Developers of small nuclear reactors and advanced battery technologies are targeting data centers to meet the increasing electricity demands of AI applications. Clayton Scott, the chief commercial officer of NuScale Power, sees a huge business opportunity in providing carbon-free power solutions to data centers. A Goldman Sachs report predicts a 160% rise in data center power needs due to AI applications, creating an energy crisis as the existing grid capacity is strained.
NuScale Power is working on small modular nuclear reactors that could generate 77 carbon-free megawatts continuously, allowing data centers to operate independently of the grid. The company is partnering with data center developer Standard Power to supply 24 SMRs capable of generating nearly 2 gigawatts of electricity. However, the company is still navigating regulatory processes and commercial deployments are not expected until later this decade.
Other nuclear and battery startups are also focusing on providing energy solutions to AI data centers. Companies like Microsoft are exploring the use of small modular reactors to power their data centers, while startups like Oklo and Helion are developing alternative nuclear power technologies. Microreactors, such as those developed by Radiant Industries, offer a flexible and quick power solution for data centers with varying power needs.
In addition to nuclear power, companies like Natron Energy are developing sodium-ion batteries that can help stabilize power surges from AI servers in data centers. These batteries are designed to handle rapid power demands and prevent grid instability. Despite being more expensive than lithium-ion batteries at the moment, sodium batteries offer advantages in terms of recharging speed and safety, making them an attractive option for AI data centers.
In the quest for sustainable energy solutions, some companies are exploring geothermal systems and green hydrogen fuel cells as alternatives to nuclear power and traditional battery technologies. Fervo, a startup working on geothermal energy systems, is being tested by Google for potential use in data centers. While these technologies may not be as fully developed as nuclear power or batteries, they offer promising alternatives for meeting the energy needs of AI data centers in the future.
Overall, the convergence of AI technology and data center power demands is driving innovation in nuclear power, advanced battery technologies, and alternative energy sources. As data centers continue to expand and consume more energy, the need for sustainable and reliable power solutions becomes increasingly critical. Developers and startups are rising to the challenge by creating innovative products that could shape the future of energy consumption in the age of AI.
