IN Brief:
- Antares, BWXT, General Atomics, Radiant, and Westinghouse have each been paired with an initial US Army installation.
- The agreements carry a combined ceiling of about $2.2 billion across FY2027–2031, with payments tied to technical milestones.
- The Army expects private investment to help deliver more than 20 contractor-owned and operated microreactors across defence installations.
The US Army has selected five nuclear-energy companies and five initial installations for its Janus microreactor programme, committing up to a combined $2.2 billion to the development, construction, and operation of small nuclear power systems on military bases. Government funding will be released against technical milestones, with private investment expected alongside the Army’s contribution.
Antares Nuclear has been paired with Fort Bragg in North Carolina, BWXT Advanced Technologies with Fort Campbell in Kentucky, General Atomics Electromagnetic Systems with Fort Hood in Texas, Radiant Industries with Fort Benning in Georgia, and Westinghouse Government Services with Fort Drum in New York. The agreements run across fiscal years 2027 to 2031 and are based on Other Transaction Authority mechanisms rather than conventional fixed-price reactor purchases.
The Army expects the government and private funding to support more than 20 microreactors across Department of War installations. That does not mean all five companies are guaranteed to reach deployment. Each supplier has a separate technical programme and will receive government money after achieving agreed milestones, leaving the Army able to stop or reshape individual projects if engineering or financial progress falls short.
Janus was launched in October 2025 with the Defense Innovation Unit to accelerate commercially developed reactor technology towards sustained operation on military installations. The programme uses a contractor-owned and operated model, with the Army providing technical oversight and regulatory support rather than acquiring every reactor as government-owned plant. An executive-order deadline requires the Army to place at least one regulated advanced reactor into operation at a domestic military installation by 30 September 2028.
The five suppliers are developing materially different systems. General Atomics, for example, is advancing its Tactical Energy System, a liquid-metal-cooled microreactor with a baseline net electrical output of about 5MW and an architecture that the company says can scale towards 20MW. Its design uses uranium-zirconium hydride fuel, natural-circulation cooling, and modules intended to be transported by truck or rail.
Radiant’s Kaleidos follows another route, with the company working on a transportable 1MWe reactor. The diversity is intentional: Janus is not standardising immediately around one design and then asking several companies to manufacture copies. The Army is instead funding competing commercial approaches through technical and site-specific milestones, giving the programme several routes towards the September 2028 target while retaining the option to narrow the field.
Military installations provide a demanding early market for the technology. Bases depend on civilian electrical networks for most routine supply, while communications, command infrastructure, computing, maintenance facilities, sensors, and security systems still require power if the surrounding grid is disrupted. Diesel generators provide conventional backup, but fuel storage and resupply create additional dependencies, particularly at remote or strategically important locations.
A microreactor able to provide continuous on-site baseload power offers a different resilience model, but the reactor vessel is only one part of the industrial problem. Deployment requires nuclear fuel, specialised materials, heat-transfer equipment, instrumentation and control, shielding, site preparation, security systems, grid interfaces, transport arrangements, trained operators, waste plans, and a regulated maintenance system. Those supply chains have to develop alongside the reactor designs if Janus is to progress beyond prototype demonstrations.
The Army has already linked Janus explicitly to the wider US nuclear industrial base. Its original programme structure includes support for the uranium fuel cycle and nuclear supply chain, while the milestone model is intended to help companies establish commercially viable products rather than defence-only designs. The stated objective is that successful reactors should ultimately be capable of being sold to other customers rather than relying indefinitely on military demand.
That commercial objective changes the programme’s engineering incentives. A bespoke reactor meeting one base’s requirements could achieve a military deployment milestone without creating an economical production line, whereas a standardised product manufactured repeatedly has a better chance of spreading design, licensing, and factory costs across several customers. The challenge is retaining enough commonality while adapting each installation to local grid, safety, environmental, security, and civil-engineering conditions.
Army regulation is another distinctive element. The Army Reactor Regulatory Office has statutory authority over reactors deployed for military missions, and Janus is using that route while drawing on Department of Energy expertise and safety frameworks developed through earlier programmes. Officials have said the aim is to maintain standards sufficiently aligned with the civilian sector that suppliers do not end up with military-only reactors requiring extensive redesign before commercial deployment.
Janus now has named suppliers, named sites, and a funding mechanism, but the most difficult programme milestones remain physical. Design completion, fuel availability, site preparation, regulatory approval, factory production, transport, construction, commissioning, and sustained operation all sit between the 26 August awards and an operating reactor. With the first deadline just over two years away, progress will be measured less convincingly in announced contract ceilings than in how quickly those engineering stages begin to close.
