of each other on very different terms. CoreWeave’s $2.6 billion facility carries a 5.50% margin over the Secured Overnight Financing Rate (SOFR), the US benchmark for floating-rate loans. Nebius’s smaller $775 million facility carries 2.50%.

    The gap does not show lenders favoring one company’s AI prospects over the other’s. It shows lenders pricing the specific collateral and customer contracts behind each loan. Nebius’s loan sits on a single contracted deployment with an investment-grade customer, meaning one with a strong credit rating. CoreWeave’s loan finances committed deployments across a broader customer pool and gives the company more flexibility, and it pays more for that flexibility.

    This matters beyond the two companies. AI infrastructure is being built largely with borrowed money. When credit markets start pricing each buildout on its own contracts and collateral, the cost of capital becomes a competitive variable, and it could separate winners from strugglers if GPU rental prices or utilisation weaken.

    Same Industry, Very Different Price of Money

    Floating-rate corporate loans charge a benchmark rate plus a fixed margin, or spread. The benchmark here is SOFR, which tracks the cost of overnight borrowing secured by US Treasuries. Because both CoreWeave and Nebius pay SOFR, the spread is the part that reflects how lenders view the risk. CoreWeave’s 5.50% is 2.2 times Nebius’s 2.50%, a gap of three percentage points. Describing it as ‘triple’ is loose shorthand. Once the shared benchmark is added back, the difference in all-in borrowing cost is proportionally smaller, though still substantial.

    The structure also matters. A delayed-draw term loan lets the borrower tap committed funds in stages as spending arrives, which suits GPU deployments that come online in phases. CoreWeave has drawn $1.2 billion of its $2.6 billion. As simple arithmetic before fees, three extra percentage points on the drawn balance comes to roughly $36 million a year. On the full facility it would be roughly $78 million a year.

    Collateral and Contracts, Not Confidence, Set the Spread

    Nebius’s facility resembles project finance. It is secured by hardware already deployed and by cash flows from an investment-grade customer, and Nebius said it covers more than 100% of that deployment’s capital expenditure. Lenders are underwriting a known asset with a known payer, which is why the margin is low. The model works best if Nebius can repeat it, and the company points to more than $40 billion of additional customer commitments as the pipeline for doing so.

    CoreWeave’s loan spreads risk across a broader set of committed deployments and runs to September 2031. That breadth gives CoreWeave room to serve shorter contracts and higher-margin enterprise work that a single-customer structure would not accommodate. The trade-off is a higher hurdle rate, the minimum return each financed GPU must earn before it adds value for shareholders. Neither structure is inherently better. One buys cheap money through narrow contracts, and the other pays for flexibility.

    Capital Hunger on Both Sides of the Comparison

    Neither company is borrowing from a position of surplus. Nebius’s $5.75 billion convertible note sale in August dwarfs its $775 million facility. Convertible notes are debt that investors can later swap for shares, so they carry potential dilution. Nebius is also adding demand-side relationships. Its September 8 filing describes Palantir naming Nebius its preferred sovereign AI infrastructure partner and the two working to deploy new capacity, including modular data centers at sites where power is already available. That filing disclosed no financial terms, contracted capacity or timeline.

    Investor positioning is mixed. CoreWeave’s hedge fund holder count rose to 71 in the second quarter of 2026 from 63, even as Magnetar Capital trimmed its stake 24% to 52,062,927 shares. Nebius’s count rose to 86 from 60, while Orbis Investment Management cut 22% to 6,494,719 shares. On August 31, 17.89% of Nebius’s float was sold short. With billions of convertible debt outstanding, part of that short interest likely comes from convertible holders hedging their positions rather than outright bets against the company.

    Why the Pricing Gap Could Matter More If Demand Cools

    While AI compute is scarce and rental prices hold, a three-point spread difference is easily absorbed. The risk appears when utilisation or pricing slips, a concern that resurfaced after September 14. Expensive debt then consumes a larger share of revenue, and strong top-line growth can coexist with weak returns to shareholders.

    The likely beneficiaries are providers that can attach financing to long contracts with highly rated customers, and the lenders who have learned to tell those deals apart. Providers relying on broader, flexible facilities are not disadvantaged today, but they carry more exposure to a downturn. If lenders keep pricing buildouts deal by deal, a lower cost of capital becomes a durable advantage that is hard for rivals to copy quickly.

    Background

    CoreWeave and Nebius are leading examples of neoclouds, companies that build and rent out GPU computing capacity for training and running artificial intelligence models. CoreWeave, which listed on Nasdaq in 2025, is US-based. Nebius Group, listed on Nasdaq as NBIS and headquartered in Amsterdam, is led by founder and CEO Arkady Volozh and was formed from the international businesses of the former Yandex N.V.

    Both companies need large amounts of capital, because GPUs, power and data center space must be paid for well before customers’ rental payments arrive. That has made debt markets central to the sector, and lenders have increasingly looked at the specific contracts and collateral behind each deal rather than lending against the AI growth story as a whole.

    Sources

    Source: CoreWeave draws $2.6B AI loan at SOFR +5.5%, triple Nebius’s rate (Dealroom), comparing CoreWeave’s August term loan with Nebius’s July facility.

    Primary sources: Nebius Group N.V. Form 6-K filed 2026-09-08; Exhibit 99.1 to 6-K filed 2026-09-08: Palantir and Nebius partner to deliver a complete sovereign AI stack to Palantir customers.

  • Everyone’s Watching New AI Silicon. CoreWeave Found 19.8% in Racks It Already Had.

    Everyone’s Watching New AI Silicon. CoreWeave Found 19.8% in Racks It Already Had.

