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

Author: Deepak Jain

  • Hillsboro’s 120-Day Pause Spares 7 Data Center Sites; AI Builders Face the Rewrite

    Hillsboro’s 120-Day Pause Spares 7 Data Center Sites; AI Builders Face the Rewrite

    TL;DR · 30-second read

    The Short Version

    Hillsboro, Oregon, a Portland suburb, already has 35 data centers. These are giant warehouse-style buildings full of computers that run the internet and artificial intelligence. The city has now voted to stop accepting new proposals for them for 120 days.

    Nothing already running, being built or under review is affected. What changes is the rulebook for whatever comes next.

    The pause also covers big battery sites that store electricity for later. The local power company asked the city to leave those out.

    Oregon Public Broadcasting reported that the Hillsboro City Council unanimously adopted a 120-day moratorium on new land-use applications for data centers and battery energy storage system (BESS) sites at a special meeting on Monday, July 27, 2026. It is the first such moratorium in Oregon. Hillsboro has 35 data centers, according to the website Data Center Maps, which is more than any other city in the state. The city says at least seven more sites are under construction or moving through permitting.

    The pause does not affect existing data centers or projects whose land-use applications were already submitted. The council may extend it once for another 120 days. The city gave one day’s notice of the meeting, which Mayor Beach Pace said was meant to keep companies from rushing applications in before the cutoff.

    Executive Summary

    Hillsboro, the center of the cluster of technology firms around Portland known as Oregon’s Silicon Forest, has temporarily stopped accepting new land-use applications for data centers and grid-scale battery sites. The city says it will use the pause to study the facilities’ impacts and amend its community development code to regulate them.

    The headline word is moratorium, but the practical effect is narrower. Operating campuses and already-filed projects continue, so Hillsboro’s near-term data center pipeline keeps moving. The lasting change will be the zoning rules that emerge from the study. Those rules will govern every new campus proposed in one of the Pacific Northwest’s most established data center markets, at a time when AI computing demand is driving much of the industry’s new capacity.

    The vote also adds to a broader tightening in Oregon. The state has paused new data center tax breaks, approved higher power rates for large users, and set up a governor’s advisory council on data centers.

    What the Pause Actually Freezes

    The moratorium is narrower than the word suggests. It stops Hillsboro from accepting new land-use applications, which are the zoning filings a developer needs before a project can move through local approvals. It does not touch the city’s roughly 35 operating data centers, and it exempts any project whose application was already on file. By the city’s own count, at least seven more sites were under construction or moving through permitting at the time of the vote. Projects at those stages have, in the ordinary course, already passed the point the moratorium blocks.

    Mayor Beach Pace’s reason for calling the meeting on a day’s notice points the same way. The concern was that companies would rush applications in before the cutoff. That suggests the city expected a queue of projects ready to file, and that the pause is aimed at the next round rather than at capacity already committed.

    In the near term, then, Hillsboro’s data center footprint keeps growing. The moratorium does not change the buildings going up now. It changes the terms for whatever comes after them.

    The Code Rewrite Is What AI Builders Should Watch

    The city’s stated goal is to study the impacts of data centers and battery sites and amend its community development code to regulate them. That is the outcome that will last. The moratorium itself expires after 120 days, or 240 if the council uses its single extension. City Attorney Chad Jacobs told the council that state law limits moratoriums to 120 days, and the ordinance allows one renewal. Zoning amendments, by contrast, stay on the books and govern every future application.

    For developers, that moves the main risk from delay to design. A four-to-eight-month pause is an inconvenience on a multi-year data center schedule. New siting standards, depending on what they cover, could change which parcels are viable and what a campus must include to be approved. Across the industry, much of the next wave of capacity is being driven by AI computing demand. That typically means larger, more power-dense campuses, which are the projects most exposed to new local rules on land, power and water.

    The council has not said what the amendments will contain, and the testimony pulled in different directions. That uncertainty, more than the 120 days, is what a company weighing a new Hillsboro site has to factor in.

    Batteries Caught in the Same Net

    The most contested decision was to include battery energy storage systems. These are large banks of batteries that absorb electricity when it is plentiful and release it when demand peaks. Portland General Electric and clean-energy advocates asked the council to exclude them. Adam Capage of Renewable Northwest said the facilities “play a vital role in Oregon’s energy future” and already support resilience and affordable electric service in Hillsboro.

    Councilors Cristian Salgado and Rob Harris acknowledged that battery sites and data centers are different, though they often come together. They argued the city needs to understand both before writing rules. That is a reasonable argument about order of operations, but it has a cost. Storage is one of the tools utilities use to manage demand peaks as large new loads connect. Pausing new battery applications at the same time as the loads leaves an open question for grid planning in the area during the study period.

    A Market Tightening From Several Directions

    Hillsboro’s vote follows several state-level changes. Earlier this year the Oregon Legislature enacted a one-year pause on new data center tax breaks through the Enterprise Zone program. In July the Public Utility Commission approved Portland General Electric’s 29.7% rate increase for large energy users, while residential rates fell 1.3%. Governor Tina Kotek has said the state is a “cheap date” for data centers and suggested more changes are coming.

    Each measure targets a different cost: taxes, power prices and now land use. Together they show that Oregon is renegotiating its terms for large facilities. The Washington County Chamber of Commerce, the Westside Economic Alliance and three building-trades unions argued for letting established planning processes and the governor’s data center advisory council finish first. Most residents who testified supported the pause. Both positions have merit. The chamber’s point about sequencing is reasonable, and so is the city’s view that it cannot write sound rules for facilities whose local impacts it has not studied.

    Background

    Hillsboro, in Washington County west of Portland, is the center of Oregon’s Silicon Forest. The area has long been home to semiconductor manufacturing and has become the state’s densest cluster of data centers. Many large operators have received long-running property tax breaks through Oregon’s Enterprise Zone program. Those breaks became a local flashpoint this spring as residents learned how long they would last.

    State policy has been shifting as well. The Legislature paused new Enterprise Zone tax breaks for data centers for one year. Regulators approved higher electricity rates for large power users such as data centers under a new state law. Governor Tina Kotek has said Oregon is a “cheap date” for the industry and suggested more changes are coming, and her data center advisory council is still working on recommendations.

    Sources

    Source: Hillsboro approves data center moratorium, Oregon Public Broadcasting’s report on the Hillsboro City Council’s July 27, 2026 vote.

  • Israel Eased Data Center Planning. Its 27GW Grid Queue Shows Where the Bottleneck Moved

    Israel Eased Data Center Planning. Its 27GW Grid Queue Shows Where the Bottleneck Moved

    TL;DR · 30-second read

    The Short Version

    Israel has stopped accepting new requests from large data centers to connect to the power grid for about four and a half months. Data centers are the warehouse-sized buildings full of computers that run cloud services and artificial intelligence.

    The reason: builders have asked for nearly three times as much electricity as the whole country normally uses. Officials want time to separate real projects from ones that may never be built.

    At least one developer has already dropped its plans. How Israel sorts this out will decide who gets power, and when.

    Israel’s Electricity Authority has frozen the processing of new applications to connect data centers of 8 megawatts (MW) or more to the power grid for 140 days, DatacenterDynamics reported on July 28, 2026. The authority described the pause as an “emergency brake” to prevent the system from making connection commitments it may not be able to fulfill. The freeze is expected to run until early December.

    The country’s transmission operator, Israel Independent System Operator (NOGA), received 19 gigawatts (GW) of new data center connection requests within two months, bringing the pending total to roughly 27GW, nearly triple Israel’s average baseline electricity consumption. During the pause, the Electricity Authority, the Ministry of Energy and NOGA will jointly develop a new policy for allocating electricity between data centers and the rest of the economy.

    Executive Summary

    Israel has imposed a national pause on new grid connection processing for data centers of 8MW and above, one of the clearest cases yet of a grid operator stopping the queue rather than working through it. The trigger is scale: roughly 27GW of pending requests against a grid whose average baseline consumption is about a third of that.

    The surge followed the National Cyber Defense Bill, passed by the Knesset in February, which designated data centers as national infrastructure to remove planning and construction constraints. With those hurdles lowered, the grid connection became the constraint that still binds, and it absorbed the full weight of demand, speculative filings included.

    The 140-day review is meant to map the requests, set eligibility criteria and build a mechanism to tell legitimate projects from speculative ones. The outcome will shape which developers can build in Israel, how quickly, and on what terms. VisionWave has already abandoned a hyperscale project, citing the pause.

    Easing Planning Moved the Bottleneck to the Grid

    In February, Israel’s parliament passed the National Cyber Defense Bill. The law classified data centers as national infrastructure specifically to remove planning and construction constraints, with the aim of speeding up an industry the government considers strategic. What followed was a flood of grid applications. NOGA received 19GW of new data center connection requests within two months, and the pending total now stands at roughly 27GW, nearly triple the country’s average baseline consumption. For scale, a gigawatt is roughly the output of a large power station.

