Overview

OpenAI has identified five U.S. locations for new AI data centers tied to the Stargate initiative, a nationwide buildout of computing capacity developed in partnership with Oracle and SoftBank. The projects are expected to add roughly 7 gigawatts of capacity and advance plans well ahead of the year-end timeline the companies set.

OpenAI’s announcement this week confirmed the two technology giants are behind the massive project, which was initially proposed by lesser-known BorderPlex Digital Assets out of Austin and its development partner Stack Infrastructure, which is based in Colorado.

Launched in January by President Donald Trump, Stargate is designed to accelerate U.S. AI infrastructure with a target of $500 billion in total investment and 10 gigawatts of compute. OpenAI says its first wave of sites puts the program within striking distance of that goal.

Oracle partnership and deployments

Three facilities will be developed with Oracle in Shackelford County, Texas; Doña Ana County, New Mexico; and an undisclosed Midwest location to be named later. The Oracle collaboration—part of a larger $300 billion tie‑up with OpenAI—also includes expanding the flagship Stargate hub in Abilene, Texas.

• Three new Oracle‑aligned sites: Shackelford County (TX), Doña Ana County (NM), Midwest (TBA)
• Abilene, TX: flagship hub expansion as part of the same collaboration
• Estimated added capacity from Oracle deployments: ~5.5 GW

SoftBank-backed sites and scaling

Two more data centers are planned with SoftBank in Lordstown, Ohio, and Milam County, Texas. OpenAI says the SoftBank‑backed sites can scale to 1.5 gigawatts over the next 18 months.

• Planned locations: Lordstown (OH) and Milam County (TX)
• Projected scaling horizon: ~18 months to 1.5 GW
• Combined with Oracle projects: ~7 GW total—three short of the 10 GW Stargate target

Employment and economic impact

Beyond compute, the companies project significant employment gains. OpenAI expects more than 25,000 on‑site roles across the confirmed facilities and tens of thousands of jobs nationwide when including construction and related supply chains. Additional locations are under evaluation, suggesting further expansion to come.

Site selection criteria

OpenAI received over 300 submissions from prospective host communities in 30 states, and the chosen sites went through a multi‑stage selection process. Factors typically considered for hyperscale builds include:

• Access to power and water; renewable energy options
• Grid interconnects and transmission capacity
• Land availability and permitting readiness
• Long‑haul fibre connectivity and network diversity
• Local workforce readiness and training pipelines

Vision and next steps

OpenAI CEO Sam Altman called the effort “historic progress” toward broad access to AI, arguing that expanding compute is essential to unlock future advances. In a personal blog post, he suggested that at the scale of 10 gigawatts, AI could help tackle major challenges such as accelerating cancer research or delivering personalized tutoring globally.

Altman also described a longer‑term ambition to standardize deployment—likening the approach to a factory model that could add about a gigawatt of AI infrastructure each week.

The future outlook

If the current slate of projects proceeds on schedule, Stargate will move substantially closer to its national capacity objective while concentrating major investments in Texas, New Mexico, Ohio, and a forthcoming Midwest site. The buildout underscores how AI leaders and cloud partners are racing to secure power, land, and supply chains to meet surging demand for advanced model training and inference.