Oracle has agreed to lease a major new data center in Abilene, Texas, equipped with 400,000 of Nvidia’s GB200 chips, as part of a landmark partnership with OpenAI. The move underscores intensifying investment in AI infrastructure amid surging global demand for compute power.
The 875-acre complex will serve as the first U.S. site for OpenAI’s Stargate initiative, with the chips forming the core of a $40 billion hardware deployment. Once fully operational in mid-2026, the facility will deliver up to 1.2 gigawatts of power—placing it among the most energy-intensive data centers in the world.
Oracle has signed a 15-year lease for the entire facility, which is being developed by Crusoe, an AI infrastructure firm, and Blue Owl Capital. Financing for the project includes $9.6 billion in loans led by JPMorgan Chase, with a $7.1 billion tranche announced this week.
Following installation, the Nvidia hardware will be transferred to OpenAI, allowing the company to diversify its computing resources beyond Microsoft’s Azure platform. The two companies ended their cloud exclusivity agreement earlier this year as rising demand outpaced Microsoft’s capacity.
“The Stargate hub in Texas is a critical step toward building out OpenAI’s independent compute capacity and ensuring resilience as AI workloads scale,” one person familiar with the matter said.
The Abilene site is one of several hyperscale developments underway. Elon Musk is expanding a similar complex in Memphis that will host up to one million Nvidia GPUs, while Amazon (AMZN.O) is planning a facility in Virginia that will exceed 1 gigawatt of power.
Backers of Stargate—including OpenAI, Oracle, SoftBank, and Abu Dhabi’s MGX—have signaled ambitions to invest up to $500 billion in AI infrastructure globally over the next four years. A companion site in Abu Dhabi is already in planning stages.
The massive order of GB200 chips is one of the largest single AI hardware procurements to date. Analysts say demand for Nvidia’s high-performance silicon is being driven by governments and enterprises racing to develop next-generation AI models.
Nvidia’s recent surge in market value, including a brief stint as the world’s most valuable public company in late 2024, reflects its dominant position in the global AI chip market.
Source: Financial Times