OpenAI's AI Data Centers to Consume Power Equivalent to Major U.S. Cities

OpenAI has announced plans to build AI data centers that will consume up to 17 gigawatts of power, equivalent to the combined electricity usage of New York City and San Diego. This development has raised concerns among experts about the environmental impact and energy demands of AI infrastructure.

Andrew Chien, a computer science professor at the University of Chicago, highlighted the significant shift in computing's energy consumption. He noted that computing, once a minor part of the economy's power use, is now becoming a substantial share. Chien expressed both excitement and concern, stating that computing could account for 10% to 12% of the world's power by 2030.

OpenAI's collaboration with Nvidia aims to build AI data centers consuming up to 10 gigawatts of power, with additional projects totaling 17 gigawatts already in progress. This is comparable to the peak power demand of Switzerland and Portugal combined. Fengqi You, an energy-systems engineering professor at Cornell University, emphasized the magnitude of this consumption.

Sam Altman, CEO of OpenAI, has advocated for nuclear energy as a solution to meet AI's growing power needs. He has invested in both fission and fusion startups, believing that reactors can provide the steady, concentrated output required for AI. However, Chien pointed out that the amount of nuclear power that could be added to the grid before 2030 is less than a gigawatt, suggesting that wind, solar, natural gas, and new storage technologies will dominate in the near term.

The environmental implications of such massive energy consumption are significant. Chien noted that companies may struggle to meet their clean and net-zero promises in the face of AI growth. You warned that data centers could stress local ecosystems by consuming resources like water and disrupting biodiversity.

The financial investment in these projects is substantial. Each OpenAI site is valued at approximately $50 billion, totaling $850 billion in planned spending. Nvidia has pledged up to $100 billion to support the expansion, providing millions of its new Vera Rubin GPUs.