OpenAI, Oracle and SoftBank announced plans to build five new artificial intelligence data centers across the United States, expanding their Stargate infrastructure initiative to meet rising demand for large-scale compute capacity. The companies said the new sites will be located in Shackelford County, Texas; Doña Ana County, New Mexico; Milam County, Texas; Lordstown, Ohio; and a fifth site in the Midwest that has not yet been disclosed. Three of the facilities will be jointly developed by Oracle and OpenAI, while two will be constructed through the partnership with SoftBank’s SB Energy.
The Stargate program was first unveiled in January during a White House event with a target of deploying 10 gigawatts of AI computing infrastructure across the country within four years. The initial plan called for up to $500 billion in investment. The five new facilities, combined with ongoing expansion at the flagship campus in Abilene, Texas, are expected to bring total planned capacity to nearly seven gigawatts. According to the announcement, the expansion at Abilene alone will add 600 megawatts to the project. The Abilene site is already partially operational on Oracle Cloud Infrastructure, with initial training and inference workloads underway.
Oracle has started deploying racks of Nvidia GB200 GPUs at the location and is building several large halls there to support additional compute demand. The companies said the construction of the new sites will generate more than 25,000 direct jobs during the build phase, in addition to thousands of indirect jobs in related industries. The new facilities are scheduled to come online in phases over the next three years, with the Lordstown site among the first expected to begin operations. SoftBank’s SB Energy unit will provide powered infrastructure for the two sites it is leading in Milam County and Lordstown.
Five US sites confirmed under latest Stargate expansion plan
The projects are designed to be scaled rapidly to keep pace with the deployment timeline. Oracle is responsible for building and operating the three other sites alongside OpenAI, leveraging its data center and cloud experience. The expansion is supported by a multibillion-dollar agreement with Nvidia, under which the chipmaker will invest up to $100 billion in the Stargate project while supplying its latest compute and networking hardware. Nvidia’s involvement aligns with the staged rollout of each gigawatt of infrastructure, ensuring that the new data centers will have immediate access to advanced AI training and inference hardware.
Stargate network advances four year capacity target in US
The announcement follows the companies’ earlier commitment to build what they described as one of the world’s largest dedicated AI data center networks. By adding five new sites across multiple states, the project advances toward its four-year capacity target. Company executives have highlighted the importance of securing adequate power and infrastructure to support the buildout. Each of the new facilities is being developed with access to large-scale electricity supply and connectivity to regional grids. The Stargate initiative is now positioned as one of the largest coordinated AI infrastructure programs in the United States.
With the addition of the new projects, the consortium of OpenAI, Oracle and SoftBank will expand its footprint across four states, reinforcing its role in scaling the country’s AI computing capacity and supporting the growing demand for artificial intelligence infrastructure nationwide while creating jobs, boosting local economies and advancing the Stargate initiative across the United States as part of one of the largest AI infrastructure programs ever launched. – By Content Syndication Services.