    The mechanism is straightforward. A data center project passes through several filters: land, zoning, construction permits, financing, customers and power. Lower the planning filters and it becomes cheaper to put a project into the pipeline, while a place in line for scarce electricity stays just as valuable. The grid connection becomes the one gate that still binds, so it absorbs all of the demand, including requests from projects that may never be built. A planning rule can be rewritten in a single legislative session. New transmission lines and generation take years.

    That is the logic behind the authority’s “emergency brake” framing. A connection offer is a promise to deliver power years from now. Issued against a queue that far exceeds national demand, such promises would either go unkept or crowd out other users. The review’s stated purpose, allocating electricity between data centers and the economy at large, puts that trade-off in plain terms: every megawatt reserved for a server hall is a megawatt not available to homes, factories or transport.

    A 27GW Queue Is a Signal, Not a Forecast

    No grid adds three times its existing load in a few years, and nothing in the announcement suggests Israel expects to. A queue of this size says more about the incentives facing developers than about the power that will eventually be consumed. Developers often file for several sites at once to keep their options open, and an early filer can hold capacity that a better-financed project later needs. The government’s plan to map requests, set eligibility criteria and separate legitimate projects from speculative ones amounts to an acknowledgment that the queue, in its current form, cannot guide grid planning.

    The hard part is that genuine demand is buried inside the inflated number. The pause comes as AI training and inference reshape data center demand worldwide. These workloads concentrate large, steady power draws on single campuses, which is exactly the profile that stresses transmission planning. Last month Crusoe, a developer focused on AI infrastructure, said it would invest around $10bn in Israeli data centers over ten to 15 years through leases. Commitments of that kind compete for the same limited grid capacity as filings with little behind them, and that is the problem the sorting exercise is meant to solve.

    Who Wins and Who Waits

    The 8MW threshold matters. Enterprise server rooms and smaller edge facilities fall below it. Hyperscale and AI campuses sit well above it. The freeze therefore lands on the segment the February law was designed to encourage. For those developers, 140 days carries a real cost. Land options, equipment reservations and tenant negotiations all lose value over time, and a project without a connection date is hard to finance or lease.

    VisionWave has already dropped plans for a hyperscale data center in Israel and cited the pause as a key factor. Other developers may choose to wait rather than leave. Rules built to separate legitimate from speculative projects would favor applicants that can show land, financing and customers. That would benefit mature developers and well-capitalized tenants over applicants holding queue positions on the chance that one becomes valuable.

    The broader lesson is not that Israel is closing to data centers. It is that easing one constraint without adding grid capacity moves the queue somewhere else. Israel’s experience suggests that any market courting AI infrastructure needs a clear, enforceable rule for allocating power alongside its permitting reforms.

    Background

    Israel’s Electricity Authority is the power sector’s regulator and sets the rules for how customers connect to and pay for the grid. NOGA, the Israel Independent System Operator, runs the transmission network and processes connection requests for large loads. Israel’s grid is largely isolated from its neighbors’, so electricity demand has to be met mainly by generation inside the country. That makes a sudden surge in requested load especially hard to accommodate.

    Israel has been positioning itself as a regional data center hub. In February the Knesset passed the National Cyber Defense Bill, which designated data centers as national infrastructure to ease planning and construction. Developers followed: Crusoe announced in June a roughly $10bn lease-based investment plan over ten to 15 years. The grid freeze is the first major check on that expansion.

    Sources

    Source: Israeli Electrical Authority freezes grid processing of new data centers for 140 days (DatacenterDynamics): Israel pauses grid connection processing for data centers of 8MW or more after 27GW of requests.

  • NV Energy’s 22 GW Queue Shows Why Data Center Power Priority Is Worth Suing Over

    NV Energy’s 22 GW Queue Shows Why Data Center Power Priority Is Worth Suing Over

    TL;DR · 30-second read

    The Short Version

    Nevada’s main electric company, NV Energy, is suing Tract, a company that builds huge campuses for data centers. Data centers are the warehouse-sized buildings full of computers that run websites, apps and artificial intelligence.

    The fight is over who gets electricity first, how much, and who pays for new power plants and lines. Requests for new power, driven largely by data centers, already add up to more than twice what the utility’s whole system uses on its busiest day.

    The utility says households could get stuck with the bill. Tract says it has promised nearly $1 billion toward the utility’s equipment and is paying its fair share.

    NV Energy, Nevada’s primary electric utility, has sued data center developer Tract in Washoe County’s Second Judicial District Court, The Nevada Independent reported on July 28, 2026. The complaint was filed Friday, July 24, and much of it is redacted because it was filed under seal. It alleges that Tract started private arbitration over when and how much power the utility must provide to its Peru Shelf and South Valley projects at the Tahoe Reno Industrial Center, and that this bypasses the Public Utilities Commission of Nevada (PUCN).

    NV Energy says Tract’s demands would shift infrastructure and energy costs onto existing customers. Tract says the complaint “seriously misstates the issues, the facts, and the contracts NV Energy signed.” It also says it has never sought subsidies, that it has committed nearly $1 billion to support NV Energy’s infrastructure, and that those contracts commit the dispute to arbitration.

    Executive Summary

    The suit pits Nevada’s dominant utility against one of the largest landholders in the state’s data center market. It turns on a question every fast-growing power market will eventually face: when a developer’s contract and a utility’s regulated resource plan collide, who decides the order in which new loads get electricity? NV Energy says only the PUCN can. Tract says the parties agreed to arbitrate.

    The backdrop explains the stakes. NV Energy’s latest integrated resource plan says it would need about 22 gigawatts (GW) just to serve existing requests for service, more than twice its current system peak, and it attributes that demand largely to data centers. When the grid is oversubscribed by that margin, a place in the queue has real economic value. Tract is already planning 362 megawatts (MW) of temporary gas and diesel generation because the grid cannot serve its projects right away.

    For ratepayers, the dispute is about who pays for growth. For developers and their financiers, it tests whether a large-load contract signed with a utility can be enforced on its own terms, or whether it stays subordinate to the regulator’s planning process.

    Twice the Peak: Why Queue Position Is Now the Prize

    NV Energy’s most recent integrated resource plan (IRP) says the utility would need roughly 22 GW of capacity just to meet requests for service already on file. That is more than twice the current system’s peak demand, and the utility says data centers are the main driver. A gigawatt is a thousand megawatts, and a single megawatt is a large industrial load on its own. When requests exceed what the grid can deliver by that much, the utility cannot connect everyone at once. The order of service then decides whether a campus opens on schedule, and it matters as much as the price of power.

    That order is exactly what this dispute concerns. According to the complaint, Tract’s arbitration seeks to settle when and how much power NV Energy should provide, how the utility generates and buys that power, and the priority of service among customers in its queue. Tract’s own contingency plan shows what a later place in line costs. NV Energy cannot serve the projects immediately, so Tract has proposed temporary natural gas and diesel plants: 144 MW for South Valley and 218 MW at Peru Ridge. These would be “behind-the-meter,” meaning on-site generation that feeds the campus directly rather than power drawn from the utility’s regulated grid.

    Three groups are affected. Other large customers in the queue are affected because capacity promised to one developer is capacity another does not get on that date. Existing ratepayers are affected if accelerated grid upgrades end up in general rates. Developers and their lenders are affected because a campus’s value depends heavily on when it can be energised. With that much at stake, both sides have escalated. Tract has gone to arbitration, and NV Energy has gone to court to stop the arbitration from deciding matters it says belong to the regulator.

    Contract Forum Versus Regulatory Forum

    NV Energy’s legal theory rests on jurisdiction. Generation, power purchasing and service priority are set through IRPs, which the utility must file with the PUCN and have approved at least every three years. The utility argues that an arbitrator cannot decide those questions outside that framework, especially while the Commission is weighing the same issues. That includes NV Energy’s proposal for a uniform large-load service agreement meant to standardise terms across data center customers. “Defendants cannot now short-circuit that regulatory process,” the filing says.

    Tract’s position rests on contract. It says the agreements NV Energy negotiated commit disputes to arbitration, and that it simply wants the utility “to live up to the commitments it made.” The agreements were executed under Rule 9, which governs how construction costs and responsibility are split between the utility and a new customer, weighing the project’s size, duration and risk. NV Energy says Rule 9 does not address the supply of electricity. The unresolved question is therefore whether the utility made supply or timing commitments in those contracts, and if it did, whether they can be enforced privately. With the agreements sealed, neither side’s reading can yet be checked publicly.

    The implications reach beyond Nevada. If construction-cost agreements can carry enforceable supply priority through arbitration, large-load contracts become a channel that runs parallel to resource planning. If they cannot, developers get less contractual certainty at the moment they are raising capital for multi-hundred-megawatt campuses. Either outcome carries costs for someone.

