Google has teamed up with Elementl Power to develop three advanced nuclear power projects in the US, totaling 1.8 gigawatts. The projects aim to support the growing energy needs of Google’s AI data centers. Google plans to expand its AI data centers in 2025 with a $75 billion budget.
The deal allows Google to buy electricity directly from the nuclear power plants at a pre-agreed price. The remaining electricity will be sold to the national grid for general use. This setup ensures Google has a stable and clean energy supply for its expanding AI data centers.
Elementl Power was founded in 2022 by Breakwater North and backed by Energy Impact Partners. The company’s team has nuclear industry experience, but this is their first advanced nuclear power project. Elementl Power hasn’t decided which small modular reactor (SMR) to use yet. They’re considering options from companies like Kairos Power, which has worked with Google before.
Kairos Power is developing a 50-megawatt nuclear power plant prototype. Their design is based on the Hermes test reactor, which was approved for construction in Tennessee in 2024. This could make it easier for Kairos Power’s SMR to get regulatory approval.
SMRs typically produce up to 300 megawatts of electricity. They’re attractive to AI companies because they can provide constant power and be installed near data centers. No SMR has been built outside China yet. In the US, NuScale was the first company to try, but they faced issues with rising costs in 2023.
Nuclear Power for AI Data Centers
The partnership between Google and Elementl Power highlights the growing energy needs of AI data centers. These centers require a lot of power to operate. Nuclear energy can provide a reliable and clean source of electricity. This is important for companies like Google that are expanding their AI capabilities.
The use of SMRs could be a key part of this strategy. They’re smaller and more flexible than traditional nuclear reactors. This makes them easier to build and deploy. Companies like Kairos Power are working to develop and deploy SMRs.
The success of these projects will depend on several factors. These include regulatory approval, cost, and public acceptance. But if successful, they could provide a new source of clean energy for AI data centers.
Source: