Deep Fission is betting that the safest place for a nuclear reactor is not behind thicker concrete walls but a mile beneath our feet. The California startup wants to sink compact power plants into deep boreholes, turning the surrounding rock into a natural shield while piping clean electricity back to the surface. If it works at scale, this approach could reframe how I think about nuclear power in an era of surging data center demand and tightening climate deadlines.
Instead of sprawling surface complexes, Deep Fission envisions a fleet of buried reactors that disappear into the landscape, serving energy-hungry customers like cloud campuses and utilities with minimal visual footprint. The company’s flagship concept, a small modular reactor called Gravity, is designed to operate far below ground level, where geology and distance provide a built-in containment system that traditional plants have struggled to match.
From surface plants to “Powering Humanity” a mile down
The core of Deep Fission’s pitch is simple: move nuclear power out of sight and into the deep subsurface, where the earth itself becomes part of the safety case. On its own site, the company frames its mission as Powering Humanity from a Mile Underground, arguing that Traditional and next-generation reactors at the surface have not fully solved concerns about cost, siting, and public acceptance. By drilling narrow boreholes roughly a mile deep and lowering small modular reactors (SMRs) into them, Deep Fission aims to sidestep many of the land use and construction headaches that plague conventional projects.
In this model, the reactor core and primary systems sit far below ground, while only a modest surface facility is visible, handling power conversion and grid connection. The company describes a closed-loop configuration in which heat from the buried unit is brought up to drive turbines, then returned underground, an arrangement it promotes in its own media materials as a way to keep radioactive components isolated. That framing is designed to appeal not just to engineers but to communities that have long been wary of living next to large nuclear plants, even as they demand more reliable low-carbon power.
Gravity, the deep borehole reactor at the center of the plan
At the heart of this strategy is Gravity, the name Deep Fission has given to its deep borehole reactor. The Company formally introduced Gravity after completing a go-to-market review, presenting it as a standardized unit that can be replicated across sites rather than a bespoke mega-project. In its announcement, Deep Fission emphasized that Gravity is designed around off-the-shelf components, existing supply chains, and readily available low-enriched uranium fuel, a deliberate choice to avoid the bottlenecks that have slowed other advanced nuclear concepts.
By standardizing Gravity as a modular product, Deep Fission is trying to make nuclear feel more like industrial equipment and less like one-off infrastructure. The company’s featured updates describe how Gravity is intended to operate a mile underground with a built-in containment system formed by the surrounding rock and engineered barriers. That combination of geological isolation and factory-style repeatability is central to the firm’s claim that it can deliver safer, faster, and cheaper nuclear projects than the surface plants that have defined the industry for decades.
How a mile-deep SMR actually works
Deep Fission’s technical concept starts with drilling a narrow but very deep borehole, similar in depth to some oil and gas wells but optimized for nuclear hardware instead of hydrocarbons. Into that shaft, the company plans to lower a compact SMR module that sits roughly one mile below the surface, where stable rock formations and hydrogeology can help contain any potential release. The reactor then circulates a working fluid in a closed loop to the surface, where turbines and generators convert the heat into electricity, a configuration the firm highlights in its own plans to sink reactors deep underground.
Because the nuclear core is buried, the visible footprint is limited to a relatively small surface plant that handles power conversion, control systems, and grid interconnection. Reporting on the design notes that Deep Fission wants to avoid the massive concrete containment domes and cooling towers associated with traditional reactors, instead relying on the depth of the borehole and engineered seals as a built-in containment system. Coverage of the concept explains that the company’s underground reactor concept envisions the unit being installed hundreds of metres underground via a borehole, with the surrounding geology forming part of the safety envelope.
Why California’s Deep Fission is chasing data centers first
Deep Fission is not shy about its first big target market: the data centers that underpin cloud computing, artificial intelligence, and streaming. The company is based in California, and reporting on its strategy notes that the California nuclear energy startup is advancing a reactor it plans to bury a mile underground for safety while eyeing power-hungry digital infrastructure as an anchor customer. One account of the firm’s trajectory describes how California-based Deep Fission is positioning its underground units as a way to deliver reliable, around-the-clock power directly to these facilities without overloading local grids.
That focus is echoed in technical coverage that describes how Deep Fission Inc plans to place SMRs in boreholes a mile deep and send the power back to the surface to feed large computing campuses. One analysis of the approach notes that Deep Fission Inc wants to use these underground reactors to power data centers and other high-demand customers, framing the technology as a way to pair nuclear reliability with the compact footprint and modularity that digital infrastructure developers expect. For an industry that has been scrambling for firm, low-carbon power in places like Northern Virginia and central Texas, a buried 15 megawatt unit that can sit near or even within a campus boundary is an attractive proposition.
Engineering safety into the rock itself
Safety is the central selling point of burying a reactor, and Deep Fission leans heavily on the idea that geology can do work that concrete and steel have struggled to guarantee. By placing the core roughly a mile underground, the company argues that any potential release would have to travel through dense rock and engineered seals before reaching the surface, a journey that would dramatically slow and dilute contaminants. Its own descriptions of the concept emphasize that the deep borehole acts as a built-in containment system, a point reinforced in featured materials that describe a built-in containment system formed by the surrounding geology.
