Partnerships with Google and Amazon could help, but getting new reactors up and running still requires more investment — and time.
A conventional nuclear power plant in Mochovce, Slovakia. Small modular reactors aim to be cheaper, safer, and faster to build than such traditional plants. Credit: Janos Kummer/Getty
In October 2023, tech giants Google and Amazon separately announced partnerships supporting ‘advanced’ nuclear energy as a means to achieve their carbon-neutral goals.
Google announced it would purchase electricity from a nuclear reactor developed by Kairos Power (based in Alameda, California, USA). Meanwhile, Amazon is investing approximately US$500 million in X-Energy Reactor Company in Rockville, Maryland, and has agreed to purchase electricity from its designed reactors to be built in Washington state.
These moves are part of a broader green picture as tech companies begin to address the escalating energy demands of artificial intelligence (AI) data centres and computing clusters. Previously, Microsoft also announced it would purchase power from a utility company that plans to restart a decommissioned 835-megawatt reactor at the Three Mile Island nuclear plant in Pennsylvania.
The partnerships with Google and Amazon involve start-up companies leading the design of ‘small modular reactors’ (SMRs). These reactors are expected to be assembled from prefabricated components and are anticipated to be smaller, cheaper, safer, and faster to build than reactors in traditional nuclear power plants. The designs pursued by X-energy, Kairos, and several other companies (some funded by agencies like the US Department of Energy and the European Commission) differ significantly from those of established energy companies, but they are still a long way from becoming reality.
Nature spoke with nuclear energy research experts about the significance of these large tech company investments and their potential impact.
Can these partnerships drive innovation in the nuclear industry?
Building nuclear power plants — a process often plagued by complex licensing procedures, construction delays, and cost overruns — already carries high financial risks, and betting on unproven technology is even more uncertain. However, partnerships with Google and Amazon can provide a “huge” boost to Kairos and X-energy, says Jacopo Buongiorno, director of the Advanced Nuclear Systems Center at MIT, USA, and a nuclear engineer. He notes: “A lot of value is in the announced trust, and of course, it comes with money.” This announcement could help these companies attract more funding, he says, allowing them to cross the “valley of death for innovation” that often prevents novel ideas from reaching commercial success.
However, the details of these agreements are not clear, and the support provided by Amazon and Google might be “a drop in the bucket” compared to the billions of dollars these start-ups need, says Edwin Lyman, a physicist and director of nuclear safety at the Union of Concerned Scientists in Washington, D.C., USA. He says, “The PR machine is running on overdrive, but private capital doesn’t appear ready to take on that kind of risk.”
Allison Macfarlane, director of the School of Public Policy and Global Affairs at the University of British Columbia, Canada, and former chair of the US Nuclear Regulatory Commission, raises another question posed by the rapid advancements in computer science: “Will AI really be this energy-intensive in 15 years?”
How do small modular reactors work?
Several start-ups — and established players including Toshiba and Rolls-Royce — are developing small reactors, each claiming unique originality and advantages. Most companies are pursuing designs different from existing power generation facilities.
Almost all types of nuclear reactors generate energy from the fission of uranium atoms. The nucleus of the unstable isotope uranium-235 splits when struck by a neutron, releasing more neutrons that then strike other nuclei, triggering a chain reaction. Nuclear reactions release energy in the form of heat, which traditional nuclear power plants absorb by pumping cold water into the reactor core, producing pressurized steam to drive turbines and generate electricity.
X-energy’s design uses helium instead of water, while Kairos plans to use molten salt. Both abandon traditional nuclear fuel rods in favour of thousands of spherical fuel pebbles. Fuel pebbles are continuously added from the top of the reactor, and spent ones are removed from the bottom, somewhat like a gumball machine.
Do small modular designs have safety advantages?
“In theory, very small reactors can have a high degree of passive safety,” says Lyman. Compared to the reactor cores that melted down in the Fukushima Daiichi nuclear power plant accident after the 2011 Japanese tsunami, small reactors have less residual heat and radioactivity when shut down.
These companies also claim that the proposed pebble-bed reactors are inherently safer because they do not require pressurization, and their design allows for coolant circulation without pumps (the three reactors at Fukushima lost control due to pump failures caused by power outages).
However, Lyman argues that relying solely on potentially unpredictable passive cooling without active cooling options as a backup is risky. As reactor volume decreases, their efficiency also declines. Another start-up, Oregon-based NuScale Power, initially designed a small modular reactor (certified by the US Nuclear Regulatory Commission) with a planned capacity of 50 megawatts, but later shifted to a larger 77-megawatt design. Lyman points out that the pursuit of economic benefits “undermines the credibility of passive safety.”
Do small modular reactors pose additional risks?
In some cases, small modular reactors “may actually push nuclear energy in a more dangerous direction,” says Lyman, “advanced doesn’t necessarily mean better.”
Lyman specifically notes that the pebble-bed reactors designed by X-energy and Kairos require high-assay low-enriched uranium (HALEU) containing 10-20% uranium-235 — in contrast, most existing reactors (including NuScale’s) only require a 5% enrichment level. While HALEU is still classified as low-enriched fuel (relative to highly enriched uranium used for nuclear weapons), Lyman states this classification is misleading. In June 2024, he and his collaborators — including physicist Richard Garwin, a designer of the first hydrogen bomb — wrote in Science that only a few hundred kilograms of HALEU could be used to build a nuclear bomb without further enrichment [1].
Amazon data centre in Ashburn, Virginia, USA. Tech companies are investing in nuclear power to meet the growing energy demands of these facilities. (Image source: Nathan Howard/Getty)
According to Macfarlane and collaborators’ 2022 study [2], small reactors may also produce more nuclear waste and have lower nuclear fuel utilization efficiency. In full-sized reactors, most neutrons generated by uranium fission pass through a large amount of nuclear fuel, making it highly probable they will strike another nucleus rather than the reactor vessel wall or dissipate into surrounding structures. Macfarlane states: “When you shrink the reactor, you have less fuel, so you have more neutron leakage.” These escaping neutrons can be absorbed by other nuclei and become radioactive.
NuScale explained that this study was partly based on the company’s now-abandoned 50-megawatt reactor design, but Macfarlane and others believe this issue may apply to most small reactors.
Will small reactors be cheaper to build?
The ability to manufacture components on an assembly line can significantly reduce reactor construction costs. But Buongiorno points out that there are also inherent economies of scale in building larger reactors. He says, “Don’t blindly believe” when you hear that small reactors can produce cheaper energy; nuclear energy indeed has many advantages, but “it’s not cheap” — and that’s unlikely to change much.
However, Buongiorno adds that once the technology is proven and mature, individual small reactors should be cheaper and faster to build than traditional large reactors. This could make them more attractive to investors and accelerate their adoption.
Meanwhile, Lyman and others worry that the hype around small modular reactor technology — and the pursuit of cost reduction — could lower safety standards. For example, some companies claim their reactors are safe enough not to require reinforced concrete containment structures.
Can all these efforts help address climate change?
“We shouldn’t shut down existing nuclear power plants. We desperately need them, and we desperately need to move away from fossil fuels,” says Macfarlane. Even some staunch opponents of nuclear power have had to concede this point.
However, whether rapid emissions reduction must rely on building new reactors remains controversial. Macfarlane points out that solar panels and wind turbines can be deployed faster. Other assessments, including a report from the International Energy Agency, indicate that in many parts of the world, relying solely on intermittent solar and wind power might be too expensive, even with large-scale battery storage — and always-available sources like nuclear energy will still play an important role in future energy supply [3].