Amazon Web Services and ChatGPT owner OpenAI have entered into a multi-year strategic partnership valued at USD38 billion, one of the largest infrastructure deals in artificial intelligence to date.
The agreement enables OpenAI to leverage the advanced cloud and computing infrastructure of AWS to run and scale its AI workloads, including those powering ChatGPT.
Under the seven-year pact, OpenAI will gain access to hundreds of thousands of NVIDIA GPUs through AWS, with capacity to scale up to tens of millions of CPUs by 2027. This partnership underscores the surging global demand for computing power as AI models grow increasingly complex.
AWS, which operates some of the world’s largest and most secure AI clusters, will provide OpenAI with the performance, scalability, and reliability needed to advance its next-generation AI systems.
The infrastructure being built for OpenAI will feature sophisticated architectural designs optimized for AI performance, clustering NVIDIA GB200 and GB300 chips via Amazon EC2 UltraServers. This setup enables low-latency communication across interconnected systems, allowing OpenAI to efficiently handle diverse workloads—from training large-scale models to serving inference for ChatGPT.
“Scaling frontier AI requires massive, reliable compute,” said Sam Altman, OpenAI’s co-founder and chief executive officer. “Our partnership with Amazon strengthens the compute ecosystem that will power the next era of AI.”
“As OpenAI continues to push the boundaries of what’s possible, Amazon’s best-in-class infrastructure will serve as a backbone for their AI ambitions,” said Matt Garman, CEO
Of AWS.
All compute capacity under this agreement is expected to be deployed by end of 2026, with potential expansion through 2027 and beyond.
