Amazon Web Services signs USD38b deal to supply OpenAI with Nvidia chips
The news: Amazon’s cloud unit, Amazon Web Services (AWS) has agreed to a USD38 billion ($58.21 billion) deal to supply OpenAI with access to hundred of thousands of Nvidia graphics processing units (GPUs) as part of a seven-year arrangement.
The numbers: AWS said on Monday that under the new USD38 billion agreement OpenAI will access AWS compute comprising hundreds of thousands Nvidia GPUs, with the ability to expand to tens of millions of CPUs to rapidly scale agentic workloads.
The context: OpenAI will begin using AWS computing power effective immediately, it said on Monday, with all targeted capacity to be provided before the end of 2026. The deal also gives OpenAI the option to expand its partnership with AWS in the following years.
AWS will deploy hundreds of thousands of chips, including Nvidia GB200s and GB300s, in clusters to support training next generation AI models or to power ChatGPT’s responses to user prompts. OpenAI may also use the AWS data centres to power agentic AI.
While the deal is small compared the agreements OpenAI has reached with other cloud players including a USD300 billion deal with Oracle and a USD250 billion deal with Microsoft, the partnership is a key first step in Amazon’s efforts to become an AI leader.
Earlier on Monday, Microsoft announced a USD9.7 billion deal with IREN, to deal to buy AI cloud capacity from the Australian data centre operator, making it the Australian company’s largest customer.
What they 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."
The source: OpenAI press release