    TL;DR · 30-second read

    The Short Version

    CoreWeave rents out computers that run artificial intelligence. On September 16 it published results from an independent industry test and said its machines handled AI requests faster than those of rival rental companies.

    The more interesting result is quieter: on machines it installed months ago, it got about 20 percent more work out of each chip purely by improving its own software. No new hardware.

    That matters because those chips cost a fortune. Squeezing more out of the ones you own, instead of buying more, is where the profit in this business actually sits.

    CoreWeave on September 16, 2026 published its submissions to MLPerf Inference v6.1, the industry benchmark round released the same day by MLCommons. Writing on the company blog, Shadi Saba said CoreWeave entered the Datacenter Closed division’s Available category across four NVIDIA platforms — HGX B200, HGX B300, GB200 NVL72 and GB300 NVL72 — and four model families: Qwen3-VL-235B-A22B, DeepSeek-R1-671B, GPT-OSS-120B and Llama 2 70B. Headline figures include 16,635 tokens per second per GPU on GPT-OSS-120B in the offline scenario, more than 1.19 million tokens per second from a single GB300 NVL72 rack, and 944,902 tokens per second on Llama 2 70B in the server scenario. NVIDIA, in a blog post of its own, announced that its next-generation Vera Rubin NVL72 made its MLPerf debut in the same round.

    One day later, on September 17, CoreWeave filed an 8-K disclosing a proposed $3.0 billion offering of convertible senior notes due 2033, with an option for a further $500 million, alongside an Equity Distribution Agreement covering at-the-market share sales and collared forward sales through eleven banks. The two disclosures, a day apart, describe the same business from opposite ends: what the fleet produces, and what it costs to keep building.

    Executive Summary

    CoreWeave’s MLPerf claims break into two categories, and they are not equally interesting. The absolute throughput records — 1,196 queries per second on the multimodal Qwen3-VL model, over 1.16 million tokens per second on GPT-OSS-120B from one rack — are a function of running NVIDIA’s newest shipping silicon, GB300 NVL72, at scale. Any operator with the same racks and the same power envelope is in that fight.

    The second category is harder to replicate. CoreWeave says that comparing its 72-GPU v6.1 submission against its 64-GPU v6.0 submission on GB200 NVL72, derived per-GPU server throughput on DeepSeek-R1-671B rose 19.8% in the five months between rounds. That gain came from serving-stack tuning, scheduling and operations, not from a new chip. If it holds in production, it changes the revenue a depreciating asset can generate across its life — which is the number that underwrites everything else.

    That includes the financing. The 8-K filed September 17 shows CoreWeave raising $3.0 billion in converts on top of an existing stack of senior notes carrying coupons from 8.500% to 9.750%. Capital that expensive is serviced by output per installed accelerator, and software gains on hardware already paid for are the cheapest output there is.

    The 19.8% That Didn’t Require a New Chip

    Strip the benchmark round down and one number does real work. CoreWeave reports that its DeepSeek-R1-671B per-GPU server throughput on NVIDIA GB200 NVL72 rose 19.8% between MLPerf Inference v6.0 and v6.1 — five months apart — through changes to its serving stack, scheduling and operations rather than new hardware. The mechanisms it names are specific: topology-aware scheduling through SUNK, which pins an inference workload inside a single high-bandwidth NVLink domain so that tokens moving between accelerators never leave the fastest interconnect in the rack, and fleet-level health monitoring through Mission Control. For a mixture-of-experts model like DeepSeek-R1 — an architecture that routes each token through a subset of specialised sub-networks, so traffic between chips is constant — placement is not a detail. It is the workload.

    Who this affects is everyone financing GPUs on a multi-year schedule. An accelerator bought in early 2026 has a fixed capital cost and a depreciation clock that does not care how efficiently it runs. If per-GPU output rises roughly a fifth without capex, the cost per token served falls by a comparable amount across the remaining life of that asset, and the residual economics of the installed base improve rather than decay. That is a materially different asset story from the one where value arrives only with the next generation of silicon.

    The claim deserves its caveats stated plainly, and CoreWeave states some of them itself. Per-GPU throughput is a derived metric — total system throughput divided by accelerator count — and the company notes it is not verified by MLCommons. The comparison also spans two differently sized submissions, 64 accelerators in v6.0 against 72 in v6.1, so scaling behaviour and configuration changes are folded into the same percentage as the software work. The direction is credible and the mechanism is named; the precision of the figure is CoreWeave’s own arithmetic, not the benchmark’s finding.

    The Unit of Account Moved From the Chip to the Rack

    The GB300 NVL72 is not a server. It is a rack in which 72 accelerators share one NVLink domain, meaning they address each other at memory-like speeds rather than over conventional networking. CoreWeave’s results read accordingly: over 1.16 million tokens per second in the server scenario and over 1.19 million offline on GPT-OSS-120B from a single rack, 944,902 and 1,136,100 tokens per second on Llama 2 70B, and 1,196 queries per second on Qwen3-VL-235B-A22B. The company’s GB200 NVL72 submission reached 901,058 tokens per second on GPT-OSS-120B in the server scenario — the same architecture, one generation back, and a visible step down.

    The consequence lands on the people who build and power facilities rather than the people who buy models. When the smallest sensible deployment unit is a rack drawing far more than a legacy hall was designed for, throughput per rack becomes a proxy for how much revenue a given square metre and a given megawatt can carry. Operators comparing a rack-scale NVL72 deployment against clusters of HGX B200 or B300 servers — which CoreWeave also submitted, and which it says led all HGX entries from cloud providers — are really comparing two different power-density and liquid-cooling commitments, not two spec sheets.

