Google agrees to multi-reactor power deal with nuclear startup Kairos

The power needs of Google’s insatiable data centers might eventually be met by small modular nuclear reactors — emphasis on might and eventually.

Search giant Google and startup Kairos Power have signed what looks to be the first corporate agreement to develop a fleet of small modular reactor (SMR) projects in the United States. The plan is to bring Kairos’ first SMR online by 2030, followed by additional reactor construction through 2035. There are currently no SMRs online in the U.S., and only one design — not Kairos’ — has received full regulatory approval to date.

In total, this deal will provide ​“500 megawatts of new 24/7 carbon-free energy to the U.S. electricity grids,” said Michael Terrell, Google’s senior director of energy and climate at a press briefing on Monday — but the exact terms of the power purchase, such as pricing, location, or number of the reactors, were left unanswered.

Still, a customer like Google committing to an order book for a series of reactors is a big deal.

The U.S. Department of Energy views this sort of multi-reactor agreement as key to bringing down the costs of nuclear and deploying more reactors, though it’s keener on large reactors than small ones these days.

While widely deployed in the nuclear navy, SMRs in service of civilian power is a relatively new idea. For around a decade, SMR champions have cited this approach as a way for the nuclear energy industry to reduce its prohibitive costs and accelerate its sluggish timelines.

In theory, SMRs have the potential to lower the costs of nuclear energy through the power of mass production — 50- to 350-megawatt reactors can be manufactured in factories and transported to various locations instead of being expensively constructed on-site.

This theory hasn’t had much of a chance to be proven out.

While dozens of companies have aspired to commercialize SMRs, only the startup NuScale has managed to win design approval from the U.S. Nuclear Regulatory Commission — but it hasn’t yet connected a reactor to the grid.

The NRC has issued Kairos a construction permit to build a demonstration reactor, a first for an advanced SMR design like the startup is pursuing: a 35 megawatt (thermal) reactor using a molten fluoride salt coolant and a higher-concentration uranium fuel recipe. Kairos aims for the project to be online by 2027.

The U.S. has 93 reactors in operation today, all of which exclusively use water for cooling, average about 1,000 megawatts in size, and rely on a low-concentration fuel.

The company’s first commercial deployment would be 50 megawatts. From there, Kairos would scale up to 75 megawatts per reactor, a plan that’s eerily similar to NuScale’s roadmap.

The U.S. nuclear industry has been stagnant in recent decades, bogged down by cost and timeline overruns. But there are now some rays of hope. Chief among them is the trend of firms exploring or attempting to reopen closed reactors, upgrade plant capacity, and extend operation of existing plants to 60 or 80 years.

The U.S. government is squarely behind this renewed push around nuclear energy.

The DOE Loan Programs Office has $64.89 billion in applications for nuclear projects under consideration as part of a new IRA program that backs efforts intended to repower or retool idled energy infrastructure. The Inflation Reduction Act included a 30 percent investment tax credit that can be expanded to 50 percent with adders for using domestic content and being located in an​“energy community” that has hosted energy infrastructure in the past.

There is also specific government support for advanced nuclear technology development. Billions in grants and incentives are aimed at building advanced reactors and creating the sort of highly enriched uranium fuel reactor that firms like Kairos need. The startup is also receiving up to $303 million to build its demonstration reactor through the DOE’s Advanced Reactor Demonstration Program.

Despite that insurance, Kairos has an enormous set of challenges ahead: It’s trying to develop, build, and permit a commercial nuclear reactor by 2030 using a different architecture, coolant, and fuel in an industry not known for its efficiency.

But as Google’s Terrell said, ​“To really get grids carbon free around the world, it’ll take more than just wind, solar, and lithium-ion storage. You’re going to need this next set of advanced technologies that are firm and more dispatchable and can help fill in some of the variability that we see with wind, solar, and storage.”