The world’s largest technology companies are pouring billions of dollars into next-generation nuclear projects to power their rapidly expanding fleets of artificial intelligence data centers. Microsoft, Amazon, Google and Meta are all backing small modular reactor (SMR) startups as electricity demand from AI workloads outstrips traditional grid capacity.
Why AI Needs Nuclear
Training and running frontier AI models consumes extraordinary amounts of electricity. A single hyperscale data center can draw more power than a medium-sized city, and AI-optimised facilities running on Nvidia GPUs push those demands even higher. Existing grids — particularly in the United States and Europe — cannot scale fast enough to meet the load.
The SMR Bet
Small modular reactors promise compact, factory-built nuclear units that can be deployed alongside data centers. Microsoft signed a deal to restart the Three Mile Island Unit 1 reactor. Amazon Web Services committed billions to SMR projects on its existing data-center footprints. Google and Meta have signed long-term power purchase agreements with nuclear startups including Kairos Power and X-Energy.
A Decade-Long Gamble
Analysts say nuclear may be the only energy source that can scale fast enough to meet AI’s appetite without blowing through climate commitments. The bet is long-term: SMRs will not come online at meaningful scale until the late 2020s and early 2030s. But the capital now flowing into the sector is transforming an industry that had stagnated for decades into one of tech’s most urgent strategic priorities.
Leave a Reply