OpenAI has entered a landmark seven-year deal worth around $38 billion with Amazon Web Services (AWS) to support its expanding artificial intelligence infrastructure needs. Through the agreement, OpenAI gains access to hundreds of thousands of state-of-the-art Nvidia GPUs, along with the option to scale into “tens of millions” of CPUs, enabling budgets large enough to train and run so-called “agentic AI” systems.
The deal signals a significant strategic shift: OpenAI is diversifying away from its earlier exclusive reliance on Microsoft and Azure, choosing AWS as a major infrastructure partner. Implementation begins immediately, with full capacity deployment targeted by end of 2026 and further expansion into 2027 and beyond.
The agreement underscores the accelerating race for compute power in the AI industry—and raises questions about how such massive infrastructure commitments will be funded and sustained.