AWS will power OpenAI’s next-generation model in a $38 billion deal as AI companies move beyond Microsoft.
summary
- OpenAI signs $38 billion deal with Amazon, giving it access to the company’s computing
- AWS uses UltraServers with Nvidia GPUs for OpenAI models
- Skeptics warn of similar trades, increasing signs of AI bubble
Tech giants are signing multibillion-dollar deals to power their AI models as demand for powerful infrastructure explodes. On Monday, November 3, OpenAI signed a seven-year, $38 billion deal with Amazon Web Services. This agreement gives OpenAI access to the vast computing power of AWS.
“Scaling frontier AI requires reliable computing at scale,” said Sam Altman, co-founder and CEO of OpenAI. “Our partnership with AWS strengthens the broader computing ecosystem that powers this next era and brings advanced AI to everyone.”
OpenAI will have access to thousands of Nvidia GPUs. These include the GB200 and GB300 series of Amazon EC2 UltraServers, which are optimized for AI computing. According to Amazon, this infrastructure can scale to tens of millions of CPUs for both model training and execution.
“As OpenAI continues to push the boundaries of what is possible, AWS’s best-in-class infrastructure will serve as the backbone of our AI ambitions,” said AWS CEO Matt Garman. “The breadth and immediate availability of optimized compute shows why AWS is uniquely positioned to support OpenAI’s vast AI workloads.”
You may also like: Interview | Big tech companies are training AI on junk data: Intuition
OpenAI deal raises questions about AI bubble
The deal is part of a $1.4 trillion investment to expand OpenAI’s infrastructure capacity. CEO Sam Altman said the company plans to develop 30 gigawatts of power capacity. For comparison, one nuclear power plant has a capacity of about 1 gigawatt, but 30 gigawatts can power 26 million homes for a year.
OpenAI’s previous trading has raised questions about the possibility of an AI bubble. Critics have warned that AI companies’ rising valuations and capital spending are not commensurate with the demand and profitability of AI.
read more: Is SLM the future of AI? Nvidia researchers think so. The reason is as follows
