Are underground nuclear reactors the future of energy

Deep Fission just started drilling in Kansas for an underground nuclear reactor test. This could flip nuclear power on its head.

Instead of building massive plants above ground, the company wants to put small nuclear reactors about a mile underground. If it works, the same drilling rigs used in the oil industry could be used to build nuclear power stations.

The California startup plans to lower small modular pressurized water reactors deep underground where the surrounding rock would serve as natural radiation shielding and containment.

The first phase involves three wells for site analysis and engineering validation. The first test well will reach roughly 6,000 feet and measure about eight inches in diameter. Engineers will use it to gather geological and safety data needed for licensing and final reactor design.

Chief executive Liz Muller says the project marks the shift from concept to physical hardware. Deep Fission is aiming to have its first reactor reach criticality by July 4. The company has already secured $80 million in new funding and a fuel supply agreement with Urenco USA.

The disruption behind the news: This startup could make nuclear power behave like oil drilling.

Nuclear power plants are massive construction projects that often take 10 years or more to complete and cost billions.

Today’s nuclear plants are large concrete structures built above ground. They require complex containment buildings, large security zones, and years of regulatory construction oversight. Costs often spiral because every reactor becomes a one-off megaproject.

Deep Fission is approaching the problem from a completely different direction. Instead of building nuclear plants like skyscrapers, it wants to install them the way oil companies drill wells.

Oil and gas companies routinely drill wells deeper than 20,000 feet. The rigs, crews, and supply chains already exist. If reactors can be lowered into drilled shafts, nuclear construction starts to look more like industrial drilling and less like large-scale civil engineering.

A modern nuclear plant can cost $6–10 billion partly because all the risk sits inside one large project that has to work perfectly the first time. A drilling-style approach breaks that risk across hundreds of smaller wells.

In oil, a single horizontal well might cost $8–12 million and be drilled in a matter of weeks. If borehole reactors follow a similar learning curve where each additional unit becomes cheaper as crews repeat the process, nuclear energy could start behaving more like a manufacturing system than a giant infrastructure project.

That’s where the economics become interesting. A one-mile borehole uses the surrounding rock as radiation shielding and containment. That removes some of the most expensive parts of a nuclear plant. There’s less concrete, less land, less visible equipment, and fewer structures above ground.

The design also dramatically reduces the land footprint. A traditional nuclear plant may require thousands of acres. A borehole reactor site could operate on a much smaller area.

Deep Fission says it already has customer interest totaling 12.5 gigawatts of future power capacity. That’s roughly equivalent to the output of more than 10 large nuclear reactors on today’s grid.

But the bigger disruption is industrial. If this model works, oilfield services companies could become part of the nuclear infrastructure supply chain. The same drilling industry that builds energy wells could start installing nuclear reactors. That possibility is gaining attention as tech leaders meet at the White House to address data center power costs, with demand for reliable electricity becoming a major strategic issue.

What to watch next

The Kansas borehole test is the first major proof point.

Regulators and engineers will focus on containment, heat transfer, and how maintenance would work underground.

If those hurdles are cleared, deployment could accelerate quickly.

The biggest open question is operational access. Traditional reactors allow engineers to perform direct maintenance. A reactor sitting a mile underground would require new engineering systems for servicing or retrieval.

The second issue is licensing speed. The Nuclear Regulatory Commission built its rules around above-ground reactors. Approving a buried reactor design will test how flexible nuclear regulation really is.

Then comes scale. If Deep Fission proves the concept, the limiting factor may not be reactor design. It could be drilling capacity.

The U.S. already runs thousands of drilling rigs across the country. If even a small slice of that industry shifted toward nuclear deployment, it could mean dozens of new reactors getting installed every year.

Energy markets usually move along cost curves. Solar took off once panels became mass-manufactured products. Oil production scaled because drilling turned into repeatable infrastructure.

If nuclear can plug into that same drilling economy, its costs could finally start coming down.

That’s why this Kansas well matters. It’s a test of whether an underground nuclear reactor can move beyond the slow, one-off megaproject model and become something the energy industry installs again and again.

P.S: A similar push toward unconventional infrastructure is showing up elsewhere too, including a startup planning underwater data centers powered by wind turbines.