    Same Principle, Different Math

    Both parties invoke the same principle. “Customers that create new costs should cover new costs, and those costs should not be shifted,” said a person familiar with the matter. Tract says it “has committed to paying our fair share for infrastructure and generation needs” and never asked to be subsidised. The disagreement is not about whether growth should pay for itself. It is about what a fair share amounts to, and what the nearly $1 billion Tract cites actually buys in capacity, timing and priority.

    The growth figures explain why that math is contested now. NV Energy estimates it needs 52 percent more energy statewide than it forecast only two years ago. Data centers account for about 5 percent of its sales today, and the utility projects that share will reach 64 percent by 2046. On that trajectory, cost-allocation rules written now will govern who pays for most of the system’s future expansion. Questions that are traditionally settled in rate cases and resource plans are here being argued in court over the terms of a contract. That is one case, not yet a pattern, but it is a sign of how much weight these contracts now carry. The lawsuit also places the utility’s stated position near local concerns about data center electricity and water use, a shift in public posture for a utility that is simultaneously planning to serve that load.

    Land Is Not Power

    Tract has assembled more than 12,000 acres in Northern Nevada, including 1,060 acres in Lyon County that it says can support 1.6 GW of data center capacity. Land has not been the constraint, and grid access keeps proving to be one. Last year, Switch sued Tract to block certain developments on the same Peru Shelf parcels, arguing they would affect Switch’s ability to connect future projects to utilities. The court ruled in Tract’s favour. Two separate lawsuits over the same ground have now turned on access to power.

    For site selectors, investors and tenants, the practical lesson is that acreage, zoning and even signed construction agreements do not establish an energisation date. On-site gas and diesel generation is becoming part of the plan rather than an emergency fallback. That brings its own permitting, fuel and emissions questions, and it does not settle when the grid will take over.

    Background

    NV Energy is Nevada’s primary electric utility and a subsidiary of Berkshire Hathaway Energy. Like other regulated utilities, it plans new generation and power purchases through integrated resource plans filed with and approved by the PUCN at least every three years. It already has large-load agreements with multiple data centers. In its current plan it is asking regulators to approve a uniform service agreement for large-load customers, consistent with those existing contracts.

    Tract develops master-planned data center parks and has bought more than 12,000 acres across Northern Nevada. In 2024 it acquired 8,600 acres in Storey County between the Tahoe Reno Industrial Center (TRIC) and the Virginia City Highlands, bringing its holdings to 11,000 acres. It later added 1,060 acres in Lyon County just south of TRIC, which it says can support 1.6 GW of capacity. Last year Tract prevailed in a lawsuit brought by Switch, a longtime TRIC operator, over development of Tract’s Peru Shelf parcels.

    Sources

    Source: Nevada’s electric utility is suing a major data center company. Here’s why. (The Nevada Independent), on NV Energy’s lawsuit challenging Tract’s arbitration over power-supply agreements for two Storey County data center projects.

  • Infinium’s 1 MW in 250 Sq Ft Shows Waterless AI Cooling’s Real Test Is Heat Rejection

    Infinium’s 1 MW in 250 Sq Ft Shows Waterless AI Cooling’s Real Test Is Heat Rejection

    TL;DR · 30-second read

    The Short Version

    A Sacramento company best known for making clean fuel has launched a cooling system for the computer buildings behind artificial intelligence. Its selling point is that it uses no water.

    Instead of blowing cold air over computers, it sinks them in a liquid that does not conduct electricity. The company says one unit can handle up to a megawatt of computing, roughly the electricity several hundred homes use, in a space about the size of a large bedroom.

    The catch is that the heat still has to go somewhere, and the company has not said how the system copes on a hot day.

    Infinium, a Sacramento clean-fuel company, has unveiled the EdgeSites Vector ONE, a self-contained data center cooling system that uses no water, the Sacramento Business Journal reported on July 27, 2026. The system submerges servers in dielectric fluid, a liquid that does not conduct electricity, rather than cooling them with air conditioning.

    According to the report, Vector ONE supports up to one megawatt of computing power in 150 to 250 square feet, and the company says it allows computing in 70% less space than air-cooled equipment. No pricing, customers, shipment dates or energy-efficiency figures were given.

    Executive Summary

    Infinium is entering data center cooling with an immersion system, a design in which servers sit in tanks of non-conductive liquid that carries heat away from chips far better than air does. The product, sold under the company’s EdgeSites and Infinium Edge names, is pitched on two claims: it uses no water, and it packs up to a megawatt of computing into a footprint the size of a large room.

    Both claims land on live issues in AI infrastructure. AI servers run hotter and denser than conventional servers, pushing air cooling to its limits, and data center water consumption has become a flashpoint with local communities and regulators. A product that addresses density and water at once is well aimed.

    What the announcement does not yet show is the part that decides whether waterless cooling works at scale: how the system rejects a megawatt of heat to the outdoors, how much electricity that takes, and how its capacity holds up in hot weather, which is exactly when evaporative cooling earns its keep.

    A Megawatt in a Large Room Is the Real Headline

    The most concrete number in the announcement is density. One megawatt in 150 to 250 square feet works out to roughly 4,000 to 6,700 watts of computing per square foot of floor. That is far beyond what traditional raised-floor, air-cooled data halls are built to handle, and it is the kind of concentration modern AI hardware demands. Graphics-processor servers used to train and run AI models pack far more power into each rack than the general-purpose servers most facilities were designed around.

    Immersion cooling makes that density possible because liquid absorbs and moves heat far more effectively than air. The claim of 70% less space follows naturally: without the aisles, air handlers and spacing air cooling requires, servers can sit much closer together. For operators short on floor space, or trying to fit AI capacity into an existing building or a small edge site, that is a genuine advantage.

    Waterless Moves the Hard Problem Outdoors

    Immersion solves the problem of getting heat off the chip. It does not make the heat disappear. Every watt a server draws ends up as heat in the fluid, so a fully loaded Vector ONE has to shed up to a megawatt of heat continuously, and it has to do so from a very small footprint. That step, known as heat rejection, is where the water question is actually decided.

    Many large data centers use evaporative cooling because evaporating water is an extremely effective way to dump heat, especially in hot, dry weather. A system that uses no water must reject heat by dry means, typically fans pushing outside air across heat exchangers. Dry systems avoid water consumption entirely, but their capacity depends on outdoor air temperature: the hotter the day, the harder they work and the more electricity their fans draw. Immersion helps here, because the fluid can run warmer than the air in a conventional data hall, which widens the gap between fluid and outdoor air and makes dry rejection more practical. How much it helps is a question of engineering data Infinium has not published.

    This is why the density figure and the water claim are linked. Packing a megawatt into 250 square feet concentrates the heat-rejection burden, and the outdoor equipment, its size, noise, power draw and hot-day rating, becomes the component that determines whether the megawatt figure holds in July in Sacramento or Phoenix. Site developers in water-stressed regions, utilities sizing grid connections, and buyers comparing total energy costs are the people for whom that answer matters most.

    Water Is the Pressure Point, Energy Is the Trade-Off

    Positioning around zero water use is commercially astute. Data center water consumption has drawn scrutiny in drought-prone areas, and a cooling approach that removes water from the conversation can ease permitting and community relations. That value is real even before any efficiency numbers are considered.

    The efficiency claim as reported, that immersion cools servers more efficiently than air conditioning, is a comparison with air cooling, a fairly low bar for liquid systems. The more relevant comparison for AI-scale sites is against water-assisted cooling, where operators trade water for lower electricity use. The industry-standard way to express that trade is power usage effectiveness, or PUE, the ratio of total facility power to computing power, alongside water usage effectiveness. Neither figure accompanied the announcement, so the claim that Vector ONE is both waterless and efficient remains to be demonstrated.

    An Edge Product From an Unexpected Supplier

    Infinium’s core business is electrofuels, synthetic fuels made from renewable electricity, and the company has recently landed an airline fuel deal and begun work on what it calls the world’s largest electrofuel plant. A cooling product is a notable step sideways, though both lines of business sit at the intersection of energy and heavy industrial engineering.

    The EdgeSites and Infinium Edge branding, and the one-megawatt unit size, point toward modular and edge deployments rather than the hundred-megawatt campuses hyperscalers build. That is a sensible niche: edge sites are often in places where water hookups, space and cooling infrastructure are hardest to come by, and a self-contained unit that arrives ready to run fits those constraints. It also means Vector ONE competes with established immersion and modular data center vendors, where track record, server-maker support and service networks weigh heavily in purchasing decisions.

    Background

    Infinium is a Sacramento-based clean-fuel company that develops electrofuels, synthetic fuels produced with renewable electricity. It has announced an airline fuel supply deal and started construction on what it describes as the world’s largest electrofuel plant. The EdgeSites Vector ONE, marketed under its Infinium Edge name, marks its move into data center cooling.

    Immersion cooling is an established technique that has moved from niche uses toward mainstream interest as AI hardware pushes power density beyond what air cooling handles comfortably. At the same time, data center water consumption has become a siting and permitting issue in water-stressed regions, increasing demand for cooling approaches that avoid evaporative water use.

