Big Tech Bets on SMRs and Geothermal Power for AI Energy

As artificial intelligence models grow more complex, the data centers powering them require an unprecedented amount of electricity. Because traditional renewable sources like wind and solar are intermittent, major technology companies are aggressively seeking 24/7 carbon-free energy to maintain their climate pledges while expanding their computing capacity. To solve this impending energy crisis, tech giants are investing billions of dollars into two emerging clean baseload technologies: Small Modular Reactors (SMRs) and advanced geothermal power.

The core challenge for AI infrastructure is the need for uninterrupted power. Training large language models and running continuous AI queries causes massive, constant power draw. While solar and wind energy have formed the backbone of tech companies' green initiatives over the past decade, they cannot generate electricity when the sun isn’t shining or the wind isn’t blowing. Current battery storage technology is not yet capable of bridging these gaps at the scale required by hyperscale data centers. Consequently, companies like Google, Microsoft, and Amazon are looking at nuclear and geothermal energy to provide a steady, round-the-clock power supply.

Small Modular Reactors (SMRs) have become a primary focus for the tech sector. Unlike traditional nuclear plants, which take decades to build and require massive footprints, SMRs are designed to be built in factories and assembled on-site. This makes them faster to deploy, scalable, and easier to locate directly next to data centers.

Tech companies are actively funding the development of these reactors:

Google signed a landmark agreement with Kairos Power to purchase energy from a fleet of SMRs, aiming to bring the first reactor online by 2030 and scale up to 500 megawatts by 2035.
Amazon invested heavily in X-energy, a developer of small modular nuclear reactors, and signed agreements with utilities in Washington and Virginia to explore SMR projects capable of generating hundreds of megawatts of carbon-free energy.
Microsoft is actively hiring nuclear energy specialists to lead a global strategy for powering its cloud and AI data centers with SMRs and microreactors, while OpenAI CEO Sam Altman has personally backed the advanced fission startup Oklo.

Alongside nuclear power, technology companies are turning to Enhanced Geothermal Systems (EGS). Traditional geothermal energy relies on naturally occurring underground pockets of hot water and steam, which limits plants to very specific geographic locations. EGS, however, borrows advanced drilling techniques from the oil and gas industry—such as horizontal drilling and hydraulic fracturing—to tap into the Earth's heat almost anywhere. By pumping water into artificially created fractures in hot, deep rock, EGS creates a closed-loop steam system that drives turbines continuously.

The push for advanced geothermal is already yielding real-world results for AI data centers:

Google partnered with geothermal startup Fervo Energy to develop an enhanced geothermal power project in Nevada. This project is already operational and actively supplying carbon-free electricity to the local grid that serves Google’s data centers in the region.
Meta recently announced a partnership with Sage Geosystems to harness next-generation geothermal power. The deal aims to deliver up to 150 megawatts of new, carbon-free baseload energy to power Meta's expanding data center footprint before the end of the decade.

While both technologies hold immense promise, they face distinct hurdles. SMRs must navigate strict regulatory approval processes, supply chain bottlenecks for specialized nuclear fuel, and the reality that most commercial deployments will not be ready until at least 2030. Advanced geothermal, meanwhile, has a faster deployment timeline and faces fewer regulatory barriers, making it a critical bridge for the late 2020s, though it still requires significant capital to scale down drilling costs. By investing in both SMRs and geothermal simultaneously, Big Tech is hedging its bets, ensuring that the rapid evolution of artificial intelligence will not be bottlenecked by an outdated electrical grid.