Nuclear power is in a hole. To get out, this start-up kept digging

Nuclear power is in a hole. To get out, this start-up kept digging

They’re betting that they can take the cost out of nuclear power—and solve its waste problem — by putting small reactors a mile underground

By Mark Harris

Let the best of Anthropocene come to you.

In the spring of 2024, a giant robot began attacking an underground nuclear reactor near Stockholm. The victim was the Ågesta Power Plant, Sweden’s first commercial pressurized water reactor, and the motive, as so often, was money.

The reactor was built in 1963 to power Ågesta, a ritzy suburb of soaring trees and sparkling lakes. Its builders thought they were being smart by excavating an air-tight cavern half as tall as the Statue of Liberty, then squirreling away the reactor beneath 50 feet of solid bedrock.

Not only would the rock minimize the risk from accidents, they thought, it would also protect Ågesta from attack, a constant worry in the chill of the Cold War.

But the real threat came from something far more mundane. Excavating, constructing and operating an underground reactor was ruinously expensive, and by the mid-1970s, economics forced Ågesta to close.

Engineers in the US and elsewhere returned to the idea of underground nuclear power stations over the decades but there was always that nagging problem of expense. Digging a cavern large and deep enough to add real safety would blow gaping holes in already colossal budgets.

Now one startup thinks it has cracked the problem.

Deep Fission, based in Berkeley, California, offers the tantalizing promise of ultra-safe underground nuclear reactors that produce power as cheaply as solar farms. What’s more, its version of a nuclear renaissance uses the very tools developed to extract fossil fuels, to replace them instead.

The problem with earlier reactors like Ågesta, thinks Elizabeth Muller, Deep Fission’s CEO, is that they couldn’t dig anywhere near deep enough. And thanks to fracking, that’s all changed.

“We’re able to draw on work that’s been done by the oil and gas industry to drill holes a mile, two miles, four miles deep,” says Muller. “If you’re in a borehole mile underground, you have the exact conditions to create a pressurized water reactor without the construction costs, which are 80% of the cost of current nuclear plants.”

And because most of it is underground, this is a nuclear power plant that looks nothing like any nuclear power plant you’ve seen before.

Nuclear power is in a hole To get out this start up kept digging

Deep Fission’s micro-reactor is an extraordinary shape: picture something as narrow as a garbage can and as tall as a house.

Fission 101

To get a sense of the vision here, a quick lesson in nuclear power is a good place to start. Large amounts of energy are stored in the bonds between protons and neutrons in every atom in the universe. In a few elements like uranium, those bonds can be broken relatively easily, such as by a stray neutron. That causes the uranium to decay into a lighter element, releasing heat, radiation and more neutrons along the way. Under the right conditions, a chain reaction continues until all the uranium has decayed.

The job of a nuclear reactor is to keep that chain reaction ticking over to provide a constant stream of heat, without either fading out or catastrophically melting down. In pressurized water reactors (PWRs) like Ågesta, which account for about two thirds of all reactors worldwide, the heat is absorbed in water pressurized to about 150 atmospheres. This keeps it in a liquid state even at high temperatures.

The superheated water, now slightly radioactive, flows out of the reactor and is used to boil clean water in another part of the plant. That steam then drives turbines to generate zero carbon electricity.

Deep Fission’s idea is to put just the dirty, dangerous parts of a PWR—the uranium fuel and most of the radioactive water—deep underground, leaving just the clean steam and electricity generation on the surface.

In a mile-deep hole, the weight of water alone naturally creates about 150 atmospheres of pressure at the reactor. That should allow Deep Fission to use some already proven and licensed PWR technologies, without having to maintain the high pressure artificially.

“Uranium fuel is not that expensive,” says Muller. “What makes nuclear power so expensive is all the concrete and steel that goes into building the pressure, and then the containment in case anything goes wrong.”

Putting the active reactor so far underground means that safety should be much less of an issue. Deep Fission calculates that there would be ten billion tons of rock shielding between the reactor and the surface. That means it would be much safer from tornadoes, tsunamis, airplane crashes or deliberate attack, says Muller. “Even for earthquakes, a mile underground, you could have some shaking, but you’re not going to have the same potential for things toppling over.”

Another cost-saving measure is to use the same standard fuel assemblies as existing PWRs. But where a current PWR might use 150 or 200 assemblies to produce a gigawatt of electricity or more, Deep Fission’s micro-reactor will have just four, generating 15 megawatts. That’s a limitation of the borehole size the company hopes to use, a mere 30 inches across. The reactor itself would be slightly skinnier, and about 30 feet tall.

“Most oil and gas wells are much narrower, in the 4 to 8 inches range, “ says Muller. “But 30 inches is well within the ‘we’ve done this before’ capability. We’ve gotten solid quotes on how long it’s going to take to drill these holes and we’re talking about a couple of weeks.”

Muller envisages drilling multiple holes for multiple reactors at a single location, perhaps as many as a hundred over three acres. The steam from these would feed into generating plants, say one for every five reactors, to produce electricity. Instead of a traditional nuclear plant with monolithic buildings and massive cooling towers, a Deep Fission facility would look more like a modest chemical plant or oil refinery—and might not need to be isolated miles from other buildings.

In fact, Deep Fission recently signed a commitment to develop two gigawatts of nuclear energy to power nearby, purpose-built, AI-ready data centers. It hopes the first reactors could be operational in 2029, at a location still to be finalized.

