OpenAI has committed to a $300bn cloud computing contract with Oracle, expected to last nearly five years, according to a report by the Wall Street Journal citing undisclosed sources.

This deal represents one of the largest cloud agreements in history. It indicates a significant investment in AI data centres amid discussions about potential market saturation, according to the publication.

Scheduled to commence in 2027, the contract requires Oracle to provide 4.5 gigawatts (GW) of power capacity. This demand is comparable to the output of more than two Hoover Dams or the consumption of roughly four million homes.

Oracle’s share prices surged by 43% on 10 September after reporting an additional $317bn in future contract revenue for the quarter ending 31 August 2025. Oracle’s CEO Safra Catz disclosed that these figures resulted from new agreements with three different clients during that period.

The increase in share prices has notably impacted the wealth of the firm’s chairman, Larry Ellison, adding more than $100bn. This surge places him among the wealthiest individuals globally, alongside Elon Musk, with an estimated net worth approaching $400bn.

However; the agreement entails significant risks for both OpenAI and Oracle. OpenAI, currently a loss-making entity, reported annual revenues of approximately $10bn as of June 2025, which falls short of the average annual payment obligation of $60bn under this contract.

For Oracle, concentrating a substantial portion of future revenue on a single client may necessitate further debt to secure the AI chips required for its data centres.

Indications of this substantial deal first appeared in June when Oracle’s filings revealed a cloud services contract expected to generate over $30bn annually beginning in 2027. As additional data centres become operational, Oracle anticipates increasing annual revenue streams from OpenAI.

Previously reliant on Microsoft for computing resources, OpenAI has sought alternative providers due to supply limitations. In June, it partnered with Google Cloud to enhance its computational infrastructure critical for operating AI models. Furthermore, in March 2025, OpenAI entered into an $11.9bn contract with CoreWeave to bolster global AI infrastructure capabilities.

In parallel developments, OpenAI plans to produce its own AI chips by 2026 to address growing demand and reduce reliance on Nvidia, reported The Financial Times. Developed in collaboration with US semiconductor company Broadcom, these chips are expected to be shipped next year, according to sources familiar with the partnership.