
OpenAI is significantly expanding its access to computing infrastructure in the United States through a major agreement with Oracle, a move that underlines the soaring demand for artificial intelligence compute power and highlights OpenAI’s long-term strategy under its Stargate initiative.
According to sources familiar with the arrangement, OpenAI will lease approximately 4.5 gigawatts of data center capacity across the U.S. from Oracle, making this one of the largest known infrastructure procurement deals for AI operations to date. This scale of power allocation is comparable to that of large-scale utility projects and represents a substantial investment in next-generation AI infrastructure.
The partnership began gaining attention earlier this year when the White House mentioned Stargate as part of broader AI development efforts. As part of the project, Oracle and development partner Crusoe Energy have already established a large data center in Abilene, Texas, which currently operates at around 1.2GW and is expected to scale up to 2GW. Additional U.S. locations under consideration for expansion include Michigan, Wisconsin, Wyoming, and Texas, with exploratory efforts underway in New Mexico, Georgia, Ohio, and Pennsylvania.
The Stargate initiative is positioned as a multiyear strategy to ensure OpenAI has the computational resources needed to scale its models, meet global demand, and support ongoing development of generative AI systems. Although specifics may evolve, sources confirm that the current and planned expansions fall under the Stargate umbrella. Neither OpenAI, Oracle, nor Crusoe have provided official comments on the reported details.
Financial markets responded favorably to the news. Oracle shares rose by as much as 3.9% on July 2, hitting an all-time high. The company has seen a 36% increase in its stock price in 2025, driven largely by enthusiasm surrounding its cloud business and its pivot toward AI-focused customers.
AI-focused Enterprises
Earlier this week, Oracle revealed it had signed a $30 billion cloud agreement to begin generating revenue in its fiscal 2028 year. While Oracle did not disclose the customer, insiders suggested the Stargate arrangement with OpenAI forms a significant part of that deal.
Oracle, historically known for its database software, has been steadily gaining ground in the cloud services sector by targeting enterprises engaged in AI-related workloads. The company’s recent efforts, including partnerships with Nvidia, Cisco Systems, and G42 – a UAE-based AI firm backed by the country’s sovereign wealth fund – reflect its broader ambition to position itself at the center of AI’s infrastructure revolution.
In May, OpenAI confirmed it was also working with Oracle, Crusoe, Nvidia, and G42 on a Stargate data center project in the United Arab Emirates, signaling that its infrastructure ambitions are global in scope.
With AI workloads requiring increasingly massive processing resources, OpenAI’s partnership with Oracle marks a notable shift in how hyperscale infrastructure is being developed and deployed to support the next phase of artificial intelligence innovation.