That’s an ambitious goal for any new reactor, let alone so novel an idea. “It’s too early to tell if this concept stands up technologically or commercially,” says Adam Stein, director of the Nuclear Energy Innovation program at the broadly pro-nuclear, non-profit Breakthrough Institute. “They avoid some costs for building a very robust containment building by putting it deep underground but increase some other costs, such as drilling that deep bore hole and the piping and cabling to get the reactor down there.”

Stein points out that the design has some things in common with advanced geothermal power generation, which aims to use naturally hot subterranean rocks to heat water and generate electricity at the surface, without any reactor at all. “They can learn from each other and take advantage of not only advances in that area, but the fact there’s already an infrastructure in the US and elsewhere to do this type of drilling,” he says.

Muller herself quips: “I sometimes refer to Deep Fission as ‘assisted geothermal’. We’re just creating the heat at the bottom rather than relying on it from the rock.” (In case you’re wondering why you need the reactor at all, a fission system would produce much hotter water and thus more electricity than natural geothermal activity).

In a mile-deep hole, the weight of water naturally creates 150 atmospheres of pressure at the reactor, avoiding the need for a large, expensive containment vessel.

But is deep enough?

If only it were that simple. Ten years ago, Kellen Giraud, now a reactor projects engineer at Idaho National Laboratory, studied the feasibility of underground nuclear power. He says Deep Fission’s idea is “definitely doable but does present design challenges.”

Up until now, new reactors have had standardized designs that enable them to be built on almost any plot of remote land. “But with an underground nuclear plant, you have to consider the different kinds of rock in order to modify that standardized design for different areas,” Giraud explains.

That could be a problem when it comes to getting the sign-off from the US Nuclear Regulatory Commission (NRC), which has “historically been pretty inflexible,” says Giraud.

In July 2024, Deep Fission presented the initial conceptual design for its reactor to the NRC. The regulators’ questions, captured in a document in August, show that many were not entirely convinced.

Several raised the possibility of corrosion from salt in groundwater, the fragility of a mile-long system piping water to and from the surface, and the difficulty of testing and classifying each borehole before use. Others noted that there is little science about how heat, neutrons and potentially radioactive water from the reactors might spread through the surrounding rock.

And then there’s the fate of the left-over nuclear waste. Like above-ground PWRs, Deep Fission’s micro-reactors would need refueling about every two years, and leave spent fuel that will be dangerously radioactive for thousands of years. In 1987, the US Congress selected Yucca Mountain in Nevada as the long-term, geological repository for much of America’s nuclear waste. But sustained opposition from local politicians has left that project stalled, leaving waste from today’s reactors stored semi-permanently on sites around the country.

“Right now we’re in a deadlock until Congress moves forward with some solution,” says Stein. “It is likely that we will, if not this year, then in the next few years, see some progress in a direction. It’s just hard to tell what direction that will be.”

Deep Fission’s preferred option would be to simply lower its reactors’ waste deeper into the very same boreholes. That would have the advantage of never transporting waste by road, rail or air, with all the attendant risks of spillage, accidents, or hijacking.

“And once we’ve done it for Deep Fission, I think it also paves the way for other people to understand how great boreholes are for nuclear waste disposal,” says Muller. “Hopefully, this will unlock the solving of the nuclear waste problem.”

(Perhaps not coincidentally, it could also unlock some income for Deep Fission’s sister company, Deep Isolation, which has been developing deep boreholes for spent fuel storage for the better part of a decade).

The NRC had some questions about this innovation, too. Regulators were worried that Deep Fission had “an overly simplified view that ‘deep’ is all that is needed for safety,” and warned the company against making decisions about disposal after the reactor was operating. Allowing permanent storage in boreholes would be “a major effort” involving new rulemaking and exemptions.

But that could still be easier than trying to revive Yucca Mountain, argues Stein. He points out that the public gets benefits from a nuclear power station in terms of jobs, clean electricity, and less pollution. A community that just stores radioactive waste gets none of these.

Deep Fission wants to store waste underneath the reactors, in the same boreholes. That would avoid the risks of accidents or hijacking that come with transportation.

“In part, this is why there has been more resistance to a spent fuel facility than to an actual power plant,” he says. “But if you co-locate generation and storage, you can continue to have public benefits in that same facility.”

Muller believes that Deep Fission is right on track to kickstart a new generation of carbon-free electricity. “The unsolved waste problem and the cost are the two big things holding back the future of nuclear power,” she says. “It’s sort of a crazy idea, but you can take the cost out of nuclear power by putting it a mile underground, and have really, really great containment.”

The company is now working to finalize its technical designs, essentially shrinking a reactor the size of Ågesta to something that can be made in a factory and shipped out on a truck.

Deep Fission’s timing might just be perfect. Giraud points out that deep geological storage of nuclear waste is no longer just theoretical. The world’s first such facility is nearing completion in Finland, and construction on a second has started in Sweden.

Some of the first material deposited there could be the most dangerous waste from Ågesta, including the reactor’s 10-foot long silver control rods. These were snipped into segments by the decommissioning robot using a pair of giant scissors.

“Sadly, we will have quite a lot of radioactive silver that we cannot make rings out of,” lead engineer Nicklas Tjernlund said last year.

If that goes smoothly, Deep Fission will be hoping that burying nuclear systems far underground soon becomes something that happens at the start of their useful life, rather than at the end.

________________