The Army has selected five suppliers and aims to deploy at least one first microreactor at a military installation by September 30, 2028. Contracts could total up to approximately $2.2 billion over FY2027-FY2031, subject to milestone payments. The Department of Defense is imposing an accelerated schedule on the sector. What is at stake goes beyond military logistics: the Janus program extends Project Pele and aims at commercial deployment of microreactors on military installations, without constituting the first full-scale trial of this technology.
The Essentials
- The U.S. Army is financing the deployment of nuclear microreactors on its bases, with contracts potentially totaling up to approximately $2.2 billion over FY2027-FY2031, five suppliers selected, and a deadline set for September 2028 (U.S. Army Janus Program).
- In 2026, Project Pele was in manufacturing and preparation for testing; it is not documented as having achieved criticality.
- Regulatory obstacles exist: procedures for transfer between military certification and commercial licensing have not yet been formalized in a clear manner.
- The military model compresses timelines by leveraging its own regulatory authority and guaranteed orders—two levers absent from the civilian market.
- Two trajectories are open: a regulatory bridge that creates a template for rapid deployment in the private sector, or a persistent split between military and civilian standards.
The Army as an Early Customer
The U.S. Army’s Janus program rests on a simple bet: a customer that pays upfront, that accepts its own regulatory framework, and that is not afraid of risk can advance an entire sector where the market hesitates. This is precisely the role the Department of Defense played for early navigation satellites or early high-performance chips.
By September 30, 2028, the goal is to have at least one microreactor capable of producing electricity at an Army installation. The timeline is tight. The civilian nuclear sector typically measures its timelines in decades, from initial design to commissioning of a facility. Here, the horizon is less than three years from contract signature.
What makes this bet credible is the nature of the product itself. A microreactor produces between 1 and 20 megawatts of electricity—enough to power a mid-sized military base without dependence on a vulnerable external grid. The Army has strategic reasons to move quickly: a base that depends on the civilian electrical grid is a base whose adversary can cut off power. Energy independence is here an operational question, not merely an environmental one.
Project Pele: Manufacturing and Testing Ahead
In 2026, Project Pele was in manufacturing; its integrated testing remained ahead following delivery and fuel loading. Criticality, a technical term designating the moment when the nuclear fission reaction becomes self-sustaining, would constitute validation that the physics delivers on its promises at real scale.
Project Pele is being developed by BWXT Advanced Technologies for the Department of Defense. The reactor is designed to be transportable: it fits in several standardized containers, can be moved by truck or cargo aircraft, and starts in a matter of hours. This mobile character fundamentally distinguishes it from a conventional power plant. On the military side, this means a deployed unit in a region without infrastructure can carry its own power source.
The integration testing that will follow delivery will constitute a demonstration phase, opening thereafter to deployment qualification and industrialization. The five Janus program suppliers are working on varied designs, which will allow the Army to compare actual performance and avoid dependence on a single supplier. The Defense Innovation Unit coordinates some of these partnerships, applying rapid acquisition methods typically reserved for software or drones.
Two Regulatory Pathways That Do Not Yet Communicate
The U.S. Army has its own nuclear safety authority. It does not need approval from the Nuclear Regulatory Commission, the NRC, to operate a reactor on a military base. This exemption regime is old: it dates back to the first military reactors of the 1950s. It has allowed the Army to move faster, but it has also created a regulatory separation between military and civilian standards.
The problem becomes clear when you imagine what comes next: a military microreactor that works well, whose costs have been amortized by the state, and that a civilian operator would want to buy to power a data center or an isolated mine. This transfer would require an NRC license. The procedures for moving from military certification to commercial licensing have not been formalized in a clear and harmonized framework. Safety standards, documentation requirements, inspection procedures are different. The civilian operator would face significant certification work, even for a reactor already in operation.
The Department of Energy launched in March 2026 its Nuclear Energy Launch Pad program to accelerate commercialization of new designs. In February 2026, the Federal Register published a regulatory framework applicable to fusion reactors, a signal that the NRC is beginning to adapt its tools to technologies that its existing regulations did not cover. These two initiatives point in the right direction. The problem of military-to-civilian transfer for fission microreactors remains partially unresolved.
This bottleneck is the real test of the program. Financing covers prototypes owned and operated by suppliers, while also aiming at industrial development and future deployments.
Who Wins, Who Waits, and What Each Risks
The five suppliers selected by the Janus program have obtained something rare in the nuclear industry: a committed customer, massive public financing, and a short timeline. In exchange, they assume considerable reputational risk. If a design fails to meet the timeline or satisfy military safety requirements, the consequences extend beyond the contract: they weigh on the company’s ability to raise private capital for its civilian activities.
