Amazon Web Services (AWS) and OpenAI have entered a long-term strategic partnership valued at $38 billion, aimed at expanding the computing capacity required for training and deploying large-scale AI models.
Under the terms of the agreement, OpenAI will gain access to AWS’s high-performance cloud infrastructure, including Amazon EC2 UltraServer clusters equipped with hundreds of thousands of NVIDIA GPUs. The setup is designed to scale to tens of millions of GPU cores as model training workloads grow. Deployment is expected to be completed by the end of 2026, with expansion planned into 2027.
OpenAI CEO Sam Altman described the collaboration as a pivotal step for the industry:
“Building the next generation of AI models requires massive and reliable computing power. Our partnership with AWS strengthens the broader ecosystem that will fuel the next era of artificial intelligence — making it more accessible and impactful.”
AWS CEO Matt Garman emphasized that Amazon’s cloud division is positioned to become the underlying infrastructure for OpenAI’s future breakthroughs:
“As OpenAI continues to push the boundaries of what’s possible, AWS’s global compute infrastructure will serve as the foundation for their AI ambitions.”
The new compute environment is engineered to minimize latency between distributed components — a critical factor for both training frontier models and serving high-volume inference workloads, such as ChatGPT.
This collaboration builds on the earlier integration of OpenAI Foundation Models in Amazon Bedrock, which allows thousands of AWS customers — including Peloton, Thomson Reuters, Comscore, and Verana Health — to deploy OpenAI models for automation, analytics, and research.
The deal comes amid rapid growth at OpenAI. Recent reports suggest the company is exploring a public listing with a potential valuation approaching $1 trillion. Altman also recently stated that OpenAI’s annual revenue is now “well above $13 billion”.
Meanwhile, Amazon has been accelerating its automation strategy — the company is preparing to replace more than 600,000 warehouse positions with robotics, aiming to automate up to 75% of its logistics operations by 2027.