    Sources

    Source: Infinium unveils self-contained data center cooling system that uses no water (Sacramento Business Journal, July 27, 2026), reporting on Infinium’s EdgeSites Vector ONE waterless immersion cooling system.

  • Pentagon Offers AI Data Centers Land at 12 Bases; the Real Price Is Off-Grid Power

    Pentagon Offers AI Data Centers Land at 12 Bases; the Real Price Is Off-Grid Power

    TL;DR · 30-second read

    The Short Version

    The US military is inviting private companies to build giant computer warehouses for artificial intelligence on at least 12 of its bases, from Texas to Alaska. Two projects, each estimated at about $2 billion, have been tentatively awarded.

    The military pays nothing up front. In exchange for the land, it gets to use the computers. The catch is that each project must make its own electricity instead of plugging into local power lines, and must not drain local water. In effect, the builders are being asked to build their own power plant too.

    Some proposed sites sit half a mile from family homes, and Congress is still arguing over the rules.

    Military.com reported on July 27, 2026 that the Army and the Department of the Air Force have opened at least a dozen military installations to commercial data center developers over the past year and a half. Only two deals have been awarded, and both are conditional. In late March the Army selected Carlyle for roughly 1,384 acres at Fort Bliss, Texas. It selected CyrusOne, a data center operator jointly held by funds managed by KKR and BlackRock, for a roughly 1,201-acre parcel at Dugway Proving Ground, Utah. The Financial Times estimated each project at about $2 billion.

    The other sites are still solicitations. The Army is considering Fort Hood, Texas, and Fort Bragg, North Carolina. The Air Force has sought bids at Arnold, Edwards, Joint Base McGuire-Dix-Lakehurst, Davis-Monthan and Robins, and more recently at Joint Base Elmendorf-Richardson, Eielson Air Force Base and Clear Space Force Station in Alaska. Proposals must include net-zero water usage and a power plan that does not draw on the local electrical grid. Initial operating capability is projected for fiscal 2027 at Fort Bliss and fiscal 2029 at Dugway.

    Executive Summary

    The Pentagon is turning spare base land into a currency for AI computing. Under the Army’s Enhanced Use Lease program, private developers finance, build, operate and eventually decommission hyperscale data centers on military land. Hyperscale refers to the warehouse-sized campuses built for the largest cloud and AI workloads. In return the service gets access to computing power, at what the Army calls “no upfront cost to taxpayers.” Army Secretary Dan Driscoll has described Fort Bliss as “the first hyper-scale data center that the Pentagon has ever done.”

    The terms matter more than the land. Unlike commercial projects off base, these bids must bring a power plan that does not rely on the local grid and must achieve net-zero water use. For developers, that changes the job from building a data center to building a data center plus its own utility. For base commanders and the surrounding communities, siting and power have become part of the procurement itself.

    The program is also politically unsettled. A House defense bill provision would bar leases unless developers exclude certain components from China, Russia, Iran and North Korea, and the Army has objected that the rule would penalize building on federal land.

    The Land Is Free. The Power Plant Is Not.

    In the commercial market, the scarcest input for a large AI campus is usually a grid connection, not land. A grid connection is the utility’s commitment to deliver large, steady amounts of electricity, and it often comes only after a long wait in an interconnection queue, the line of projects waiting for approval to plug in. The Army and Air Force solicitations rule that path out. Proposals must include a power plan that does not draw on the local electrical grid, a requirement that does not apply to identical projects built off base.

    That changes what is actually being procured. The Army’s side of the bargain, land in exchange for computing access, is accurate in cash terms, but it moves the heaviest capital item onto the bidder. Whoever builds at Fort Bliss or Dugway is committing to finance and run generation as well as the data halls. The Army’s long-term vision, as described by Col. John Oliver, is a campus with a commercial side, a classified military data side and onsite power generation. The roughly $2 billion estimate for each project has not been broken down publicly, so nobody outside the negotiations can yet say how much of it is power plant and how much is computing.

    The rule is likely to favor sponsors with deep capital and power-development experience, such as a global investment firm like Carlyle or a KKR- and BlackRock-backed operator like CyrusOne, over developers whose model depends on buying utility power. It also changes the utility relationship. El Paso Electric attended the Army’s spring listening session at Fort Bliss, and the Army says excess power generated on base could potentially be sold back to civilian grids. A would-be large customer could become a supplier instead. For residents, the rule speaks directly to the concern behind a Virginia state-funded study, which found monthly household energy bills there could rise by $14 to $37 by 2040 as data centers grow.

    Water Is the Second Utility the Bidder Must Bring

    Data centers use water to cool their servers, and the net-zero water requirement is hardest to meet at the Army’s flagship site. Defense Department procurement documents reported by Data Center Dynamics rate the water risk for the El Paso area as “Extremely High.” One option under discussion is having Carlyle drill a new well to feed El Paso’s desalination plant and offset what the data center uses. Oliver was candid that this is unproven: “We don’t know if that’s an engineering solution that we can actually get to yet.”

    Water and power are linked. Cooling designs that use less water generally use more electricity. In a desert location, a strict water target therefore adds to the size of the on-site generation the developer must build and pay for. The two requirements compound rather than sit side by side.

    A Dozen Sites, Two Awards, Very Different Neighbors

    The sites vary widely. Dugway, where the military conducts its primary chemical and biological weapons testing, offers three parcels totaling about 3,466 acres, roughly an hour’s drive from the nearest community. Fort Hood has 207 acres on offer, and Fort Bliss has nearly seven times that. Contracting documents show a Fort Hood parcel within a half-mile of residential and commercial property, and several Fort Bragg locations within a mile of civilian areas and a half-mile of civilian housing. Smaller parcels close to homes leave less room for generation, and less buffer for its noise, than a remote range does.

    Opposition near homes is not a hypothetical risk. Data Center Watch counts nearly $156 billion in projects nationwide delayed or canceled after local opposition. The Army has built community engagement into the bid: developers must describe outreach plans and assess “any risks or opposition.” Brookings senior fellow Darrell M. West said communities that accept these facilities tend to be the ones told the details up front, including energy source, water use, cost and noise.

    Congress Is Still Writing the Rules

    Rep. Cory Mills secured a provision in the House fiscal 2027 defense authorization bill barring data center leases unless developers avoid certain printed circuit boards, advanced semiconductors and chipsets made in China, Russia, Iran or North Korea. The Army objected that it would create a “federal land penalty,” imposing rules on installations that would not apply to identical projects anywhere else. Rep. John Garamendi’s proposal to require evaluation of energy, water, noise, light and security effects before any deal was rejected by House Armed Services Republicans. A Senate amendment to pause leasing until the Pentagon finalizes a data center strategy was also defeated.

    Each condition adds to the same balance sheet. Free land has to pay for off-grid generation, net-zero water and possibly a restricted supply chain, and the Army’s own argument shows it knows the margin is not unlimited. Whether the economics still work at scale is the question the two conditional awards will answer first.

    Background

    AI workloads have pushed demand for hyperscale data centers faster than utilities can connect them, and local opposition has become a real constraint, with Data Center Watch counting nearly $156 billion in projects delayed or canceled. The Pentagon holds large tracts of underused land at remote and secure installations, and a Trump executive order directed agencies to open non-excess federal land to data center development. The Army’s Enhanced Use Lease program, with the U.S. Army Corps of Engineers handling lease negotiations and environmental review, is the vehicle.

    Carlyle is a global investment firm. CyrusOne is a data center operator jointly held by funds managed by KKR and BlackRock. Dugway Proving Ground in Utah is where the U.S. military conducts its primary chemical and biological weapons testing, and Fort Bliss in El Paso is home to the Army’s 1st Armored Division.

    Sources

    Source: Pentagon Plans AI Data Centers at Military Bases Across Multiple Branches (Military.com), on Army and Air Force plans to lease base land to commercial AI data center developers.

  • Vantage’s 95-Ton NOx Permit Shows AI Self-Power Is Judged by Ceiling, Not Output

    Vantage’s 95-Ton NOx Permit Shows AI Self-Power Is Judged by Ceiling, Not Output

    TL;DR · 30-second read

    The Short Version

    A data center in Virginia makes its own electricity with eight gas-burning turbines, each about the size of a bus, close to people’s homes.

    A Harvard researcher estimated its air pollution could cause tens of millions of dollars a year in health harm. Virginia’s environmental regulator moved quickly to challenge that estimate, even though one of its own experts said judging health effects was outside his training.

    Why it matters: more of the buildings that run artificial intelligence want their own power plants, and fights like this one will decide where they can be built.

    Politico reported on July 25 that internal emails, obtained through a public records request, show Virginia’s Department of Environmental Quality (DEQ) organized a rebuttal to a health analysis of Vantage Data Centers’ self-powered facility in Sterling, Loudoun County, before that analysis was published. DEQ Director Michael Rolband forwarded the Piedmont Environmental Council’s report to seven air-permitting staff within five minutes of receiving it in early March. He asked for “a defensible critique that will stand the test of public and critical review.”

