Oracle is set to invest approximately $40 billion in Nvidia’s top-tier GB200 chips to support the development of OpenAI’s massive new data center in the United States.
According to a report by the Financial Times, located in Abilene, Texas, the facility is part of the Stargate Project — a strategic U.S.-led initiative aimed at strengthening the nation’s position in the rapidly advancing global artificial intelligence landscape.
The cloud services giant will purchase around 400,000 of Nvidia’s most advanced chips and lease the computing capacity to OpenAI, the report said, citing sources familiar with the deal. OpenAI and Oracle have not responded to requests for comment, while an Nvidia representative declined to speak on the matter.
The data center, expected to be fully functional by mid-2025, will be leased by Oracle for a 15-year term, according to the report.
Financial backing for the project includes $9.6 billion in debt financing from JPMorgan across two loans, and about $5 billion in equity investment from site owners Crusoe and U.S. investment firm Blue Owl Capital.
This massive infrastructure push will help OpenAI reduce its reliance on Microsoft — its largest financial supporter — as demand for computing power continues to exceed what Microsoft alone can supply.
For Oracle, the deal signals a major leap forward in its ambitions to expand its cloud computing footprint and better compete with dominant players like Microsoft, Amazon, and Google.
The U.S. Stargate project is not Oracle and OpenAI’s only collaboration. The companies, along with Nvidia, are also participating in a similar initiative in the Middle East. A sprawling AI data center is planned for the United Arab Emirates, which will likely require more than 100,000 Nvidia chips. The first phase of that facility is expected to be operational by 2026.