Oracle Strikes Over $30 Billion/Year to Power OpenAI’s AI Workloads
In a major milestone for the cloud industry, Oracle has secured a $30 billion/year agreement with OpenAI, further increasing as data centers come online, to provide the computing power behind the next wave of artificial intelligence. The deal, announced in July 2025, will see OpenAI use Oracle’s cloud infrastructure to support its rapidly growing demand for data center capacity.
Some headlines reported it as $300B which reconciles as follows: $30B/yr = Oracle’s disclosed annual run-rate beginning FY28; $300B/5y = reported total value of the full, ramping commitment. Same deal, different lenses.
Starting in Oracle’s fiscal year 2028, the contract commits OpenAI to using Oracle’s cloud services at an unprecedented scale — around 4.5 gigawatts of capacity. That’s the equivalent power of multiple massive data centers, all working behind the scenes to train and run AI models.
The partnership is part of OpenAI’s broader “Stargate” initiative — a bold plan to build out one of the largest AI computing networks in the world, with backing from investors like SoftBank and MGX. For Oracle, it’s a chance to prove that it can go head-to-head with giants like Amazon Web Services and Microsoft Azure.
To meet the demand, Oracle plans to expand existing sites like its Abilene, Texas hub and build new infrastructure across several U.S. states, including Michigan, Wyoming, and New Mexico.
This isn’t a funding round or an investment — no money changed hands between investors and Oracle. Instead, it’s a commercial agreement that guarantees Oracle billions in future revenue, and gives OpenAI the capacity it needs to keep pushing the boundaries of what artificial intelligence can do.
In a world where AI is reshaping industries, this deal marks a major turning point — not just for Oracle and OpenAI, but for the cloud ecosystem as a whole.
Source:
A.M.
Revision by B.S.
