4 min readNew DelhiOct 16, 2024 08:12 AM IST
First published on: Oct 15, 2024 at 03:05 PM IST
Nearly a decade-and-a-half after pioneering the first corporate purchase agreements for renewable electricity over a decade ago, technology major Google has yet another global landmark: signing the first ever corporate agreement to purchase nuclear energy from multiple small modular reactors (SMRs) to be developed by Alameda, California-based Kairos Power.
This new deal aims to use small nuclear reactors to generate the huge amounts of energy needed to power Google’s artificial intelligence (AI) data centres and the tech major has said the pact with Kairos Power will see it “start using the first reactor this decade” and bring more online by 2035. Both companies did not give any details about how much the deal is worth or where the plants will be constructed.
In 2010, Google had entered its first power purchase agreement with a 20-year pact with a 114-megawatt-capacity wind farm in Iowa, marking a global first at that time.
The Google-Kairos deal
Technology firms such as Google are increasingly turning to nuclear sources of energy to supply the power needed to run vast data centres that drive AI. While renewables have been the first choice for companies, including for Google, there are hurdles that green power poses. Nuclear offers a solution, especially since the grid needs new electricity sources to support AI applications and nuclear solutions offer a clean, round-the-clock power source that can tide over the limitations of renewables – not generating power when the sun’s not shining or the wind is not blowing, and there not being enough storage options at this point in time to tide over these shortfalls.
Big boost for SMRs
SMRs – small reactors with a capacity of 30MWe to 300 MWe per unit – are increasingly seen as important for nuclear energy to remain a commercially competitive option in the future, especially in the wake of surging power demand from technology companies given the massive incremental electricity requirement coming in from AI machine learning applications and data centres.
As of now, two SMR projects have reached the operational stage globally. One is an SMR named Akademik Lomonosov floating power unit in Russia that has two-modules of 35 MWe (megawatt electric) and started commercial operation in May 2020. The other is a demonstration SMR project called HTR-PM in China that was grid-connected in December 2021 and is reported to have started commercial operations in December 2023. Apart from Kairos, other global companies that are in the fray with SMR technology include New Jersey-based Holtec, Rolls-Royce SMR, NuScale’s VOYGR SMR, Westinghouse Electric’s AP300 SMR and GE-Hitachi’s BWRX-300.
India too is working to get into the manufacturing value chain of small reactors, both as a way of fulfilling its commitment to clean energy transition, and bundling SMRs as a technology-led foreign policy pitch.
Kairos’ SMR tech
Kairos Power’s technology uses a molten-salt cooling system, combined with a “ceramic, pebble-type fuel”, to efficiently transport heat to a steam turbine to generate power. This is a different technology as compared to the most commonly used light water model, where reactors use water as moderator and coolant. Kairos claims its SMR’s passively safe system “allows the reactor to operate at low pressure, enabling a simpler more affordable nuclear reactor design”. The company says it will complete “multiple successive hardware demonstrations” ahead of its first commercial plant to enable critical learnings and efficiency improvements that accelerate reactor deployments, as well as greater cost certainty for Google and other customers.
Earlier this year, Kairos Power broke ground on its Hermes non-powered demonstration reactor in Tennessee, the first American advanced reactor project to receive a construction permit from the US Nuclear Regulatory Commission.
Technology firms such as Google are increasingly turning to nuclear sources of energy to supply the power needed to run vast data centres that drive AI. (Reuters photo) 
Share
