OpenAI will lease approximately 4.5 gigawatts of data centre power from Oracle under a $30 billion‑per‑year agreement set to begin delivering revenue in Oracle’s fiscal year 2028. This significant capacity boost forms part of the expansive “Stargate” initiative—an AI infrastructure venture launched in January aiming to deploy up to $500 billion globally.
Oracle plans to construct and expand multiple U. S. data centres to support this deal, including enhancements to its 1.2 GW facility in Abilene, Texas, scaling it to 2 GW. Additional site candidates span Texas, Michigan, Wisconsin, Georgia, New Mexico, Ohio and Pennsylvania. By comparison, the 4.5 GW allotted to OpenAI constitutes roughly a quarter of total U. S. data centre capacity, enough to power millions of homes.
The deal elevates Oracle’s cloud infrastructure business significantly—its existing data centre revenue in fiscal 2025 was $10.3 billion, making the OpenAI agreement nearly triple in scale. Oracle intends to support this growth with investments including $7 billion into Stargate and $25 billion in capital expenditures in 2026, alongside a major Nvidia GB200 chip acquisition valued at $40 billion.
Oracle chief executive Safra Catz previously disclosed in a Securities and Exchange Commission filing that the company had signed multiple large cloud services contracts expected to generate over $30 billion annually from fiscal 2028. Analyst commentary suggests this contract with OpenAI constitutes net-new revenue—significant new business rather than renewals.
Investor response has been strongly positive. Oracle’s share price surged to record levels, buoyed by expectations of over 50 per cent year‑on‑year revenue growth in fiscal 2028 and projections of surpassing $104 billion in revenue by 2029. TD Cowen analysts raised their price target from $250 to $275, while Citizens JMP maintained an outperform rating with a $240 target, noting Oracle’s “sophisticated” GPU cluster offerings.
The Stargate initiative, originally unveiled at a White House event in January 2025 alongside President Trump, is a collaboration between OpenAI, Oracle, SoftBank and Abu Dhabi’s MGX sovereign fund. So far, $50 billion has been committed by founding partners, with long-term funding targets set at $500 billion for global AI infrastructure development.
OpenAI’s strategy reflects a deliberate move to diversify its compute ecosystem beyond Microsoft Azure, its largest investor and former exclusive cloud provider. The firm has since added other partners—including Google Cloud and CoreWeave—alongside the new Oracle deal.
Industry analysts view this sprawling infrastructure expansion as essential to maintain leadership in the escalating AI arms race. Larry Ellison, Oracle co‑founder, has emphasised the company’s ambition: “We will build and operate more cloud infrastructure data centres than all of our cloud infrastructure competitors”.
Yet the sheer scale of this power draw raises questions around sustainability and resource allocation. A related analysis suggested a 2 GW facility in West Texas could consume about 40 million litres of water daily—placing strain on local supplies already under pressure.
As the U. S. data centre landscape evolves rapidly, Oracle and OpenAI’s partnership underscores the escalating importance of infrastructure in AI advancement. The unfolding execution of this agreement will shape both companies’ market positions and broader resource landscapes, even as the narrative of artificial intelligence continues to accelerate globally.
