AWS and OpenAI Sign $38 Billion Multi-Year Compute Deal

Amazon Web Services and OpenAI have signed a $38 billion, multi-year strategic partnership that gives OpenAI immediate access to AWS infrastructure to run and scale its core AI workloads.

The deal, announced today, commits OpenAI to a $38 billion agreement with continued growth over the next seven years. OpenAI starts using AWS compute immediately, with all capacity targeted for deployment before the end of 2026 and options to expand further into 2027 and beyond.

The arrangement matters beyond the two companies. The rapid advancement of AI technology has created what AWS describes as unprecedented demand for computing power, and frontier model providers are increasingly turning to AWS for performance, scale, and security as they push their models toward higher levels of intelligence. A deal of this size signals where a substantial share of the industry's compute spending will flow through the end of the decade.

The hardware behind the deal

AWS will provide OpenAI with Amazon EC2 UltraServers featuring hundreds of thousands of state-of-the-art NVIDIA GPUs — both GB200s and GB300s — with the ability to scale to tens of millions of CPUs for OpenAI's advanced generative AI workloads.

The infrastructure deployment features an architectural design optimized for AI processing efficiency and performance. Clustering the NVIDIA GPUs via EC2 UltraServers on the same network enables low-latency performance across interconnected systems, according to the announcement. The clusters will support a range of workloads: serving inference for ChatGPT, training next-generation models, and scaling what AWS calls agentic workloads.

AWS says it brings unusual experience to the task, noting it already operates clusters topping 500,000 chips securely and reliably at scale.

What the executives said

"Scaling frontier AI requires massive, reliable compute," said OpenAI co-founder and CEO Sam Altman. "Our partnership with AWS strengthens the broad compute ecosystem that will power this next era and bring advanced AI to everyone."

"As OpenAI continues to push the boundaries of what's possible, AWS's best-in-class infrastructure will serve as a backbone for their AI ambitions," said Matt Garman, CEO of AWS. "The breadth and immediate availability of optimized compute demonstrates why AWS is uniquely positioned to support OpenAI's vast AI workloads."

An existing relationship, deepened

The partnership builds on work already underway between the two companies. Earlier this year, OpenAI's open weight foundation models became available on Amazon Bedrock, bringing additional model options to millions of AWS customers.

OpenAI has quickly become one of the most popular publicly available model providers in Amazon Bedrock. Thousands of customers — including Bystreet, Comscore, Peloton, Thomson Reuters, Triomics, and Verana Health — use its models for agentic workflows, coding, scientific analysis, and mathematical problem-solving.

The stakes

For OpenAI, the deal rapidly expands compute capacity while, per the announcement, benefitting from the price, performance, scale, and security of AWS. For AWS, landing one of the largest AI workloads in the industry — the infrastructure behind ChatGPT, which serves millions of users — is a competitive win in the cloud market where hyperscalers compete aggressively for frontier-lab contracts.

AWS positions the combination of its cloud infrastructure leadership and OpenAI's generative AI advancements as a way to help millions of users continue getting value from ChatGPT.

With capacity ramping through 2026 and expansion possible into 2027 and beyond, the agreement locks in a significant piece of OpenAI's compute roadmap on AWS for years — and ties the availability of ChatGPT's next generation of models, in part, to the pace of Amazon's data center buildout.