External reporting on the design underscores that the reactors are intended to be installed in boreholes drilled a mile deep, using the natural geological properties of the subsurface to provide shielding and isolation. One technical overview explains that the company’s units are meant to be lowered into boreholes drilled a mile deep, avoiding some of the large surface structures used in aboveground reactors. That approach does not eliminate the need for robust engineering, but it does change the risk profile, shifting from massive, highly visible containment domes to a system where distance and rock layers are part of the defense-in-depth strategy.
Cost, LCOE, and the promise of faster builds
Beyond safety, Deep Fission is pitching its underground SMRs as a way to tame the runaway costs that have haunted large nuclear projects. Traditional surface plants often face unforeseen construction costs and delays, a pattern that one analysis bluntly summarizes by noting that, unfortunately, surface level nuclear power plants frequently run into overruns due to complex civil works and regulatory hurdles. In contrast, Deep Fission’s plan to drill standardized boreholes and drop in modular units is presented as a way to avoid many of those site-specific headaches, a point highlighted in coverage of its intention to build nuclear power plants below the surface of the earth for large-scale data centers and utilities.
Cost competitiveness is often distilled into the levelised cost of electricity, or LCOE, and Deep Fission’s backers are keen to show that underground SMRs can compete with other low-carbon sources. Reporting on the company’s site selection work notes that LCOE estimates for different energy sources vary, but the International Energy Agency has said the LCOE for advanced nuclear can be favorable compared with other firm low-carbon options. In that context, Deep Fission’s decision to pursue projects in three US states is framed as a way to demonstrate that its underground units can deliver attractive LCOE once they move from concept to operation, especially if the company can keep construction timelines short and capital costs predictable.
Texas, Utah, Kansas and a 12.5 GW customer pipeline
Deep Fission’s ambitions are not limited to a single pilot site. The company has already identified three US states, Texas, Utah, and Kansas, as candidates for its initial underground SMR deployments, signaling that it wants to prove the concept across diverse grids and regulatory environments. Reporting on this effort notes that Letters of Intent have been signed in each of the three states to pursue joint development projects, a step that gives the firm a clearer path from engineering drawings to real-world installations. One account of the process highlights that Letters of Intent are now in place in Texas, Utah, and Kansas, positioning those regions at the front of the line for underground SMRs.
On the commercial side, Deep Fission says it has built a substantial pipeline of potential customers even before its first reactor is in the ground. The company reports that it has expanded its customer pipeline to 12.5 gigawatts, a figure that reflects interest from data centers, utilities, and other large energy users looking for firm, low-carbon power. In its own update, Deep Fission describes how this 12.5 gigawatts of potential demand is tied to projects in Texas, Utah, and Kansas, underscoring how quickly the idea of buried reactors has resonated with energy buyers that are under pressure to decarbonize without sacrificing reliability.
Timelines, pilots, and the road to 2026
For all the excitement around the concept, Deep Fission still has to prove that it can deliver a working underground reactor on a credible timeline. The company has publicly discussed plans to build 15 megawatt reactors a mile underground by 2026, positioning that first wave as a pilot that can validate both the technology and the regulatory pathway. One detailed report on the firm’s trajectory notes that the US nuclear startup intends to construct 15 MW reactors a mile underground by 2026, with a first pilot targeted by July of that year to serve data centers and other customers worldwide.
Other coverage of the company’s roadmap reinforces that Deep Fission is working toward a near-term demonstration rather than a distant, speculative project. Reporting on its progress describes how the firm is advancing a reactor it plans to bury a mile underground for safety, with a clear focus on the road to 2026 and beyond as it seeks to speed growth. One account of the company’s strategy explains that it is already planning for future expansion once the initial unit is operating, outlining plans and the road to 2026 that include scaling up manufacturing and streamlining regulatory approvals if the first projects succeed.
What deep-buried reactors could mean for the nuclear debate
If Deep Fission can execute on its vision, burying reactors a mile underground could reshape public debates about nuclear power, especially in communities that have long resisted large surface plants. The company’s concept of Powering Humanity from a Mile Underground is explicitly framed as a response to the perception that Traditional and next-generation reactors have not fully addressed concerns about safety, cost, and siting. By moving the most sensitive components out of sight and into deep rock formations, the firm is betting that it can make nuclear feel less intrusive and more like critical infrastructure that quietly supports data centers, factories, and cities from below, a narrative it advances in its own Powering Humanity materials.
At the same time, the idea of drilling boreholes a mile deep to host nuclear reactors raises its own set of questions about geology, long-term waste management, and regulatory oversight. Technical reporting on the concept notes that the units are meant to be installed hundreds of metres underground via a borehole, with the surrounding rock acting as part of the containment system, a design that regulators and local communities will need to scrutinize carefully. One early look at the technology describes how Deep Fission’s underground reactor concept aims to help with the clean energy transition, but also acknowledges that the company must demonstrate that difficult underground conditions can be managed over the full life of each unit. For now, the promise of compact, buried reactors is compelling, yet the proof will come only when Gravity and its successors are operating safely, reliably, and at scale beneath the surface.
More from MorningOverview