The Oracle-OpenAI $300 Billion Cloud Deal

Oracle and OpenAI just announced a five-year, $300 billion cloud computing agreement, one of the largest contracts in technology history. Beginning in 2027, Oracle will provide OpenAI with vast computing resources to support AI research and model development. This builds on their earlier July 2025 partnership to add more than 5 gigawatts of new U.S. data center capacity under the Stargate project, equivalent to the energy needs of millions of households.

Under the terms of the contract, OpenAI is expected to spend an average of $60 billion per year on Oracle’s cloud infrastructure, gaining access to millions of GPUs and expanded facilities for training advanced AI systems. The arrangement diversifies OpenAI’s computing base beyond Microsoft Azure and aligns with broader efforts to scale AI infrastructure, which include large-scale chip purchases and data center investments. Oracle, for its part, reported a significant increase in future revenues and a sharp rise in its market valuation following the announcement.

The deal, however, raises concerns about feasibility. OpenAI’s projected 2025 revenue is roughly $13 billion, far below the annual annual spending required, leading to questions about funding sources. Oracle also faces challenges in securing sufficient hardware, financing expansion, and addressing power grid constraints. Critics caution that the commitments may be unrealistic or overstated, reflecting broader uncertainty about the economics of scaling AI at this magnitude.