Google Taps Kairos Power for Small Modular Nuclear Reactors to Fuel AI
The tech giant will purchase up to 500 MW of carbon-free energy to sustain its energy-intensive AI data centers.
Google has entered into a first-of-its-kind corporate agreement with Kairos Power to purchase electricity from a fleet of small modular reactors (SMRs). The deal secures a stable, carbon-free energy supply to meet the massive power requirements of Google's artificial intelligence data centers.
Under the terms of the agreement, the partnership aims to add up to 500 MW of new 24/7 carbon-free power to electricity grids across the United States. The deployment schedule is phased, with the first reactor intended to be online by 2030 and additional units continuing to roll out through 2035.
The AI Energy Crunch
This move comes as the rapid expansion of AI technologies drives a surge in electricity demand for data centers. While tech companies have historically relied on variable renewable sources like wind and solar, those sources cannot provide the constant, round-the-clock power required for high-intensity computing. To maintain sustainability goals while scaling AI, companies are now looking toward advanced nuclear options.
Kairos Power differentiates its technology from traditional nuclear plants by utilizing a molten-salt cooling system and ceramic pebble-type fuel. According to the company, this design allows reactors to operate at low pressure, which enhances inherent safety compared to conventional nuclear infrastructure.
Industry Implications
This agreement marks a significant shift in corporate energy procurement. By moving toward advanced nuclear power, Google is attempting to solve the intermittency problem associated with renewables. Furthermore, by creating a committed "orderbook" for these reactors, Google is effectively helping to commercialize SMR technology. This corporate backing could lower overall costs and accelerate the broader deployment of advanced nuclear energy across the U.S. power grid.
Michael Terrell, Senior Director of Energy and Climate at Google, noted that the grid requires new electricity sources to support AI technologies that are driving national competitiveness, economic growth, and major scientific advances.
The Path to 2030
As the industry watches this rollout, the primary focus remains on the 2030 deadline for the first operational reactor. The success of this initiative will likely serve as a bellwether for other hyperscalers considering nuclear energy to fuel the AI era. While the agreement sets a clear timeline through 2035, the actual speed of deployment will depend on regulatory approvals and the successful scaling of Kairos Power's modular designs.