    Available Is the Word Doing the Most Work

    CoreWeave’s submissions sit in the Datacenter Closed division’s Available category. Closed means every entrant runs the same reference implementation, so results are comparable rather than a contest in custom optimisation; Available means the system can actually be obtained. That second qualifier matters this round, because NVIDIA used the same release to announce that its next-generation Vera Rubin NVL72 made its MLPerf debut. A benchmark headline generated by a platform that is debuting is not a price a buyer can procure against today; CoreWeave’s four platforms are all Blackwell and Blackwell Ultra, the generations in production now.

    For anyone running a procurement process, the practical read is to check three fields before treating any v6.1 result as a quote: which silicon generation produced it, whether the entry sits in the Available category, and whether the metric quoted is a submitted number or a derived per-GPU figure. CoreWeave also stresses that it ran on the same production clusters and images customers get rather than a benchmark-tuned rig — a genuinely meaningful distinction, though one the benchmark itself does not audit.

    Throughput Is the Collateral Story

    The day after the benchmark post, CoreWeave’s 8-K disclosed a proposed $3.0 billion offering of convertible senior notes due April 1, 2033, sold under Rule 144A to qualified institutional buyers, with an option for up to $500 million more and capped call transactions to limit dilution on conversion. A concurrent Equity Distribution Agreement adds at-the-market share sales and collared forward sales through a syndicate that includes Deutsche Bank, Goldman Sachs, J.P. Morgan, Morgan Stanley and Citigroup. Interest and conversion terms, the release states, are set at pricing.

    The comparison worth making is internal. CoreWeave’s own press release lists the existing obligations the new notes will rank alongside: senior notes at 9.250%, 9.000%, 9.750%, 9.625% and 8.500%, and earlier convertibles at 1.75%. Converts price far below straight debt because the buyer is paid partly in equity optionality rather than cash coupon — cheaper carry now, potential dilution later, which the capped calls are bought to blunt. For a business whose costs are dominated by accelerators, the servicing capacity for any of it comes back to tokens per installed GPU.

    That is the honest link between two documents filed a day apart, and it is a structural one rather than a causal one: a benchmark result is a performance claim, not a revenue disclosure, and nothing in the filings ties the offering to the MLPerf round. But the neocloud model runs on both halves at once — demonstrable output per accelerator to justify the capital, and access to capital to buy more accelerators. Software gains of the kind CoreWeave reported on already-deployed GB200 racks are the half that does not require either.

    Background

    CoreWeave, based in Livingston, New Jersey and listed on Nasdaq as CRWV, is a specialist cloud provider built around NVIDIA accelerators for AI training and inference — a category often called a “neocloud,” distinguished from general-purpose hyperscalers by its narrow focus and its heavy reliance on debt and equity markets to fund hardware. The company was recently named a Visionary in the Gartner Magic Quadrant for Cloud AI Infrastructure.

    MLPerf, administered by the MLCommons consortium, has become the closest thing the industry has to an audited comparison of AI serving performance, which is why vendors submit and why the categories matter. Inference — running a trained model to answer requests — is now the larger and more cost-sensitive half of the AI workload, because a model is trained once but served continuously. That shift is what makes tokens per accelerator, rather than raw chip specifications, the metric operators and their lenders increasingly watch.

    Sources

    Source: MLPerf® Inference v6.1 Results: CoreWeave Leads Providers — CoreWeave’s September 16, 2026 post detailing its Datacenter Closed, Available submissions across four NVIDIA Blackwell platforms. See also NVIDIA Vera Rubin NVL72 Delivers Leading Performance in MLPerf Inference v6.1 Debut, NVIDIA’s account of the same benchmark round.

    Primary sources: CoreWeave, Inc. Form 8-K filed September 17, 2026 (Items 7.01 and 8.01, disclosing the proposed convertible notes offering and the Equity Distribution Agreement); Exhibit 99.1 — CoreWeave Announces Proposed $3.0 Billion Convertible Senior Notes Offering; and Exhibit 99.2 — CoreWeave Investor Presentation, September 2026. Full benchmark tables are published by MLCommons at mlcommons.org.

  • Backblaze’s $335M CoreWeave Deal and AI’s Storage Layer

    Backblaze’s $335M CoreWeave Deal and AI’s Storage Layer

    TL;DR · 30-second read

    The Short Version

    A company best known for cheap online backup has agreed to store data for one of the largest artificial intelligence computing firms. Backblaze will hold data for CoreWeave under a deal worth $335 million over five years.

    The volume is measured in exabytes. One exabyte is a billion gigabytes — about a million laptop hard drives. Backblaze shares jumped roughly 25% on the news.

    Why it matters: teaching a computer to answer questions requires mountains of data, and that data has to sit somewhere cheap while the expensive machines do the thinking.

    Backblaze has signed a $335 million, five-year agreement to provide storage capacity to CoreWeave, the specialist artificial-intelligence cloud provider, according to Stocktwits, which reported the deal and the market reaction. Backblaze will expand CoreWeave’s storage to “multi-exabyte” levels — an exabyte being one billion gigabytes — while, in the companies’ framing, preserving CoreWeave’s high-performance resources for heavy AI workloads. Existing CoreWeave AI Object Storage customers can use the expanded capacity immediately without changing any code.

    Backblaze shares (BLZE) rose about 25% on the session, putting the stock on track for its highest level in nearly seven months; CoreWeave (CRWV) fell more than 5% the same day amid a broader technology-led selloff. “Storage is the foundation every AI workflow is built on — without it, even the world’s most powerful compute sits idle,” said Gleb Budman, Backblaze co-founder and chief executive, who described the arrangement as working with CoreWeave “on elements of their storage environment.”

