In what is shaping up to be one of the most consequential infrastructure moves in artificial intelligence history, OpenAI has reportedly inked a massive cloud computing deal with Oracle—worth an estimated $30 billion annually—to secure 4.5 gigawatts of compute capacity.
The agreement, among the largest of its kind, marks a significant expansion of OpenAI’s Stargate initiative, a global data centre program developed in collaboration with SoftBank, Oracle, and Abu Dhabi sovereign wealth fund MGX. Stargate is designed to lay the physical foundation for future frontier AI models by building out ultra-scale infrastructure capable of training and running massive neural networks.
While neither OpenAI nor Oracle have officially confirmed the agreement, signs of the deal surfaced after Oracle’s July announcement of a long-term cloud contract set to generate annual revenue beginning in 2028. If confirmed, the value would triple Oracle’s 2025 data centre infrastructure revenues—a sharp pivot into hyperscale AI infrastructure.
Why This Matters
The computing power in question—4.5GW—is staggering. That’s roughly 25% of all current operational data centre capacity in the U.S., and underscores the scale of demand as AI moves beyond experimentation into industrial-scale deployment.
Under the terms of the Stargate venture, Oracle is expected to build out multiple sites across the U.S., with priority regions reportedly including Texas, Ohio, Georgia, Michigan, New Mexico, Wisconsin, Wyoming, and Pennsylvania.
One anchor facility will be a 1.2GW site in Abilene, Texas, which Oracle is enhancing in collaboration with Crusoe, a data centre start-up specializing in energy-efficient AI infrastructure. Oracle also plans to integrate approximately 400,000 Nvidia GB200 chips, in a hardware deal alone reportedly worth $40 billion, aimed at powering next-generation AI workloads.
The Strategic Stakes
This move signals Oracle’s most aggressive push yet to challenge Microsoft Azure and Amazon Web Services in the race to become the AI era’s go-to cloud provider. The company is set to invest over $25 billion in CapEx next year, far surpassing previous forecasts, and has already committed $7 billion specifically for Stargate.
Oracle founder Larry Ellison has made no secret of the company’s ambitions: “We will build and operate more cloud infrastructure data centres than all of our cloud infrastructure competitors,” he declared earlier this year.
For OpenAI, this scale of investment is a direct reflection of the needs of frontier model development—large language models that require exponential amounts of compute. Stargate is designed to deliver that at unmatched scale, with global reach and sovereign partners that ensure redundancy, resilience, and regulatory compliance.
IMAGE: OpenAI