    The Piedmont-commissioned study estimated that soot from the facility’s eight natural gas turbines could cause $53 million to $99 million a year in health costs, and 3.4 to 6.5 premature deaths, if the site emitted the maximum its permit allows. DEQ’s own report, released in late April, said actual emissions were 28 to 69 percent of permitted levels. The agency has said its staff “did not collaborate with PEC, Vantage or any other external participants” in its assessment.

    Executive Summary

    This dispute centers on the only data center in Virginia that generates its own electricity. Its on-site power system combines eight bus-sized natural gas turbines with dozens of backup diesel generators, and it sits in a residential area. DEQ’s 2023 permits allow up to 95 tons a year of nitrogen oxides (gases that help form smog) and more than 56 tons of soot, meaning fine particles that lodge in the lungs. Both pollutants are linked to asthma and heart attacks. Those limits are typical for a minor pollution source. The dispute arose because of where the facility sits, not because of how large the limits are.

    It matters beyond one building because self-generation is becoming the industry’s answer to grid congestion. Dominion Energy disclosed in February that it receives about 10 new connection requests a month from large-load customers. It also disclosed that dozens of projects requesting a combined 70 gigawatts are waiting to connect, three times the utility’s all-time peak load. Vantage itself described its turbines to DEQ as “a temporary solution given the current power constraints of the grid.”

    The episode shows how the health case against on-site generation gets built. Critics model the legal ceiling in the permit, not the meter reading. Even when actual emissions come in well below that ceiling, the permit keeps the higher figure in play.

    The Permit Ceiling Is the Number That Gets Modeled

    Michael Cork, a biostatistics research fellow at Harvard’s T.H. Chan School of Public Health, did not have Vantage’s operating data when Piedmont hired him. He used the only public figures available, which were the permitted maximums. That produced the headline range of $53 million to $99 million in annual health costs. Vantage objected that this was a “worst case scenario,” and DEQ’s report made the same point: measured emissions of soot, volatile organic compounds and nitrogen oxides ran at 28 to 69 percent of what the permit allows.

    The correction lowered the estimate but did not end the argument. Cork reran his model with DEQ’s actual-emissions figures and arrived at $33 million to $63 million a year and two to four premature deaths. He also noted that nothing in the permit prevents the facility from increasing output back toward its ceiling. That is the mechanism the headline refers to. A permit is a legal right to emit, so outside analysts can reasonably treat it as the facility’s potential footprint, whatever the current meter shows.

    This matters for developers planning on-site generation. Operators often seek permit headroom so they can run harder when the grid is strained or when load grows. That headroom has operational value. It is also, in effect, a published number that critics can model. A project that wants its real-world performance to count in public debate may need to make that performance binding, through enforceable limits closer to expected output or regular public reporting of actual emissions, rather than relying on a regulator to vouch for it after the fact.

    Why Self-Power Keeps Spreading Anyway

    The economics pushing data centers toward their own power plants are not easing. A 70-gigawatt connection backlog at a single utility means many projects face long waits for grid power. On-site gas turbines offer a way to start operating sooner. They also answer the political charge that data centers raise household electricity bills. The federal government has encouraged the model too: President Donald Trump promoted a wave of power plants for new data centers at a White House meeting with technology companies in early March.

    The Vantage case shows the trade-off. Self-generation moves the power question off the utility’s books and onto the local air-permit docket. The cost stops appearing in rate cases and starts appearing in health-damage estimates for the neighborhood downwind. The more facilities follow this path, the more often permitting agencies will be asked to rule on disputes that go beyond their traditional remit.

    A Regulator’s Rebuttal Is Not a Health Finding

    The emails show disagreement inside DEQ about how far the agency should go. Michael Kiss, a manager in the Office of Air Quality Assessments, wrote that the magnitude of impacts was “relatively low” and not surprising. He also said he had “absolutely no comments on the health study” because “we aren’t qualified to do that review.” Robert Burnley, who led DEQ from 2002 to 2006, said the conventional defense of a permit is narrower: that the emissions do not cause a violation of air quality standards.

    Several claims in DEQ’s final report are open to challenge on their own terms. The report said health impacts were “unlikely” because Loudoun County soot levels are below EPA standards. Health experts, including former EPA air quality official Michael Koerber, said soot can harm health even at EPA-approved levels, and the World Health Organization’s guideline is nearly half the EPA figure. The report also cited 2023 state health data, collected before Vantage began operating, to describe Loudoun as one of Virginia’s healthiest localities. That data says little about the facility’s effect.

    For the industry, the lesson is practical rather than partisan. The Data Center Coalition circulated DEQ’s report to reporters and lawmakers as evidence that the Piedmont analysis was wrong. A rebuttal whose health reasoning is contested gives less protection than it appears to. The most useful thing DEQ published was arguably the actual-emissions data. It narrowed the dispute to a real number, even though that number still produced a substantial estimate.

    Siting Near Homes Is Where the Costs Land

    DEQ’s permits for Vantage are in line with other minor-source permits in the state. The concern stems from the facility’s location in a residential area. Neighbors report constant noise as well as air pollution. Vantage says it has taken steps to reduce the noise. Health-cost models of the kind Cork used scale with the number of people in a plume’s path, so the same turbines sited away from housing would generate a very different estimate.

    Virginia has more than 600 data centers, and many future self-powered sites will compete for land in areas that are already developed. Where generation sits relative to homes is likely to be as consequential for permitting risk as how many tons it is allowed to emit.

    Background

    Northern Virginia, and Loudoun County in particular, hosts the world’s largest concentration of data centers, and Virginia is home to more than 600 of them. Dominion Energy, the region’s largest utility, has struggled to keep pace with demand. Large projects now wait in a queue that is far larger than anything the utility has ever served. In response, developers have begun planning facilities that generate their own electricity on site, usually with natural gas turbines.

    Vantage Data Centers’ Sterling facility came before that trend. DEQ issued its air permits in 2023. The Piedmont Environmental Council, a Virginia conservation group, has been active on data center land-use and environmental impacts. The Data Center Coalition is the industry’s trade group.

    Sources

    Source: Emails show how Virginia regulators downplayed data center health concerns — Politico’s report on internal DEQ emails concerning Vantage’s self-powered Loudoun County data center.

  • PJM’s 3.1 GW Data Center Dropout Shows AI’s Other Grid Risk: Load That Leaves

    PJM’s 3.1 GW Data Center Dropout Shows AI’s Other Grid Risk: Load That Leaves

    TL;DR · 30-second read

    The Short Version

    When one power line failed near Washington, DC, a wave of data centers (the warehouse-sized buildings full of computers that run the internet and AI) unplugged from the power grid within seconds and switched to their own backup power.

    Together they had been using about 3% of the electricity on a grid that serves 67 million people. Losing that much demand at once threw the system off balance. Lights flickered from Virginia to Chicago, and it took more than 10 minutes to settle.

    It was twice as big as a similar scare in 2024, and data centers keep multiplying. The proposed fix is to make them stay connected through brief hiccups, often by using giant batteries.

    A single power line failure outside Washington, DC, this past week caused a wave of Northern Virginia data centers to switch to backup power at nearly the same moment, TechCrunch reported. According to PJM data, about 3.1 gigawatts of demand disappeared from the PJM Interconnection grid in roughly 30 seconds. After a partial recovery, more facilities dropped off, and the grid’s surplus peaked at 3.49 gigawatts. Stabilizing took another 11 minutes. A line fault like this one normally takes only a few seconds to recover from.

    According to Reuters, the disconnected load was about 3% of PJM’s demand at the time. Sensor data from the startup Ting Labs showed voltage spiking across the grid from Northern Virginia to Chicago. There was no blackout, but lights flickered across the region. The event was roughly twice the size of a 2024 PJM incident in which 60 data centers disconnected at once and removed 1.5 gigawatts.

    Executive Summary

    Most discussion of data centers and the power grid asks whether there is enough electricity to feed them. This event shows the opposite failure: a cluster of very large customers that all leave the grid at once when they sense trouble. Each facility’s protective logic behaved as designed and switched to on-site backup power when voltage dipped. Together, those individually sensible decisions turned a routine line fault into a regional disturbance that lasted more than ten minutes.

    Two remedies are being proposed. The first is coordination. Ali Zain Banatwala, a senior market models specialist at the Independent Electricity System Operator, argues that neighboring loads need a way to disconnect and reconnect one after another rather than all together. The second is ride-through, which means keeping facilities connected through brief disturbances. Ricardo de Azevedo, CTO of ON.Energy, whose company sells campus-scale battery systems built for that purpose, said ERCOT, the Texas grid operator, will require large loads to ride through disruptions.

