OpenAI Expands US AI Data Center Footprint with Stargate Initiative
Key Points
- OpenAI will add five new U.S. data centers under the Stargate program.
- Partnerships include Oracle for three sites and SoftBank’s SB Energy for two sites.
- Total U.S. capacity will approach seven gigawatts, with a 600‑MW expansion at Abilene.
- Projected job creation exceeds 25,000 onsite positions.
- Oracle’s Abilene facility will house over 400,000 GPUs and more than 1.4 GW of power.
- Nvidia partnership aims to deliver an additional 10 GW of capacity, with up to $100 B investment.
- Stargate aligns with U.S. AI competitiveness goals and receives support from local officials.
- International Stargate projects announced for the UK and UAE.
- Industry experts warn of environmental and accounting complexities.
- Recent reports cite trade uncertainty and hardware costs as potential delays.
OpenAI announced plans to add five new data centers across the United States under its Stargate program, partnering with Oracle and SoftBank. The expansion will bring total U.S. capacity close to seven gigawatts and includes sites in Texas, New Mexico, Ohio, and a yet‑to‑be‑named Midwest location. Oracle will operate the flagship Abilene facility, while SoftBank’s SB Energy backs two of the new sites. The project is tied to broader U.S. AI competitiveness goals, with expectations of thousands of jobs and additional collaborations with Nvidia and international partners.
Stargate Expansion Overview
OpenAI disclosed a plan to construct five additional data centers in the United States as part of its Stargate infrastructure program. The new sites are being developed in collaboration with Oracle and SoftBank, bringing the total planned capacity in the United States to nearly seven gigawatts. OpenAI CEO Sam Altman emphasized the scale of AI infrastructure compared with the internet and highlighted the Texas model as a template for rapid, cost‑effective expansion.
Location and Partnerships
Three of the new facilities are slated for Shackelford County, Texas; Doña Ana County, New Mexico; and an undisclosed location in the Midwest. These projects are being pursued together with Oracle, following a prior agreement that aims to add up to 4.5 gigawatts of U.S. data‑center capacity beyond the existing Stargate site in Abilene, Texas. The remaining two sites will be led by OpenAI and SB Energy, a SoftBank subsidiary focused on solar and battery development, and are located in Lordstown, Ohio, and Milam County, Texas.
Economic Impact and Job Creation
OpenAI asserts that the new data centers, together with a planned 600‑megawatt expansion of the Abilene facility, will generate more than 25,000 onsite jobs. The flagship Abilene site is primarily owned and operated by Oracle, with OpenAI as the main tenant. The build‑out, managed by data‑center startup Crusoe, is projected for completion by mid‑2026 and already runs on Oracle Cloud Infrastructure, supporting both training and inference workloads.
Infrastructure Details
Oracle is constructing eight data‑center halls in Abilene, each expected to support roughly 100 megawatts of power. One hall is complete, and another is nearing finalization. When fully operational, the Abilene facility will house more than 400,000 GPUs and provide over 1.4 gigawatts of power. OpenAI’s site selection process reviewed more than 300 proposals from roughly 30 states, after surveying around 700 potential locations nationwide.
Strategic Context and Government Support
The Stargate program aligns with broader U.S. policy initiatives aimed at maintaining AI leadership. The project was announced during a period of heightened emphasis on AI infrastructure by the Trump administration, which introduced an AI action plan encouraging rapid development and reduced regulatory barriers. Local officials, including the mayor of Abilene and Senator Ted Cruz, expressed strong support for the initiative, advocating for a light‑touch regulatory approach to keep pace with international competitors.
Additional Partnerships and International Expansion
OpenAI also revealed a strategic partnership with Nvidia to deploy an additional 10 gigawatts of data‑center capacity, with Nvidia committing up to $100 billion in investment. The first Nvidia‑powered system, using Vera Rubin GPUs, is expected to become operational in the second half of 2026. Internationally, OpenAI announced Stargate projects in the United Kingdom and the United Arab Emirates, with assistance from U.S. administration officials to facilitate allied‑nation deals.
Industry Perspectives and Challenges
Jonathan Koomey, a professor at UC Berkeley who studies data‑center efficiency, noted that while scale is a central strategy for OpenAI, the environmental and accounting complexities of such large‑scale projects warrant careful consideration. Recent reporting from Bloomberg and The Wall Street Journal highlighted potential delays linked to trade uncertainty and rising hardware costs, underscoring the challenges inherent in rapidly expanding AI infrastructure.