    Executive Summary

    The headline number is modest by AI-infrastructure standards — $335 million spread across five years — but the structure of the deal is more interesting than its size. CoreWeave, whose business is renting graphics processors for AI training and inference, is buying bulk capacity storage from an outside specialist rather than building the tier itself. That is a make-or-buy decision at a layer of the stack most market commentary ignores.

    For Backblaze, a company whose brand was built on low-cost backup and archival storage, this is a repositioning: from consumer and small-business backup vendor to a capacity supplier inside an AI cloud. The “no code changes” detail matters here. It implies interface compatibility with what CoreWeave customers already call, which is the practical precondition for any storage substitution at scale — customers will not rewrite data pipelines to save on cold storage.

    The broader signal is about AI economics. Accelerated compute is scarce and expensive; the bytes that feed it are abundant and cheap, but there are an enormous number of them. Training datasets, model checkpoints, inference logs and retrieval corpora all accumulate faster than they are deleted. Separating that mass from the fast storage sitting next to the processors is a straightforward cost-optimisation move, and it creates a real addressable market for vendors who are good at cheap, durable bulk capacity.

    The Cheap Tier Nobody Prices In

    Discussion of AI infrastructure economics tends to stop at the accelerator: how many chips, at what price, drawing how much power. Storage is treated as a rounding error. Inside an actual training or inference pipeline it is not. Raw training corpora, cleaned and tokenised derivatives, model checkpoints written at intervals through a run, evaluation outputs, and the document collections used for retrieval all have to live somewhere, and most of them are read rarely after they are written.

    Keeping that material on the high-performance storage physically attached to accelerator clusters is expensive and, worse, wasteful: fast flash next to a processor exists to keep that processor busy, not to warehouse last quarter’s checkpoints. The design pattern the industry has converged on is tiering — a small, very fast layer for active work, and a large, slow, cheap layer for everything else. Budman’s framing of the deal, preserving CoreWeave’s high-performance resources for heavy workloads, is a description of exactly that split.

    Object storage is the natural home for the bulk tier. It is accessed over the network by simple read-and-write requests rather than presented as a filesystem, which makes it cheap to scale to exabyte volumes and easy to address from distributed jobs. It is also, crucially, close to a commodity — which is what makes an outside supplier viable in the first place.

    What $335 Million Actually Buys

    Spread evenly, $335 million over five years is roughly $67 million a year of contract value. Against CoreWeave’s overall infrastructure spending that is small. Against Backblaze’s business it is plainly significant — significant enough that the market repriced the stock by a quarter in a session. That asymmetry is the whole story of this layer: bulk storage is a rounding error to the buyer and a franchise to the seller.

    The economics on Backblaze’s side depend on details the announcement does not settle. Multi-exabyte capacity has to be bought, racked, powered and cooled before it earns anything, so the shape of the cash flows — how much capital goes out ahead of revenue, and how quickly CoreWeave’s consumption ramps — determines whether this is an accretive contract or an expensive one to serve. Committed minimums versus consumption-based drawdown makes the difference between a revenue floor and a ceiling.

    There is also concentration risk to weigh honestly. A single customer of this size stabilises a revenue base and simultaneously creates dependence on one counterparty’s growth, pricing leverage and renewal decision in 2031. Neither risk is disqualifying; both belong in any serious model of what the deal is worth.

    Why a Neocloud Buys Instead of Builds

    Specialist AI clouds — the so-called neoclouds — have generally competed on one thing: getting scarce accelerators into customers’ hands quickly. Their capital and engineering attention goes to procurement, power, networking and cluster orchestration. Building and operating an exabyte-scale object store is a different discipline with different failure modes, and it earns thin margins even when done well.

    Choosing to buy that layer is therefore a rational focus decision rather than a concession. It also raises the question of what else in the stack is a candidate for the same treatment. Every neocloud faces the same trade-off between vertical integration, which captures margin, and partnership, which conserves capital and shortens time to market. Deals like this one are how that boundary gets drawn in public.

    The competitive read for buyers of AI capacity is mixed. More supply of cheap capacity behind a compatible interface is good for price and portability. But if bulk storage inside AI clouds consolidates around a small number of wholesale suppliers, the long-run pricing dynamic will look less like commodity competition and more like the hyperscale storage market it is meant to undercut.

    Reading the Market Reaction Carefully

    Backblaze stock rose about 25% on the day and has gained more than 74% so far in 2026 and roughly 59% over twelve months, outperforming the S&P 500. Retail sentiment on Stocktwits shifted from neutral to bullish on high message volume, with users on the platform describing the company as an emerging contender in AI cloud storage.

    A sharp move on a contract announcement reflects a revision to expectations, not a settled valuation. The useful test over the next several quarters is mundane: does reported revenue reflect the ramp, what does it cost in capital and gross margin to serve, and does the CoreWeave relationship bring adjacent business or remain a single large account. Notably, CoreWeave shares fell more than 5% the same day in a broad technology selloff — a reminder that the deal is material to one party’s outlook far more than the other’s.

    The market backdrop is genuinely large. Grand View Research valued the global AI data centre market at $147.3 billion in 2025 and projects $810.6 billion by 2033, a compound annual growth rate of 23.9% from 2026, with cloud platforms as a primary driver. Third-party forecasts of that horizon carry wide error bars, but the directional point — that capacity demand compounds — is consistent with what operators are actually building.

    Background

    Backblaze built its business on undercutting incumbent cloud storage on price, first with consumer and business backup and later with an object storage service aimed at developers and media workflows. Its pitch has consistently been cost per terabyte and simplicity rather than breadth of platform services — a positioning that historically kept it outside the hyperscale conversation, and which now maps neatly onto the bulk-capacity tier of AI pipelines.