    The stakes rise with the load. Synapse Energy Economics put data centers at about 6% of PJM’s load in 2024 and projects 24% by 2040. The potential size of a synchronized dropout grows with that share unless the facilities are built to behave differently.

    AI’s Other Grid Problem Is Load That Leaves

    An electric grid has to keep supply and demand in near-perfect balance from moment to moment. When a line goes down, the grid normally absorbs the loss and rebalances within seconds. This time, the voltage dip from the fault reached many data centers sitting close together. Their protection systems are designed to make split-second decisions to safeguard the computers inside, and they moved the facilities onto backup power within a few seconds of each other. The grid lost about 3.1 gigawatts of demand in roughly 30 seconds. Generators were suddenly producing more electricity than customers were using, and voltage jumped.

    The second wave is the detail grid planners should focus on. PJM’s grid appeared to recover partway before more loads dropped off, pushing the surplus to a peak of 3.49 gigawatts. In that pattern, the disturbance triggers disconnections, and the disconnections deepen the disturbance. That feedback is what stretched a seconds-long event into an 11-minute one. A gigawatt is roughly the output of a large power plant. In effect, the grid had to absorb the sudden departure of several power plants’ worth of customers.

    That is why the grid risk from the data center buildout is not only a question of supply. Operators are used to planning around the sudden loss of a large generator. Clustered data centers create a mirror-image contingency: a large block of customers lost all at once because of a single triggering event. Everyone on the grid is exposed to that. Households saw flickering lights, generators must ramp down fast, and the operator must hold capability in reserve for a contingency that grows as data center load grows.

    Concentration Turns a Local Fault Into a Regional Event

    Northern Virginia has the highest concentration of data centers in the world. That density is what made the disconnection synchronized, because one voltage dip reached many facilities almost at once. The effect traveled well beyond Virginia. Ting Labs runs a sensor network through household electrical outlets, and it recorded the voltage spike as far away as Chicago. PJM covers the territory from New Jersey to Illinois, serves 67 million customers, and is the largest grid operator in the United States.

    The trend line gives the event its weight. In 2024, 60 data centers disconnected together and removed 1.5 gigawatts. This time the figure was about 3.1 gigawatts, roughly double. Data centers were about 6% of PJM’s load then, and Synapse Energy Economics projects 24% by 2040. Two events are not a statistical series, and nothing here shows the next one will double again. The arithmetic still matters: if protective behavior stays the same while the share of load it governs roughly quadruples, the ceiling on a synchronized dropout rises with it.

    De Azevedo called the event “the canary in the coal mine” and said incidents involving large loads are “happening more and more.” That claim comes from a vendor of the remedy. The two PJM incidents, with their measured sizes, are the documented reference points.

    Two Fixes: Choreograph the Exit, or Don’t Leave

    The first approach accepts that data centers will disconnect but asks them to do it in order. “We need to figure a way for these loads that are located next to each other to sequentially either disconnect or reconnect,” Banatwala said. Staggered behavior would let grid operators write procedures in advance instead of reacting to a cliff. The hard part is coordination. The facilities belong to different owners, each has settings tuned to protect its own uptime, and they would need shared thresholds or a shared signal.

    The second approach is ride-through, meaning the facility stays connected through brief disturbances. ON.Energy’s version is an uninterruptible power supply, the battery backup that keeps servers running through outages, sized for an entire campus, including chillers and other equipment as well as servers. Batteries and power-conversion equipment sit between the grid and the facility, so the grid sees one steady, predictable load. The company says the system:

    • charges its batteries when the grid has surplus power;
    • discharges to the servers when supply dips;
    • follows the grid within milliseconds;
    • smooths the swings of AI training workloads that ramp power up and down.

    It says it is installing 3 gigawatts of these systems across four campuses.

    Those claims are specific enough to test. The test is whether a campus behind such a system stays connected and grid-friendly during a real fault like this one. The economics matter too. Sizing batteries and power electronics for a whole campus adds capital cost that the developer bears, while most of the benefit goes to the grid. Mandates may be what shifts that calculation, such as ERCOT’s ride-through requirement, which de Azevedo described.

    What Developers and Buyers Should Expect

    From a tenant’s point of view, the July event worked: the data centers switched to backup, and the computers presumably kept running. The cost of that decision fell on the grid and everyone else connected to it. This asymmetry is what makes the event a policy question rather than just an engineering one. Individual facilities have little reason to change behavior that protects their own uptime unless interconnection rules require it.

    For developers, the likely direction is toward grid connection terms that specify how a facility behaves during a disturbance, not just how much power it draws. ERCOT is moving that way, according to de Azevedo. PJM, where both incidents occurred, is where the question is now most visible. For suppliers of storage, power conversion and switchgear, the grid-facing layer of a campus becomes part of the design requirement. For colocation buyers, it is reasonable to ask providers how their facilities respond to voltage dips and whether ride-through requirements could change facility design or pricing.

    Background

    PJM Interconnection is a regional transmission organization, the independent body that runs the high-voltage grid and wholesale electricity market for an area stretching from New Jersey to Illinois. It serves 67 million customers and is the largest grid operator in the United States. Its territory includes Northern Virginia, which has the highest concentration of data centers in the world.

    Data centers are built never to lose power. When grid voltage wavers, they switch to on-site backup, such as uninterruptible power supplies and generators, in a split second. That design protects the computers inside. When many facilities share the same stretch of grid, however, it also means they can react to the same disturbance at the same moment. PJM saw this in 2024, when 60 data centers disconnected simultaneously, and saw it again at twice the scale this July. ERCOT, the operator of most of Texas’s grid, is among the grid managers moving to require large loads to ride through disturbances.

    Sources

    Source: One fallen power line exposed a growing AI data center problem — here’s how to fix it (TechCrunch): a report on the roughly 3-gigawatt data center disconnection on the PJM grid and proposed ride-through fixes.

  • Fort Bliss’s 1,384 Acres Show Military AI’s Hard Limit Is Power, Not Land

    Fort Bliss’s 1,384 Acres Show Military AI’s Hard Limit Is Power, Not Land

    TL;DR · 30-second read

    The Short Version

    The United States military is letting private companies build giant computer centers for artificial intelligence on its bases. One in Texas will cover more than two square miles of Fort Bliss. Another is planned for a Utah testing range.

    The catch is electricity. A local newsroom estimated the Texas site alone could use more power than all 460,000 homes and businesses on El Paso’s power system combined. Members of Congress from both parties now want the Pentagon to explain where that power and water will come from before building starts.

    Lawmakers from both parties are trying to attach conditions to the Defense Department’s program of leasing military land to private companies for AI data centers, Politico reported on July 25, 2026. The Army has signed two conditional agreements: Carlyle Group will build on 1,384 acres at Fort Bliss in El Paso, Texas, by 2027, and CyrusOne will build on 1,201 acres at Utah’s Dugway Proving Ground by 2029, according to an Army fact sheet. The Air Force has offered unused land on five installations.

    Most of the proposals would require briefings, studies or planning rather than stopping construction. A provision from Rep. Ken Calvert (R-Calif.) requiring DOD to brief Congress on community impacts before construction was adopted in the House Defense spending bill, and a provision from Rep. Cory Mills (R-Fla.) barring land deals for data centers with significant components made by U.S. adversaries is in the House-passed NDAA. Broader Democratic amendments, including one from Rep. John Garamendi (D-Calif.), were rejected.

    Executive Summary

    The Pentagon’s plan to host commercial hyperscale data centers on military bases, set in motion by President Trump’s executive order last year directing the department to identify suitable sites, has moved from policy to signed deals. The Army describes its Fort Bliss and Dugway projects as coming at “no upfront cost to taxpayers” and including “comprehensive behind-the-meter power and water solutions”, meaning the sites would generate or source their own electricity and water rather than draw on the local utility.

    That promise is now the center of the congressional debate. The measures gaining traction do not question whether the military can offer land; they ask the Defense Department to show, before construction, how these facilities will be powered and cooled without straining base infrastructure, local grids, or water supplies. For developers, that turns power and water planning from a project detail into a condition of doing business on federal land.

    The House and Senate still have to reconcile their versions of the defense authorization and spending bills, possibly after the November elections, so the final terms remain open.

    Land Is the Part the Military Can Give Away

    The Army’s pitch to developers is built on land: large, secured, already federally controlled parcels offered without an upfront cost to taxpayers. Fort Bliss alone puts 1,384 acres on the table, and Dugway another 1,201. Even supporters in Congress frame the benefit in land terms. Rep. Derrick Van Orden (R-Wis.) said he favors base siting if it means fewer data centers on farmland that could otherwise grow crops.

    But acreage is not what limits an AI data center. The binding constraint is electricity, and then water for cooling. El Paso Matters has estimated that the Fort Bliss facility could require more power than all of El Paso’s 460,000 electricity customers combined. Whether or not that estimate holds at full build-out, it sets the scale: a single campus of this kind can rival a mid-sized city’s demand. No amount of available land changes that arithmetic.

