AI Infrastructure Models

Companies including SpaceX, Blue Origin, Google, Orbital, Starcloud, and Cowboy Space are advancing proposals or tests for solar-powered data centers in low Earth orbit. The approach could reduce dependence on terrestrial power, land, water, and permitting, but large-scale deployment still depends on cheaper and more frequent launches, lightweight cooling systems, radiation-tolerant hardware, reliable communications, and workable maintenance models. Current evidence points to small demonstrations and delay-tolerant or space-generated workloads before orbital facilities can complement, rather than replace, ground-based data centers.

Blackstone and Google have announced a U.S. joint venture to sell TPU-based compute and data center capacity to external customers, backed by $5 billion of equity and an initial 500 MW rollout targeted for 2027. The signal is consistent across sources: capital, power, sites, and specialized chips are being assembled to meet rising AI infrastructure demand.

SoftBank is planning a major AI data center expansion in France, with an initial €45 billion investment expected to deliver 3.1 GW across Dunkirk, Bosquel, and Bouchain by 2031, while broader plans could reach €75 billion and 5 GW. EDF and Schneider Electric are involved, including the conversion of a former power plant and the development of data center equipment production in Dunkirk. The plan highlights France’s nuclear-heavy electricity supply and existing industrial base as factors that may support large power-intensive projects, although the larger investment and capacity figures remain planned rather than fully committed.

Nvidia is backing data center expansion by investing in IREN and Nebius while supplying the systems and technology needed to scale AI compute. IREN is pairing secured power in Texas and Spain with Blackwell deployments, while Nebius is increasing capital spending, acquiring Eigen AI, and building capacity around large Meta and Microsoft commitments. The central constraint is execution: available power and contracted demand support growth, but hardware installation, financing needs, deployment speed, and margin pressure could limit how quickly that capacity becomes profitable.

Applied Digital and Hut 8 are securing long-term leases for large AI data center campuses, with commitments ranging from 210 MW to 704 MW and contract values extending into the billions of dollars. Applied Digital also reported bringing an additional 75 MW online at Polaris Forge 1, while debt financing and high-density cooling systems support the broader buildout. The agreements improve revenue visibility for developers but also increase capital requirements, construction execution demands, and exposure to a small number of large hyperscaler customers.

SpaceX is turning its AI data center capacity into a commercial business while preparing additional facilities in Texas and developing AI1 satellites for possible orbital computing. Agreements with Anthropic and Google would generate substantial recurring revenue and use large volumes of GPU capacity, while the terrestrial expansion is intended to support further growth. The orbital concept could address power and siting constraints, but its economics, insurance requirements, launch schedule, and deployment scale remain uncertain.

Digi Power X is building out a purpose-built AI data center campus in Alabama while scaling its NeoCloudz GPU-as-a-Service business. The company has secured major equipment, finalized grid interconnection, and is targeting phased capacity additions through 2027 alongside new NVIDIA GPU deployments. The topic matters because it shows a capital-intensive AI infrastructure ramp that depends on delivery timing, financing discipline, and successful conversion of planned capacity into revenue.

Infrastructure investors are committing substantial capital to data centers as demand shifts toward high-density, lower-latency AI workloads. I Squared Capital agreed to buy 10 Cogent facilities for $225 million and plans to invest another $1 billion, while a consortium led by AIP, MGX, and BlackRock’s GIP completed the approximately $40 billion purchase of Aligned and committed $5 billion for growth. The transactions expand investor control over sites with existing power, connectivity, and expansion potential, while making grid access, cooling, and deployment timelines increasingly important.

Meta is scaling its AI computing footprint through major data center projects, external capacity agreements, and in-house accelerator development. Reported plans include expanding the Hyperion campus in Louisiana from 2.2 to 5 gigawatts, securing about 1.6 gigawatts from Crusoe sites in Texas and Missouri, and deploying 7 gigawatts of computing infrastructure in 2026. The scale of spending and the possibility of unused capacity are pushing Meta to consider leasing compute to companies such as Anthropic, although those plans and related commercial arrangements remain preliminary.

Underwater and floating data centers are moving from demonstrations toward limited commercial deployment, led by a 24-megawatt facility near Shanghai that combines seawater cooling with nearby offshore wind power. The approach could reduce land requirements and cooling energy for coastal computing sites, but maintenance, subsea cables, permitting, reliability, environmental effects, and eventual scaling remain unresolved. Current evidence supports a niche role rather than replacement of conventional land-based campuses.

Kentucky is weighing large data center proposals that promise investment and jobs but raise concerns about electricity costs, land conversion, water use, emissions, transparency, and limited statewide siting rules. Several counties have responded with moratoriums, zoning restrictions, or project opposition, making data centers a consequential local political issue. At the same time, TeraWulf has secured a 20-year Anthropic lease for a planned 401-megawatt AI campus in Hawesville, showing that Kentucky is also attracting major committed AI capacity.

Cipher Digital is advancing multiple data center projects while shifting from Bitcoin mining toward AI and high-performance computing infrastructure. The company has identified Amazon Web Services as the tenant for its 70-megawatt Stingray facility and priced approximately $810 million of senior secured notes to fund remaining construction costs, reimburse prior equity contributions, and establish debt service reserves. Barber Lake has reached structural completion, while Black Pearl remains in retrofit and planning stages. The financing supports continued buildout but leaves project completion, leverage, and the closing of the notes offering as important execution considerations.