For civilian actors, the situation is more ambiguous. Data centers, whose electrical consumption is growing faster than network operators can absorb, are looking at microreactors as a potential solution. Major technology companies have begun financing players in the advanced nuclear sector. But they are waiting to see if a transfer regime can simplify the move to a civilian framework. The question of the electrical grid as a lever of power makes this dependence even more visible: a local and controllable energy source is worth more than a grid connection when the grid itself becomes terrain for competition.
Operators of isolated mines, industrial sites off-grid in polar or desert regions, and certain ports present the same demand profile: significant consumption, difficult or costly grid connection, and higher price tolerance than the residential consumer. These are the first natural customers for a civilian microreactor sector. Their decision to invest depends in large part on the existence of a predictable regulatory framework.
The Bridge or the Split: What the Coming Years Decide
The fundamental question is not whether military microreactors will work. Elements from 2026 show technical progress, without yet demonstrating that military microreactors will operate sustainably. The question is whether military deployment creates conditions for a rapid civilian sector, or whether it simply produces a fleet of state reactors with no commercial descendants.
The first scenario, that of a successful bridge, assumes that military and civilian standards converge before 2028. A possible transfer regime would probably require interagency coordination and regulatory action by the NRC to clarify which data generated during military deployment are acceptable as safety evidence for a commercial license, what additional inspections are necessary, and who finances them. If such a framework emerges before the first Janus reactors are operational, suppliers could design their documentation files for both markets simultaneously. Certification costs would be amortized over a larger volume, timelines accelerated, and civilian actors could reduce their certification efforts.
This scenario would have concrete and rapid effects. A reactor whose certification cycle is partially covered by military investment would potentially be cheaper to deploy in the civilian sector. Economies of scale become accessible. The five suppliers can invest in serial manufacturing capacity rather than producing prototypes one at a time. It is on this condition that the microreactor can meet the demand from data centers or isolated industrial sites before 2035.
The second scenario, that of the split, is the symmetrical risk. If military and civilian standards remain separate, each supplier would have to bear additional certification efforts for each market. For many, this would represent a major obstacle. Capital would risk concentrating on the safer military market, and economies of scale would become hard to achieve.
Capital invested in one sector does not serve the other. Data centers and industrial operators wait further, or turn to other solutions: natural gas, massive battery storage, or long-distance electricity imports.
The signals to watch are precise. First, if Project Pele initiates a formal procedure with the NRC in the months following its military deployment, that is a signal that the bridge is actively being pursued. Next, if one of the five Janus suppliers announces a civilian customer before 2028, it means the market anticipates a resolution of the regulatory problem. Finally, if the NRC publishes a framework specific to microreactors emerging from military programs, it signals that the administration has decided in favor of convergence.
The Department of Energy’s Nuclear Energy Launch Pad is designed precisely to accelerate this type of transition. Its mandate includes supporting actors toward commercial certification. If it financed part of the NRC qualification costs for technologies from the military program, it could help make the transition economically viable. The acceleration of the energy transition in other sectors shows that public investments in immature technologies are often the necessary condition before the private sector takes over: the same pattern could apply here, provided that the regulatory handoff is anticipated.
Lessons from Other Sectors
The history of dual-use military and civilian technologies offers useful precedents, though the analogies are imperfect. The Internet was designed for the Department of Defense before becoming a global civilian infrastructure. GPS, developed and financed by the U.S. Army, is today the reference navigation system for billions of civilian devices. In both cases, the transition required a deliberate political decision to make the technology accessible to the civilian sector, not an automatic migration.
Nuclear presents an additional complexity that electronics did not have: safety is at the center of the economic model and regulatory framework in a way that has no equivalent in semiconductors or network software. A decision for regulatory convergence cannot be made in six months based solely on commercial opportunity. It requires operating data, feedback from real-world conditions, and validation by an agency whose independence is precisely what gives value to its certification.
This is why the 2026-2028 window is decisive. The Janus reactors are in selection and prototyping phases. They will produce real operating data: thermal behavior, reliability, maintenance, waste management. This data would be relevant for the NRC to assess an adapted certification framework. Harmonization of minimum documentation requirements between the Pentagon and the NRC could reduce administrative redundancies, while still requiring an NRC license review specific to each civilian application.
Convergence cannot be decreed, but it can be prepared. This is what the Nuclear Energy Launch Pad and ongoing discussions at the NRC on new designs are separately doing. The question that remains open: do these two processes come together before the first military reactors enter service, or after?
Sources
- U.S. Army Janus Program, Utility Dive, announcement of selection of five suppliers
- U.S. Department of Defense, Army Janus Program announcement (DoD press release, August 2026)
- U.S. Department of Energy, Nuclear Energy Launch Pad, March 2026
- Federal Register, Regulatory framework for fusion reactors, February 2026
- Defense Innovation Unit, Microreactor partnership, announcements 2025-2026
- BWXT Advanced Technologies, Project Pele, manufacturing documentation and demonstration testing
