OpenAI and Oracle ink a staggering $300 billion deal for a Stargate-sized AI future

OpenAI has reportedly signed a deal to purchase a colossal $300 billion in computing power from Oracle over roughly five years, in one of the largest cloud contracts ever conceived. First reported by The Wall Street Journal, the agreement is a cornerstone of the ambitious Project Stargate and further solidifies Oracle’s surprising comeback as a major player in the AI infrastructure wars. The contract is set to begin in 2027 and is already sending shockwaves through the market, catapulting Oracle’s stock and its founder, Larry Ellison, to new heights.

The announcement sent Oracle’s shares soaring more than 40 percent, its largest single-day gain since 1992. During a quarterly earnings call, Oracle CEO Safra Catz revealed the company had signed four multi-billion-dollar contracts with three unnamed companies in Q1, contributing to what one analyst called an “astonishing quarter.”

Oracle rides the AI wave to the top

The deal is the most visible part of a massive windfall for Oracle. The company announced its backlog of future contract revenue, or remaining performance obligations, ballooned to $455 billion, largely thanks to AI-driven demand. Catz issued an incredibly aggressive forecast, projecting Oracle’s cloud infrastructure revenue will grow from $18 billion this fiscal year to $144 billion by 2030. The market reaction was immediate, adding nearly $100 billion to co-founder Larry Ellison’s net worth in a single morning and making him the world’s richest person. “There’s no better evidence of a seismic shift happening in computing than these results that you just put up,” said Deutsche Bank analyst Brad Zelnick during the call.

Building the Stargate

This $300 billion commitment is a key piece of the even larger Project Stargate, an initiative to build massive AI data centers. OpenAI, Oracle, and Japanese conglomerate SoftBank initially revealed the project in January alongside President Trump, with a commitment to spend at least $100 billion on computing infrastructure in the US, with plans to potentially expand that to $500 billion. Construction is already underway on facilities in Abilene, Texas, as part of the effort to satisfy the AI industry’s insatiable appetite for power and processing.

OpenAI isn’t betting on just one horse

For OpenAI, the deal marks a significant strategic diversification. The AI giant has been moving away from its exclusive reliance on its primary partner and investor, Microsoft Azure, for cloud computing. Besides this massive Oracle deal, OpenAI also reportedly inked a cloud agreement with rival Google this spring and is said to be working with semiconductor giant Broadcom on a custom AI chip. This multi-pronged strategy underscores just how much compute the company believes it will need to stay at the forefront of AI development.

Under the hood, a few warning lights are flashing

Despite the headline-grabbing numbers and soaring stock, Oracle’s success comes with caveats. The company’s latest quarterly results were otherwise mixed, with revenue and profit slightly missing analyst expectations. To fund this massive data center build-out, Oracle expects to spend $35 billion on capital expenditures this year, which will lead to negative free cash flow. Furthermore, some analysts are cautious, with one noting that some business comes from rivals like Microsoft and Google who are “offloading their capacity to other data center providers,” suggesting they aren’t all “organic customers to Oracle.” This expensive arms race is also happening while the actual business returns on AI remain unproven. An MIT study found that 95 percent of organizations investing in generative AI are seeing zero return, raising questions about how long this level of spending can be sustained.