Here’s an eye-popping new price tag for generative AI’s booming demand for energy and computing power for data centers: $300 billion. That’s how much ChatGPT maker OpenAI has reportedly committed to spend with Oracle over a five-year span to continue its run of early success.
The deal, which would be one of the largest contracts for cloud computing, was announced by both The Wall Street Journal and The New York Times on Wednesday.
The contract with Oracle would entail delivery of as much as 4.5 gigawatts of power capacity, which the Journal says is roughly equivalent to two Hoover Dams or the amount used by 4 million homes. The deal would go into effect in 2027, the reports said.
OpenAI didn’t respond to our request for comment. Oracle declined to comment.
According to research reports, the number of data centers in the US nearly doubled from 2021 to 2024, and the demand for more data centers will rise 9% yearly through 2030. Those facilities are expected to consume twice as much electricity by 2035 as they do today.
Don’t miss any of our unbiased tech content and lab-based reviews. Add CNET as a preferred Google source.
Previously, OpenAI relied exclusively on Microsoft Azure, but has begun to diversify its cloud portfolio.
Watch this: The Hidden Impact of the AI Data Center Boom
In January, OpenAI announced the formation of the Stargate Project, a new company that would invest $500 billion over four years to build AI infrastructure — that is, data centers — with partners including Oracle, Microsoft, Nvidia and Softbank. In July, OpenAI and Oracle announced an agreement to develop 4.5 gigawatts of data center capacity for Stargate, beyond the 10 gigawatts promised in the January announcement.
The Stargate facility is under construction at a sprawling site in Abilene, Texas.
ChatGPT is the world’s most used chatbot, amid intense competition with Google’s Gemini, Anthropic’s Claude and Perplexity’s namesake software.
The rapid growth of generative AI products and services has even led OpenAI CEO Sam Altman to worry that we’re in an AI bubble.
