OpenAI has entered into a massive $30 billion annual agreement with Oracle to boost its artificial intelligence infrastructure capabilities. Under this deal, OpenAI will gain access to cloud computing power totaling 4.5 gigawatts, making it one of the largest cloud technology contracts in the AI sector to date.
The agreement also expands OpenAI’s “Stargate” project, launched earlier in 2025 in partnership with SoftBank, aiming to build massive global data centers. The total investment for Stargate could reach up to $500 billion, with about $50 billion already invested.
Key partners in the project include Oracle, SoftBank, and Abu Dhabi’s sovereign wealth fund MGX, though details on the funds spent so far have not been publicly disclosed.
Oracle plans to construct new data centers across multiple U.S. states such as Texas, Michigan, Wisconsin, Wyoming, New Mexico, Georgia, Ohio, and Pennsylvania. Additionally, the company will expand the 1.2-gigawatt Abilene data center in Texas, developed and financed by Crusoe.
As part of the deal, Oracle will order approximately 400,000 Nvidia GB200 high-performance AI chips, which will be used to train and run AI models.
Oracle founder Larry Ellison stated in a recent investor briefing:
“Oracle will become the largest builder and operator of cloud infrastructure data centers worldwide in the coming years.”
Oracle has committed an initial $7 billion investment to this venture, with plans to spend up to $25 billion by 2026.