    CoreWeave belongs to a newer category: providers that emerged to supply accelerated computing at a time when demand for it outran what general-purpose clouds could deliver. These operators grew rapidly by concentrating on procurement, power and cluster performance. As their platforms matured, the surrounding services — storage, networking, data movement — became decisions to make rather than assumptions to inherit, and this agreement is one visible outcome of that process.

    Sources

    Source: BLZE Stock Surges After Storage Deal With CoreWeave – Retail Bulls Call Backblaze An Emerging AI Cloud Contender — Stocktwits reporting on the $335 million, five-year storage agreement, the share price reaction and retail sentiment. Market sizing cited from Grand View Research.

  • CoreWeave’s $6 Billion Lancaster Bet on Powered Land

    CoreWeave’s $6 Billion Lancaster Bet on Powered Land

    TL;DR · 30-second read

    The Short Version

    • The artificial intelligence company CoreWeave is building a $6 billion computing campus in Lancaster County, Pennsylvania — farm country about halfway between Philadelphia and Harrisburg.
    • What buyers are really chasing is not the dirt but the electricity. Land that already has a big power hookup has sold for more than $8 million an acre in some markets.
    • Farmland nearby drew more than 50 purchase offers in a single year. Some owners are selling; others have legally locked their land to farming only.
    • Why care? These campuses use enough electricity to run a small city, and states are now arguing over who pays for the wires.

    StartupHub.ai reported, drawing on CNBC’s segment Inside The Real Estate Rush Behind AI, that CoreWeave is building a $6 billion data center campus in Lancaster County, Pennsylvania, less than 20 miles from Mount Joy farmland that attracted more than 50 purchase offers in a single year. Lancaster’s draw, according to that account, is a reliable electrical grid, state incentives and a Mid-Atlantic position within reach of several major metropolitan areas.

    The project lands in the middle of a national scramble for what the industry calls powered land — sites that already have, or can plausibly obtain, large-scale grid connections. Investors spent nearly $6 billion on U.S. land earmarked for future data centers in the first half of 2026, up 79% year over year and roughly 27% of all spending on development sites. The United States already hosts more than 4,700 data centers.

    Executive Summary

    For CoreWeave, the significance of Lancaster is not the headline number but the change in posture it implies. The company grew by renting: it filled other operators’ buildings with graphics processing units — the specialized chips that train and run artificial intelligence models — and sold that capacity by the hour. A campus of its own, at $6 billion of committed capital, moves it a long way toward being the landlord rather than the tenant.

    That shift concentrates a set of risks the company previously paid someone else to carry. Whoever owns the site owns the interconnection queue — the waiting list utilities maintain for connecting large new loads to the grid — along with the substation costs, the zoning fight, the water permit and the timing risk if energization slips. It also captures the margin that a third-party operator would otherwise take, and gains control over how quickly the next building goes up.

    The wider market context is unforgiving on both counts. Amazon, Alphabet, Microsoft and Meta spent over $167 billion combined in the second quarter of 2026, up 79% year over year, which sets the pace CoreWeave has to match to stay relevant. At the same time, Pennsylvania Governor Josh Shapiro has said that of more than 100 data center proposals in the state, only five hold permits to operate — a reminder that announced capital and delivered capacity are separated by a queue that is currently very long.

    From Renting Racks to Owning the Substation

    The economics of a specialist artificial intelligence cloud have always been awkward. Chips depreciate on a schedule measured in a handful of years; buildings and electrical infrastructure amortize over decades. A company that leases space pays a spread to the landlord and inherits the landlord’s delivery schedule, which in a market this tight is the binding constraint on growth. Owning the campus removes the spread and, more importantly, removes the queue-jumping problem: you are no longer waiting for someone else to decide which tenant gets the next 50 megawatts.

    The trade is that development risk stops being a line item in someone else’s business. Site work, transformer procurement, interconnection studies and permitting all move onto the builder’s balance sheet, and none of them are hourly-billable until the first customer racks go live. The release notes that near-term power shortages already add a year simply to assess distribution to a site — a delay that costs nothing to a tenant and a great deal to an owner carrying $6 billion of committed spend.

    Scale gives some perspective on the commitment. The campus’s stated cost is roughly equal to what every investor in the United States combined spent buying data center land in the first half of 2026 — different categories of spending, since one figure is construction and the other is raw acreage, but a useful marker of how much a single project now weighs against an entire national land market.

    Powered Land Is the Asset; Dirt Is the Wrapper

    The premium in this market attaches to electricity, not acreage. Sites with large power access have exceeded $8 million per acre in Northern Virginia and the Northeast, and the industry needs an estimated 40,000 acres of such land to meet growth through 2030. That is the whole thesis in one sentence: land is abundant, grid capacity is not, and the price gap between the two is where developers are making their money.

    It is also where the risk sits. Speculators frequently buy before they know whether permits or power will arrive, and the release is blunt about the consequence — if power does not materialize, the land stays illiquid. Boise Valley values have reached 15 times pre-pandemic levels on the strength of Micron’s two fabrication plants and a large Meta campus 10 to 15 miles away, and brokers there expect much of the remaining farmland to convert to industrial use within five to seven years. Those expectations are underwritten by an assumption about utility delivery, not by signed interconnection agreements.

    The speculative posture is visible in how developers are building. Idaho developer Mike Adler, who controls 1,000 acres and $1 billion of real estate and is expanding from 1.9 million to 4.6 million square feet, says 62% of his current Boise Valley projects are spec builds — constructed before any tenant signs a lease. That is a rational bet when demand outruns supply. It is also the classic structure through which a slowdown in one sector transmits into a regional property market.