EdgeMode is negotiating the sale of several Spanish data center development sites to Pure Data Centres Group and a separate 300MW campus in Toledo to Spark AI Foundry. The proposed deals are tied to milestones such as Ready-to-Build status, administrative approvals, and power or capacity leases, with large implied build-out figures for the Spanish portfolio. The topic matters because it shows EdgeMode shifting from owning and developing sites toward monetizing them through staged transactions, while also highlighting the power and infrastructure constraints shaping Spanish data center projects.

Digital Realty is increasing its control of major Northern Virginia facilities while developing additional capacity in Virginia and Kansas City to serve hyperscale and AI workloads. Its $3.5 billion agreement with Blackstone covers three fully leased facilities totaling 288 megawatts, while other plans include a potential 13-building Digital Dulles campus and up to two gigawatts of power near Kansas City. The developments show how data center growth is being shaped by ownership, financing, grid access, cooling requirements, and the need to demonstrate the infrastructure behind AI services.

Data center developers and technology companies are turning to existing nuclear plants, proposed small modular reactors, and long-term power contracts as electricity demand rises and grid availability tightens. Microsoft, Google, Meta, and Riot Platforms are pursuing different versions of this approach, ranging from the planned restart of Iowa’s 615-megawatt Duane Arnold plant to early-stage reactor concepts for future AI campuses. The strategy could provide dependable power, but permitting, financing, construction timelines, supply chains, and local opposition make many nuclear-powered data center plans long-term rather than immediately available.

Oracle is rapidly expanding AI data center capacity, including a planned $16 billion campus in Saline Township, Michigan, to support customers such as OpenAI and other hyperscalers. The buildout is increasing capital spending, borrowing, and exposure to construction delays, power and water limits, permitting disputes, supply shortages, and uncertain customer demand. The projects could add substantial capacity and employment, but their financial and operational returns remain dependent on timely delivery, affordable energy, and sustained AI infrastructure demand.

OpenAI is reportedly negotiating to lease a proposed 10-gigawatt data center in southern Ohio, with SoftBank developing the site and Nvidia considering up to $250 billion in financial guarantees plus additional chip financing. The project would be built in phases on federally owned land near Piketon, with initial capacity targeted for 2028 and electricity access tied to U.S. government control and Japanese funding. The proposed structure illustrates how large AI facilities are increasingly being assembled through partnerships among model developers, chipmakers, investors, utilities, and governments, while leaving substantial questions about debt exposure, approvals, power availability, and commercial demand.

Hyperscalers, colocation firms and equipment suppliers are committing billions of dollars to expand AI and cloud data center capacity in India. Projects and investment plans center on Mumbai, Hyderabad, Gujarat and other emerging locations, with Meta, Amazon and RMZ among the most prominent participants. The expansion could lift India’s capacity from roughly 1.5–1.6 gigawatts today to about 6–7 gigawatts by 2030, while increasing demand for power, cooling, grid connections and renewable energy. Meta’s separate 1-gigawatt Alberta project shows the same AI-driven expansion pattern outside India, but also highlights construction, environmental and community constraints.

Large AI-focused data center projects are moving toward hundreds of megawatts and, in some cases, gigawatt-scale power requirements. Developers and investors are committing capital earlier through joint ventures, private debt, land facilities, GPU financing, and other structured arrangements, while utility interconnection delays, permitting, labor shortages, and community resistance slow delivery. The central question is whether power and construction capacity can arrive on schedule and whether AI demand will justify the scale of current investment.

AI-driven data center expansion is accelerating across the US and Asia, but growth is increasingly limited by power access, permitting, labor, cooling, fiber supply, and grid upgrades. New projects, forecasts, and contracts all point to a buildout that is now shaped as much by physical infrastructure as by compute demand.

Chevron and Microsoft are advancing Project Kilby, a planned natural-gas power facility in Reeves County, Texas, that would provide dedicated electricity to Microsoft’s expanding data center campus under a 20-year agreement. The project is expected to reach about 2.67 gigawatts, deliver first power in 2028, and reduce reliance on the local grid, although Chevron has not yet made its final investment decision. The same reporting points to broader interest from energy companies in serving data centers directly, alongside continuing concerns over emissions and local impacts such as noise.

Environmental groups are pressing the FCC to pause orbital data center licensing until it completes a broader environmental review of cumulative impacts, including pollution from launches and reentry, light pollution, and effects on astronomy and ecosystems.

Energy companies are expanding behind-the-meter natural gas generation to serve data centers as grid access and local capacity become constraints. Kodiak Gas Services and Baker Hughes are targeting up to 1.8 GW by 2030, while Williams and Blackstone are funding a roughly 2.6 GW portfolio across Ohio and Utah, linking power generation, pipelines, and data center demand. These projects could shorten power-delivery timelines and improve reliability for large computing customers, but their execution, capital requirements, and project schedules remain important variables.

Alaska is reviewing multiple proposals that could support large data centers, including Stak Energy’s North Slope project and a potential AIDEA-led industrial district near Houston. Stak’s plan would generate at least a gigawatt from nearby natural gas for AI and cloud customers, while the Houston proposal would reserve land for utility corridors, logistics facilities and possible data centers. The projects face unresolved questions about land transfers, public acceptance, water and gravel use, connectivity, financing and final customer demand.