    That is why the Army’s promise of “behind-the-meter” power carries so much weight. Behind-the-meter means generation on the customer’s side of the utility connection, so the data center would not simply pull from the local grid. If it works, the developer absorbs the cost and risk of building a power plant. If it falls short, the load lands on the regional utility and its ratepayers, which is precisely the outcome lawmakers from both parties say they want the Pentagon to rule out in advance.

    Congress Is Asking for Proof, Not Stopping the Build-Out

    The legislative scorecard shows where the consensus sits. Proposals to halt or ban construction failed: Rep. Rashida Tlaib’s (D-Mich.) amendment to prohibit AI data centers on federal lands was rejected, as was a Senate Democratic amendment to block leasing until DOD finalizes a data center strategy. Garamendi’s pre-deal certification requirement, covering energy, water, noise, light, security and supply chains, was voted down in House Armed Services.

    What advanced is disclosure. Calvert’s provision, written into the House Defense spending bill, acknowledges that data centers “place significant strain on energy and water resources” and requires a pre-construction briefing on community impacts. A bipartisan Energy-Water package asks the Department of Energy to mitigate data centers’ effects on grid reliability and ratepayers. Even critics of Garamendi’s approach, such as Rep. Pat Harrigan (R-N.C.), suggested annual impact reports as an alternative. The live disagreement is not whether the Pentagon must account for power and water, but whether that accounting comes before a deal is signed or after.

    DOD officials argue the department has already accounted for these impacts and warn that added requirements could slow the military in a competitive AI race. That concern is legitimate; so is lawmakers’ complaint, voiced by Rep. Mike Levin (D-Calif.), that they have not received a briefing. Both claims could be tested quickly if the department published the power and water figures behind its agreements.

    The Supply Chain Clause May Bite Harder Than the Studies

    The provision with the most direct commercial consequence may be the one from Mills, now in the House-passed NDAA. It would bar DOD from providing land to data centers containing “significant components” manufactured by U.S. adversaries. An unnamed Army official told the Federal News Network that this would create an “unworkable standard” that could drive commercial partners away from Army land.

    Unlike a briefing requirement, a component rule reaches into the equipment bill of materials: servers, networking gear, power and cooling systems. How much it bites depends on how “significant” is defined and enforced, and on whether it applies to behind-the-meter generation equipment as well as computing hardware. For developers weighing base sites against private land, a sourcing rule that is hard to certify could matter more to the economics than any study requirement.

    Who Carries the Risk

    For Carlyle and CyrusOne, the emerging terms shift risk onto the developer: finance the facility, supply its own power and water, and, if Mills’ language survives conference, document the origin of key components. For base commanders, the question is whether on-site generation and water use interact with installation readiness. For neighboring communities such as El Paso, the stakes are grid reliability and electricity bills, the same concern driving the House Energy and Commerce Committee’s Ratepayer Protection Act, H.R. 9340.

    The broader lesson for anyone siting AI infrastructure is that federal land does not bypass the power question. It relocates it, and adds a layer of public accountability that private sites do not face.

    Background

    AI data centers pack large numbers of power-hungry processors into a single site, and their electricity and cooling needs have become a flashpoint for utilities, regulators and communities across the United States. Hyperscale facilities, the largest class of data center built for cloud and AI workloads, can draw power on the scale of a small city and often use significant volumes of water for cooling.

    The Defense Department entered this market after a presidential executive order last year directed it to identify military sites suitable for AI infrastructure. The Army’s model leases base land to private developers, who build and operate the facilities, and the Air Force has offered unused land at five installations. Carlyle Group is an investment firm; CyrusOne is a data center developer and operator.

    Sources

    Source: The Pentagon wants to build data centers. Congress would like a word. (Politico) — bipartisan congressional scrutiny of Defense Department land leases for AI data centers at Fort Bliss, Dugway and Air Force installations.

  • Georgia’s 1,000-Mile AI Grid Build Shows the Cost the Rate Math Misses: Easements

    Georgia’s 1,000-Mile AI Grid Build Shows the Cost the Rate Math Misses: Easements

    TL;DR · 30-second read

    The Short Version

    Georgia’s biggest power company is building more than 1,000 miles of new power lines, longer than the drive from Atlanta to New York City. About 80 percent of the new electricity it plans to add is expected to go to data centers, the giant warehouses of computers that run artificial intelligence.

    The lines must cross private land: more than 330 properties, including nearly 30 homes that will be torn down. The company starts its offers a quarter above a property’s appraised value and says it rarely forces a sale. But the right to run lines across land usually lasts forever, even after the owner sells.

    Fortune reported on July 26, 2026 that Georgia Power is contacting property owners along a 35-mile route through Fayette, Heard, Fulton and Coweta counties for the Ashley Park-Wansley Transmission Project, a new line between its Ashley Park Substation and its plant at Wansley. It is one of five projects that together make up more than 1,000 miles of new transmission in the utility’s plan. That plan spans more than 330 parcels of land and requires nearly 30 residential properties to be leveled.

    Georgia Power says land acquisition and easement negotiations are close to complete, with construction as early as 2027. It opens offers at 125% of appraised value and says eminent domain is used in fewer than 1% of its land transactions: five times in the past year, none on homes. The utility disputes describing the line as a data center project. The generation plan it agreed with the Georgia Public Service Commission in December 2025 says about 80% of nearly 10,000 new megawatts is expected to power data centers.

    Executive Summary

    Georgia Power is acquiring land and easements for one of five transmission projects that together add more than 1,000 miles of high-voltage lines to Georgia’s grid. Transmission lines are the long-distance wires that move bulk electricity from power plants to local substations. The buildout accompanies roughly 10 gigawatts of new generation that state regulators certified in December 2025, and about 80% of that generation is expected to power data centers.

    The story matters because it moves the cost of AI-era grid expansion out of abstract rate cases and onto specific parcels of land. Georgia Power argues that large data center customers will pay more, which it projects will cut residential bills by about $102 a year from 2029. That argument does not address the right-of-way. Nearly 30 homes are to be leveled, and permanent easements will cover hundreds of parcels.

    How utilities conduct those negotiations shapes how fast AI infrastructure can be powered and on what terms. So does how often they fall back on eminent domain, and how publicly they justify it.

    The Rate Math Counts Bills, Not Easements

    Georgia Power frames its public case for the buildout around electricity bills. The utility projects that the five transmission projects will save residential customers about $102 a year each starting in 2029. The mechanism is that large-load customers, meaning very big single users such as data centers, are required to pay more. Media Relations Manager Meredith Stone said those funds put “downward pressure on the rates for our residential customers.” That is a claim about tariffs, the prices customers pay, and it is the number the utility leads with.

    The land side of the same projects works differently. Delivering power over more than 1,000 miles of new line means securing rights across more than 330 parcels and leveling nearly 30 homes. Georgia Power says it opens negotiations at 125% of appraised value. Where owners grant an easement rather than sell outright, the utility gains a permanent right to use a strip of their land. David Needham, a Georgia attorney who has represented landowners in eminent domain cases for more than a decade, said those rights stay attached to the property after it is sold. They can prohibit permanent structures beneath the lines and can reserve broad rights for the utility to install additional infrastructure.

    The per-customer savings figure does not capture that cost. A bill saving is spread across the whole residential customer base and recurs every year. An easement is priced once, at the negotiating table, and its restrictions are carried indefinitely by whoever owns that particular piece of ground. For utilities and data center developers, the consequence is that AI-driven grid expansion depends on hundreds of individual land deals, not only on the generation plan regulators certified. Those deals determine how fast the expansion proceeds and how much public acceptance it wins.

    Condemnation Is Rare; the Room to Say No Is Narrow

    On the record, condemnation is uncommon. Georgia Power says eminent domain is a last resort that occurs in fewer than 1% of its land transactions. It says the power was used five times in the past year, none involving residential homes. Nothing disclosed contradicts those figures. The case that drew the most public attention also ended in a deal. It became widely known through social media posts by Ansley Brown and involved the office of U.S. Rep. Brian Jack, and Georgia Power said it reached an agreement with the family, without disclosing the terms.

    Residents’ description of the process is also consistent with how it works. “We don’t have a choice in this,” Brown said in one video. Once a route is chosen, a landowner’s options are a full sale or an easement, and keeping the line off the property is not one of them. Condemnation stands behind the negotiation as the alternative if talks fail. Needham called the resulting easement agreements “very much tilted in favor of the power company,” while noting that “there is compensation offered.” Both accounts can be accurate: forced takings are rare, and the bargaining position is uneven.

    Who the Line Serves Depends on Which Number You Read

    Georgia Power says it is “not in the data center business.” It says the Ashley Park-Wansley line will serve the broader needs of the communities and customers who rely on reliable service. CBS reported that 70% to 80% of the power from this particular line would go to AI data centers. A Georgia Power representative was unable to confirm that estimate.