    The Rules Are Being Written While the Concrete Cures

    Two governors have moved in the same direction from different angles. Shapiro signed an August executive order intended to block cost pass-throughs to residents and limit water use, alongside his observation that only five of more than 100 Pennsylvania proposals hold operating permits. In Texas, Governor Greg Abbott has directed regulators to require data centers to fully fund their own electric infrastructure. Both are answers to the same political question: when a single campus draws load comparable to a city, who pays for the wires?

    The direction of travel matters for anyone underwriting these projects. U.S. electricity consumption is projected to rise nearly 2% annually through 2030 — more than twice the rate of the prior decade — and the incremental demand is concentrated in exactly the large-load interconnections regulators are now scrutinizing. A rule that shifts substation and transmission costs onto the developer does not stop projects; it raises the effective cost per megawatt and rewards operators with the balance sheet to absorb it. That favors the largest builders and squeezes thinly capitalized speculators holding options on unpowered ground.

    What the Boom Looks Like at Ground Level

    Around Mount Joy, the conflict is a zoning question with a long horizon. Bobby Thompson and Michelle Kennedy, who have farmed there for decades, placed an easement on their land — a permanent legal restriction limiting it to farming — while neighbors have applied to rezone for an industrial park. Both are legitimate uses of private property rights, and the outcome will be settled parcel by parcel rather than by any single decision.

    The downstream economics are real and already measurable. Air Filter Superstore, which supplies the kind of filtration these facilities consume in volume, grew from under $10 million in annual revenue before 2020 to more than triple that, expanding from 7,000 square feet to over 30,000 after investing about $1 million. Multiply that across electrical contractors, mechanical suppliers and site trades and you have the local case for approval. Micron’s plan to invest $250 billion in the United States through 2035 suggests the industrial demand behind those suppliers is not a one-year phenomenon.

    The caution is that supplier expansion, spec construction and land repricing are all leveraged to the same assumption — that hyperscaler capital expenditure keeps compounding. A 79% year-over-year increase from the four largest spenders is a powerful signal, but it is a growth rate, not a floor. Communities weighing tax base against farmland, and vendors sizing warehouses against forecast orders, are implicitly taking a view on how long that rate holds.

    Background

    CoreWeave began as a cryptocurrency mining operation before repurposing its expertise in dense, power-hungry computing into a cloud business built on graphics processing units. It listed publicly in 2025 and scaled principally as a tenant, installing its hardware in facilities owned by third-party data center operators — an asset-light approach that let it grow quickly without carrying construction risk.

    The market it operates in has been reshaped by the electricity constraint. As artificial intelligence workloads pushed rack power densities far beyond traditional cloud computing, the scarce input stopped being land or even chips and became grid interconnection. U.S. electricity consumption is projected to rise nearly 2% annually through 2030, more than twice the previous decade’s rate, and that demand is concentrating in a handful of regions with available capacity — pulling multibillion-dollar industrial projects into farm counties in Pennsylvania, Idaho and elsewhere, and prompting state governments to rewrite the rules on who pays for the grid upgrades they require.

    Sources

    Source: CoreWeave $6 billion Lancaster data center — StartupHub.ai’s account of the Lancaster campus and the national competition for powered land, drawing on CNBC’s segment Inside The Real Estate Rush Behind AI.

  • Synergy: Neocloud Revenues Growing 200%+ a Year, Headed for $180B by 2030

    Synergy: Neocloud Revenues Growing 200%+ a Year, Headed for $180B by 2030

    Synergy Research Group reported on August 17, 2026 that “neoclouds” — the emerging tier of specialized GPU cloud providers built for AI workloads — are currently growing revenues at more than 200% per year. On that trajectory, Synergy forecasts the segment will reach $180 billion in annual revenues by 2030.

    Executive Summary

    Synergy Research Group, a market intelligence firm that has tracked cloud and data center markets for decades, put a striking pair of numbers on one of the fastest-moving corners of the infrastructure industry: neocloud providers are more than tripling their revenues each year, and the category is projected to become a $180 billion market by 2030.

    The forecast matters because it treats neoclouds not as a temporary arbitrage on scarce GPUs, but as a durable market tier alongside the hyperscale clouds. If Synergy is right, a business model that barely existed three years ago will, within four years, rival the size of the entire global colocation industry — with all the capital, power, and data center demand that implies. It is worth noting the syndicated item we reviewed carries the headline figures but not Synergy’s full methodology, so the underlying assumptions deserve scrutiny alongside the projection itself.

    What a Neocloud Is — and Why the Category Exists

    “Neocloud” is the industry’s shorthand for cloud providers built specifically around GPU compute for artificial intelligence — renting out clusters of accelerators for model training and inference rather than offering the sprawling general-purpose service catalogs of AWS, Microsoft Azure, or Google Cloud. Commonly cited players in the category include CoreWeave, Lambda, Nebius, and Crusoe, though Synergy’s specific inclusion list is not visible in the syndicated item.

    The category exists because AI demand outran what the traditional clouds could supply. Training frontier models requires dense, tightly networked GPU clusters, exotic power and cooling footprints, and pricing models closer to industrial capacity contracts than to on-demand virtual machines. Specialists that could secure GPUs, power, and data center space quickly found a seller’s market waiting for them.

    The Economics Behind 200% Growth

    Growth above 200% per year is extraordinary, but the arithmetic behind it is straightforward: the segment started from a small base, and demand for AI compute currently exceeds supply. When capacity sells out before it is built, revenue growth tracks how fast a provider can energize new data center capacity — which is why the neocloud story is inseparable from the power and data center construction booms.

