Oracle has announced a plan to allocate roughly $40 billion to acquire advanced chips from Nvidia, intended to power a newly built data center for OpenAI.
Located in Abilene, Texas, the facility forms part of the U.S. Stargate Project, an initiative led by top artificial intelligence companies aiming to solidify the nation's position amid rising global competition.
According to inside sources, the center will incorporate close to 400,000 Nvidia GB200 chips, with Oracle set to secure the computing resources that will subsequently be made available for OpenAI’s operations. Neither organization has offered comments at this time, while a representative from Nvidia declined to provide further information.
The center is on track to become fully operational by the middle of next year, with Oracle committing to a 15-year lease on the property.
Financial support for this venture comes from several channels. A major portion of the funding is provided by a credit agreement with JPMorgan, which has arranged debt financing totaling $9.6 billion across two separate loan packages. In addition, the site’s owners—including Crusoe and the U.S.-based investment firm Blue Owl Capital—have contributed an estimated $5 billion in cash.
The project is designed to allow OpenAI to lessen its reliance on Microsoft, as the company’s demand for large-scale processing has exceeded what its current supporter can supply.
In a related development, Oracle, OpenAI, and Nvidia are teaming up on another notable facility in the United Arab Emirates. That project is expected to deploy in excess of 100,000 Nvidia chips, with the first phase scheduled to commence operations in 2026.
This expansion marks an effort to advance capabilities in cloud computing and artificial intelligence infrastructure, positioning those involved to remain competitive in a rapidly growing technology sector.