    At the portfolio level the question is less open. In December 2025 the Public Service Commission and Georgia Power agreed to certify nearly 10,000 megawatts of new generation, and the agreement said approximately 80% of it is expected to power data centers. Transmission lines are shared assets, and the same wires feed homes, hospitals and server halls. The utility’s framing is therefore defensible for any single line. Still, the regulator-approved document says the new demand justifying this much new capacity comes mostly from data centers.

    The distinction matters because eminent domain rests on public use. The more visibly new corridors are tied to a small number of very large customers, the more utilities and regulators have to show the broader benefit. The $102 savings projection plays exactly that role.

    Georgia’s Welcome for Data Centers Is Getting Conditional

    Georgia hosts more than 200 operational data centers, nearly 170 of them in Atlanta, including facilities owned by Microsoft, Amazon and Google. A sales and use tax exemption dating to 2018 waives 100% of state and local sales and use taxes on equipment purchases for projects that meet a minimum investment of $100 million to $250 million. In 2024 lawmakers voted to suspend new exemptions, and Governor Brian Kemp vetoed the measure. A December 2025 evaluation by the University of Georgia’s Carl Vinson Institute of Government projected the exemption’s net fiscal impact at roughly negative $574 million.

    Georgia is now scrutinizing its tax incentives, charging large loads more, and facing land disputes in public view. The pattern for operators is that the state remains open to data centers but is pricing more of their wider costs explicitly. “The policy direction has clearly shifted,” said J. Thomas Perdue, a policy analyst at the Georgia Public Policy Foundation. For developers weighing sites, power in Georgia increasingly comes with public questions about who bears the cost of delivering it.

    Background

    Georgia Power is the state’s largest electric utility and a subsidiary of Southern Company. It plans its generation and transmission through an Integrated Resource Plan, a long-range forecast of demand and the resources needed to meet it. The plan requires approval from the Georgia Public Service Commission, the elected body that regulates the state’s utilities. The commission approved the 2025 plan, which cited continued power needs driven by AI investment. In December 2025 it certified nearly 10,000 megawatts of new generation. It also approved a formal process to evaluate grid-enhancing technologies that can squeeze more capacity from existing lines.

    Georgia has courted data centers since 2018 through sales and use tax exemptions and now hosts more than 200 operating facilities, most of them around Atlanta. Legislative support has cooled: a 2024 vote to suspend new exemptions was vetoed, and a 2025 state-commissioned evaluation found the incentive to be a net fiscal cost. Eminent domain, the power to take private property for public use with just compensation, is available to utilities building infrastructure. Georgia Power says it treats it as a last resort.

    Sources

    Source: ‘My childhood home is being taken’: Eminent domain for data centers targets properties across a thousand miles of grid expansion in Georgia (Fortune, July 26, 2026). The article reports on Georgia Power’s land acquisition for the Ashley Park-Wansley line and its broader 1,000-mile transmission plan.

  • LG’s 600kW Nvidia-Certified CDU Shows Why AI Cooling Is a Vetted-Vendor Market

    LG’s 600kW Nvidia-Certified CDU Shows Why AI Cooling Is a Vetted-Vendor Market

    TL;DR · 30-second read

    The Short Version

    LG, best known for televisions and fridges, just won approval from Nvidia, the company whose chips run most artificial intelligence systems, to supply equipment that keeps those chips cool.

    One cabinet of Nvidia’s newest chips uses roughly as much electricity as a hundred homes. That makes it far too hot for fans, so liquid has to be pumped right up against the chips.

    LG is the first Korean company approved for this. The approval matters because buyers rarely risk a leak next to millions of dollars of hardware.

    Startup Fortune reported that LG Electronics’ 600kW-class coolant distribution unit (CDU) has passed Nvidia’s final quality certification. That makes LG the first Korean company approved in this category of liquid cooling equipment for Nvidia-based AI server racks. LG announced the approval in late July 2026.

    LG says it will next seek certification for 1MW, 2.5MW and 4MW CDU models. It sells the unit as part of a ‘Chip to Chiller’ package spanning server-level cold plates through building-scale chillers.

    Executive Summary

    A coolant distribution unit is the pump-and-heat-exchanger that moves liquid between a data center’s cooling plant and the cold plates mounted on GPUs and CPUs. LG’s 600kW unit is now on Nvidia’s approved list. LG says it holds coolant temperature within ±0.25°C, runs redundant pumps and power supplies, and includes virtual sensors and leak detection.

    The approval matters because of rack density. Nvidia’s DGX GB200 NVL72 system draws about 120kW per rack, roughly four times the level at which air cooling alone stops working. At that density, liquid cooling is mandatory. It also sits millimetres from very expensive silicon, so buyers screen suppliers before any price is discussed.

    For LG, the certification also serves its shift toward business customers. It gets LG into the running for Nvidia-based deployments. It does not yet deliver orders: LG has named affiliate projects and earlier regional references, but no large external cloud contract.

    At 120kW a Rack, Air Is Out of the Picture

    Data Center Engineering puts the practical limit for air handling at roughly 30kW per rack. Above that, heat removal has to move closer to the chip. Nvidia’s own DGX GB200 system guide describes the NVL72 rack at approximately 120kW, with 72 Blackwell GPUs and 36 Grace CPUs in a single rack-scale design. For operators, liquid cooling at that density is a requirement, not a design choice.

    The CDU is the working heart of that liquid loop. It takes heat from coolant returning off the chips’ cold plates and passes it to the facility’s water system, while holding flow and temperature steady. On paper, a 600kW unit covers the heat of about five 120kW racks. That is a row-level building block, not a whole-hall system.

    Why Approval Comes Before Price

    This is the mechanism behind the headline. When the equipment being cooled costs millions of dollars per rack, a coolant leak or pump failure is a business interruption plus a large hardware bill, not a routine repair. Buyers therefore lean on Nvidia’s qualification as a first screen. The approved list narrows the field before bidding starts, so a cooling vendor’s first hurdle is getting qualified, and competing on price comes after.

    LG’s stated specifications match what that screen looks for: tight temperature control (±0.25°C), redundant pumps and power so cooling continues through a component failure, and leak detection. None of this is novel on its own. Together, with Nvidia’s approval attached, it lets LG compete for business it previously could not bid on.

    The effects run in several directions. Established cooling vendors gain a large, well-capitalised new competitor. Hyperscalers, the largest cloud operators, and other data center builders gain another qualified source at a time when liquid cooling demand is rising fast. Uncertified suppliers face a clearer barrier. One certification does not prove the whole market works this way, but it shows plainly where the entry gate for Nvidia-based cooling sits.

    From 600kW to 4MW: The Roadmap Is the Real Test

    LG plans to seek certification for 1MW, 2.5MW and 4MW CDUs next. The scale jump is significant: a 4MW unit could, in principle, serve the heat of more than thirty NVL72-class racks. That is the size at which large AI campuses buy cooling. The 600kW approval gets LG into the running. The megawatt-class approvals will decide whether LG can supply those campuses.

    LG frames the opportunity with a market forecast of about 26 percent annual growth, reaching roughly $133.5 billion (about 185 trillion won) by 2034. This is LG’s own projection, and it has not published the underlying methodology. Readers should treat it as the company’s view of market size, not an independent estimate. The direction is less in doubt than the number: Nvidia’s published rack power figures already require liquid cooling at the top end.

    An Appliance Maker’s Business-Customer Arithmetic

    The certification also fits LG’s broader repositioning. At a January 7 briefing, LG said business lines including B2B, webOS and direct-to-consumer grew from about 29 percent of revenue in 2021 to 45 percent in the second half of 2025. Their share of operating profit rose from 21 percent to about 90 percent. Data center cooling draws on work LG already does: HVAC, heat transfer and precision temperature control.

    The execution risk is real. LG’s stated pipeline includes AI data centers for group affiliate LG CNS, along with earlier cooling references in North America and Indonesia. These are credible starting points. They do not show that global cloud buyers, who usually demand worldwide service coverage and proven delivery at volume, will commit to LG for large deployments.

    Background

    LG Electronics built its brand on televisions, refrigerators and washing machines. For several years it has been shifting its growth strategy toward business customers, including commercial HVAC, and this year it said those business lines generate most of its operating profit. Precision cooling for data centers is a natural extension of that heat-transfer engineering.

    Nvidia’s Blackwell-generation rack-scale systems have made liquid cooling routine at the high end of AI infrastructure. The GB200 NVL72 combines 72 GPUs and 36 CPUs in one rack drawing about 120kW, far above the roughly 30kW at which air cooling reaches its limit. That has created fast-growing demand for direct-to-chip cooling components such as cold plates and coolant distribution units. Nvidia’s supplier qualification acts as the main gate into that supply chain.

    Sources

    Source: LG Electronics wins Korea’s first Nvidia liquid cooling certification as AI data centers hit a heat wall (Startup Fortune): LG’s 600kW coolant distribution unit clears Nvidia’s quality certification.