    The harder question is margin durability. Neocloud economics rest on expensive, fast-depreciating hardware, heavy debt financing in many cases, and — for several prominent players — revenue concentrated in a small number of very large AI customers. A $180 billion revenue projection says the market will be big; it does not by itself say the businesses in it will be uniformly profitable. Investors should distinguish between the size of the pie and the quality of any individual slice.

    Winners, Losers, and the Hyperscaler Question

    For data center operators, utilities, and connectivity providers, the forecast is almost unambiguously bullish: neoclouds are among the largest lessees of wholesale data center capacity and the most aggressive buyers of power. A tier growing toward $180 billion in revenue implies sustained demand for the physical layer beneath it — sites, substations, fiber, and cooling.

    For the hyperscalers, the picture is more nuanced. Neoclouds are simultaneously competitors for AI workloads and, in some well-publicized arrangements across the industry, suppliers of capacity to the hyperscalers themselves. Whether the big clouds ultimately reabsorb this demand as their own GPU fleets scale, or the neocloud tier keeps a permanent structural advantage in speed and specialization, is the central competitive question the next few years will answer.

    Can the Curve Hold to 2030?

    Extending any 200% growth rate for years produces implausible numbers, and Synergy’s own forecast implies significant deceleration: a market compounding at 200% would blow far past $180 billion by 2030 from almost any plausible base. Read properly, the projection assumes today’s hypergrowth cools into merely strong growth — a reasonable but assumption-laden path.

    The risks to the curve are the familiar ones for AI infrastructure: whether enterprise AI spending keeps converting into paid compute at current rates, whether power availability constrains buildouts, how quickly GPU generations depreciate, and whether customer concentration turns any single buyer’s pullback into a segment-wide shock. None of these invalidate the forecast; all of them are the difference between the projection and the outcome.

    Background

    The neocloud category rose to prominence after 2023, when generative AI demand created acute scarcity in GPU compute and a wave of specialists — several of them former cryptocurrency miners repurposing power-rich sites — pivoted to renting AI capacity. The segment has since attracted tens of billions of dollars in capital and become one of the largest sources of demand in the data center leasing market. Synergy Research Group, which has long published the benchmark market-share data for cloud infrastructure services, tracking the rise of AWS, Microsoft, and Google, now treats this GPU-specialist tier as a distinct market worth forecasting in its own right — itself a signal of how the AI buildout is restructuring cloud economics.

    Source: Neoclouds Currently Growing by Over 200% per Year; Will Reach $180 Billion in Revenues by 2030 — Synergy Research Group, a market forecast for the GPU-specialist cloud segment published August 17, 2026.

  • CoreWeave Named Visionary in Gartner’s 2026 Cloud AI Quadrant

    CoreWeave Named Visionary in Gartner’s 2026 Cloud AI Quadrant

    CoreWeave announced on July 6, 2026 that it has been named a Visionary in Gartner’s 2026 Magic Quadrant for Cloud AI Developer Services. The recognition places the GPU-focused cloud provider on one of the industry’s most closely watched analyst grids alongside larger hyperscalers.

    Executive Summary

    CoreWeave, best known for renting out large fleets of Nvidia GPUs to AI labs and enterprises, has picked up a Visionary designation in Gartner’s 2026 Magic Quadrant for Cloud AI Developer Services. Gartner’s Magic Quadrant is a widely referenced analyst report that plots vendors on two axes — completeness of vision and ability to execute — and Visionaries score high on vision but are typically still building out execution scale.

    The placement matters because Cloud AI Developer Services is a category traditionally dominated by the three hyperscalers, whose managed AI platforms bundle models, training frameworks, and deployment tools. CoreWeave earning a named spot signals that its pitch — purpose-built GPU infrastructure with a developer-facing stack — is being taken seriously by procurement teams that historically default to AWS, Azure, or Google Cloud.

    Why a Visionary Tag, Not a Leader Tag, Is the Story

    Being named a Visionary is a genuine analyst endorsement, but the label carries a specific meaning. In Gartner’s framework, Visionaries understand where a market is heading and often shape it with differentiated technology, but they have not yet demonstrated the operational breadth of the Leaders quadrant. For a company like CoreWeave, that reading fits the public narrative: a GPU specialist that grew explosively during the generative AI wave, but whose managed developer services are newer than the hyperscalers’ decade-old platforms.

    For buyers, the practical translation is that CoreWeave is worth a serious bake-off for AI workloads, particularly training and large-scale inference, without assuming it yet matches AWS or Azure on the breadth of adjacent services like identity, data warehousing, or global compliance tooling.

    The Competitive Frame: Specialist Clouds Versus Hyperscalers

    The Magic Quadrant category itself is worth unpacking. Cloud AI Developer Services covers the tools developers use to build, tune, and deploy AI applications — model APIs, training platforms, MLOps, and increasingly agent frameworks. The hyperscalers compete here with fully integrated stacks. Specialist clouds compete on price-performance for GPU-intensive workloads and, more recently, on time-to-capacity for scarce accelerators.

    Getting graded in the same report as the hyperscalers is a validation of the specialist thesis: that a meaningful share of AI spend will flow to providers optimized specifically for the workload, rather than to general-purpose clouds that also happen to sell GPUs. Whether that share remains large as hyperscaler capacity catches up is the open strategic question.

    What This Does — and Does Not — Prove

    Analyst recognition is a procurement lubricant. Enterprise buyers frequently cite Magic Quadrant placement to justify shortlists, and inclusion can shorten sales cycles materially. In that narrow sense, the designation has real commercial value for CoreWeave beyond the marketing headline.

