The Largest Deal in History: OpenAI Signs $300 Billion Cloud Agreement with Oracle
⬤ The five-year agreement, set to begin in 2027, ranks among the largest cloud computing contracts ever.
⬤ OpenAI will require 4.5 gigawatts of power—enough to supply four million homes—as it scales its infrastructure.
⬤ OpenAI’s commitment far exceeds its current revenue, underscoring the risks tied to rapid AI expansion
OpenAI has signed a record-breaking deal with Oracle worth 300 billion dollars over five years to secure cloud computing capacity, underscoring both the explosive growth of artificial intelligence and the massive financial scale now attached to it. Sources told The Wall Street Journal that the agreement will take effect in 2027, making it one of the largest cloud computing contracts ever signed.
The contract dwarfs OpenAI’s current annual revenue of roughly 10 billion dollars, committing the company to spend an average of 60 billion dollars per year on cloud services. The scale of the deal highlights the enormous financial pressure facing OpenAI, which reportedly told investors last year that it does not expect to turn a profit until 2029, and could incur losses of up to 44 billion dollars before then.
The agreement also reflects the enormous energy demands required to train and operate advanced AI systems. OpenAI’s partnership with Oracle will require 4.5 gigawatts of electrical capacity roughly equivalent to the output of two Hoover Dam, or enough to power about four million homes.
Oracle CEO Safra Catz told analysts that the company’s cloud division is experiencing “unprecedented demand.” In a previous disclosure earlier this year, Oracle hinted at the OpenAI contract by forecasting more than 30 billion dollars in annual revenue from new cloud services beginning in 2027.
The deal comes as OpenAI continues to struggle with an ongoing shortage of computing capacity, which has slowed product launches and delayed next-generation AI models. These challenges are not unique to OpenAI; technology companies around the world are racing to build data centers and secure high-performance chips amid surging demand.
OpenAI has pursued several initiatives to expand its computing capabilities. It struck a deal with Broadcom to develop custom chips and partnered with SoftBank on a massive infrastructure project called Stargate, aimed at building a global network of AI data centers. Although that initiative has faced early delays, OpenAI now sees its collaboration with Oracle as a crucial part of the same broader strategy.
For years, OpenAI relied exclusively on Microsoft for computing infrastructure. As capacity limits began to tighten, the company sought approval to diversify its providers. The Oracle partnership marks a major shift and represents OpenAI’s largest move beyond the Microsoft ecosystem.
The deal carries significant risks for both companies. For OpenAI, it introduces a financial commitment that far exceeds its current revenues and places its future on the continued global adoption of ChatGPT and its enterprise products. The company also faces growing competition for AI talent, regulatory scrutiny, and delicate negotiations with Microsoft over the future of their partnership.
For Oracle, the challenge lies in the concentration of so much projected revenue in a single client. The company already carries a heavy debt load, with a debt-to-equity ratio of 427 percent, compared to Microsoft’s 32.7 percent. In its most recent fiscal year, Oracle generated 21.5 billion dollars in operating cash flow against 27.4 billion dollars in capital expenditures. By contrast, Microsoft produced 136 billion dollars in operating cash flow with 88 billion dollars in capital spending, putting it in a far stronger position to absorb the costs of AI infrastructure expansion.
The Oracle-OpenAI deal signals the next phase of the global AI infrastructure race, where access to computing power is becoming as strategic as the technology itself.
