Energy Stocks Crash, Highlighting Uncertainty over AI Computing Needs

02/27/2025 - New York, NY

by Miranda Morrow

Wall Street's high-voltage bet on an AI-powered electricity boom short-circuited this week after China’s DeepSeek stunned markets with claims that its AI operations require only a fraction of the energy consumed by American rivals. The announcement triggered a broad selloff in energy sector stocks, as investors began reassessing the validity of aggressive utility load growth forecasts once used to justify hundreds of billions in new power infrastructure.

The crash was catalyzed by a cautionary op-ed from the Electric Power Supply Association's president, who flagged the irrational exuberance inflating electricity demand projections. The eGeneration Foundation has echoed caution, warning that the accuracy of China's claims have yet to be verified.

With DeepSeek touting a technological leap in AI energy efficiency—and the national security implications of misjudging AI’s infrastructure footprint—regulators and policymakers rushing to make new policy without verifying China's claims, leaves us to ask: how much electricity will AI really need?

While some experts urge restraint, others warn that underestimating demand could just as easily leave the grid unprepared. Even OpenAI’s ChatGPT weighed in, remarking, “Future AI energy needs will likely grow but at a more measured pace than early predictions suggested, driven by efficiency gains, specialized hardware, and smarter resource management.”

As the dust settles, the lesson is clear: energy investment should be informed by rigorous analysis, not speculative hype. For now, the future of AI—and the grid—is still being written.