    What it does not prove is durable margin, customer diversification, or that CoreWeave’s developer-services layer is at feature parity with incumbents. Gartner scores vision and execution against a defined market frame; it does not opine on unit economics, GPU supply contracts, or concentration risk with a small number of very large customers. Readers should treat the placement as one useful signal among several, not as a verdict on the business.

    Background

    CoreWeave began as a niche compute provider and repositioned during the generative AI boom into a specialist cloud focused on large-scale Nvidia GPU deployments, becoming a prominent supplier of training and inference capacity to AI labs and enterprises. It has since expanded into developer-facing services that sit above the raw infrastructure layer.

    Gartner’s Magic Quadrant for Cloud AI Developer Services is one of the industry’s most cited analyst reports for AI platform procurement, historically dominated by the largest hyperscale cloud providers. Inclusion for a specialist cloud reflects the broader shift of AI workloads toward providers optimized specifically for accelerated computing.

    Source: CoreWeave Named a Visionary in 2026 Gartner Cloud AI Report — CoreWeave’s announcement of its placement in Gartner’s 2026 Magic Quadrant for Cloud AI Developer Services.

  • Galaxy’s Helios Phase I Delivers 133 MW of AI Capacity to CoreWeave

    Galaxy’s Helios Phase I Delivers 133 MW of AI Capacity to CoreWeave

    Galaxy announced on July 5, 2026 that it has completed Phase I of its Helios data center campus in West Texas, delivering 133 megawatts (MW) of critical IT load to CoreWeave, the AI-focused cloud provider. Critical IT load refers to the power available to the computing equipment itself — servers and GPUs — as distinct from the total power a facility draws for cooling and other overhead.

    The completion converts a site that began life as a Bitcoin mining campus into dedicated AI infrastructure under Galaxy’s long-term lease arrangement with CoreWeave, one of the most prominent examples of the crypto-to-AI conversion trend reshaping the data center market.

    Executive Summary

    Galaxy, the digital assets and data center infrastructure firm, has finished the first phase of its Helios campus buildout and handed over 133 MW of critical IT load to its anchor tenant CoreWeave. Phase I completion moves the project from promise to delivery: Helios is now an operating revenue-generating AI data center rather than a conversion story on a slide deck.

    The milestone matters beyond Galaxy. Helios is the flagship test case for whether former cryptocurrency mining sites — which come with grid interconnections and power contracts already in place — can be economically retrofitted to the far more demanding standards of AI training and inference infrastructure. Delivering a first phase at this scale suggests the model can work, at least for sites with strong power positions.

    For CoreWeave, the delivery adds substantial contracted capacity at a time when access to powered land and energized shells — not GPUs — is widely seen as the binding constraint on AI cloud growth.

    Why Crypto Sites Became AI Real Estate

    The most valuable asset in data center development today is not land or buildings but secured power: a grid interconnection agreement and the megawatts behind it. Bitcoin mining operators spent the late 2010s and early 2020s locking up exactly that, often in low-cost power markets like West Texas. When AI demand exploded, those interconnections became worth far more serving GPUs than mining rigs, because AI tenants sign long-term leases at data center economics rather than riding volatile crypto margins.

    Galaxy’s Helios campus, acquired from a Bitcoin mining operator, is the highest-profile execution of that arbitrage. The conversion is not trivial — AI facilities require far denser power delivery, liquid or advanced air cooling, and enterprise-grade redundancy that mining sites never needed — but the timeline still beats greenfield development, where new grid interconnection requests can queue for years.

    What 133 MW Actually Buys

    133 MW of critical IT load is a substantial block of capacity by any historical standard — a few years ago it would have ranked among the larger single-tenant deployments in the world. In the AI era it is best understood as a first tranche: large frontier training clusters are increasingly specified in the hundreds of megawatts, and operators including Galaxy have discussed multi-phase expansion at Helios well beyond Phase I.

    Because the load is contracted to a single tenant, the economics resemble a triple-net real estate deal more than a retail colocation business: predictable lease revenue over a long term, with Galaxy carrying development and delivery risk and CoreWeave carrying utilization risk. That structure has become the dominant template for AI data center finance because lenders can underwrite the lease.

    Winners, Losers, and the Competitive Field

    The clearest winners are holders of energized or near-energized power positions — converted mining sites, utilities with spare interconnection capacity, and developers who queued early. CoreWeave benefits by adding capacity faster than greenfield timelines would allow, supporting its competition with hyperscale clouds for AI workloads. The pressure lands on developers still waiting in interconnection queues, and on regions whose grids cannot absorb gigawatt-class requests.

    The open competitive question is durability. Conversion sites tend to sit in remote, power-rich locations, which suits training workloads that tolerate latency. If the market shifts toward inference — which favors proximity to users — the value of remote megawatts could be repriced. Phase I’s completion answers the execution question; it does not settle the location question.

    Background

    Helios began as one of the larger Bitcoin mining campuses in the United States before Galaxy acquired the site and redirected it toward AI and high-performance computing. Galaxy subsequently signed long-term lease agreements making CoreWeave the campus’s anchor tenant, with capacity to be delivered in phases — Phase I, now complete, being the first.

    The conversion sits inside a broader industry shift: as demand for AI compute outran the pace of new grid connections, sites with existing power infrastructure — many of them crypto mining facilities in Texas and the Mountain West — became prime targets for repurposing. Helios is widely watched as the leading proof point for whether that playbook delivers at scale.

    Source: Galaxy Completes Phase I of Its Helios Data Center Campus, Delivering 133 Megawatts of Critical IT Load to CoreWeave — PR Newswire press release, July 5, 2026, announcing Phase I completion at Galaxy’s West Texas AI campus.