Twenty Kilowatts in the Dark: Why the US, Russia and China Are Racing to Put Nuclear Reactors on the Moon

The moon has a deadline now

In January 2026, NASA announced plans for a nuclear reactor on the moon by the end of 2030, following a December 2025 executive order from President Donald Trump directing the agency to build a launch-ready lunar reactor. In late August, NASA's Glenn Research Center posted draft documentation and asked contractors to prepare Lunar Reactor-1, a 20-kilowatt fission system for the lunar south pole.

The program accelerated through October. On the 8th, NASA Administrator Jared Isaacman and Energy Secretary Chris Wright signed a space-nuclear memorandum taking effect November 1; on the 20th, NASA named Sierra Lobo to build the power-conversion system and the Energy Department's Idaho National Laboratory to develop the reactor itself.

Moscow has answered with "Selena", a reactor for the China-led International Lunar Research Station announced by Roscosmos in May 2025, targeting 2036. In August, US Transportation Secretary Sean Duffy called it what it is: "a race to the moon, in a race with China."

What the two blueprints actually promise

Lunar Reactor-1 would produce 20 kilowatts, roughly the power use of 16 American homes, and run five years on the surface with no maintenance. It would weigh under 6,000 kilograms, fit a 3.7-by-5.5-metre rocket compartment, and burn high-assay low-enriched uranium (HALEU). Russia's Selena is the lighter, longer answer: up to 10 kilowatts, a decade of autonomous operation, 1,100 to 1,300 kilograms for a softer touchdown.

Both sides say the reactors will stay switched off until they reach the moon, keeping radioactive risk minimal until the last step. Both aim at the same ground: the south pole, near craters that have not seen sunlight in millions of years and hold frozen water. And both begin with a proving run; NASA's Space Reactor-1 Freedom, a nuclear-electric spacecraft slated for a December 2028 flight to Mars, releasing three Ingenuity-class helicopters, a payload called SkyFall, to survey landing sites. Freedom's lessons are meant to feed Lunar Reactor-1.

The landing nobody has solved yet

No lander has yet been developed that can safely set heavy radioactive material on the Moon's rugged surface, and there is no precedent for installing and operating a reactor in low gravity. Getting a fission system down intact, then running it with no human hands nearby, is an unsolved engineering problem.

Nor is it clear what the 2030 deadline really means. Simon Middleburgh, director of Bangor University's Nuclear Futures Institute, told Al Jazeera it likely covers building the reactor itself; while "the harder part will be landing on the moon and safely setting up the system." Launch regulation, a launch platform and a sustainable surface setup are, in his words, bigger challenges than the nuclear technology.

Open questions remain: whether HALEU fuel can be produced at the scale these programs need, it is in short supply, and whether Selena can truly be built and installed autonomously, as Roscosmos chief Yury Borisov claimed in 2024 when he told TASS the technology was "almost ready."

Why does the moon need a nuclear reactor at all?

The science case starts with water. Scientists believe frozen water sits trapped in the moon's shadowed craters, and once released, it could supply drinking water, air and rocket fuel for future bases. Mining it takes steady, heavy power. The sun shines on any one spot only half the time, so solar panels go dark for about two weeks every month; batteries and diesel would have to be launched from Earth at ruinous cost. Middleburgh estimates three or four reactors would be needed to mine ice and run a base, a permanent foothold the growing lunar economy already assumes is coming.

The strategic case is sharper. The New York Times assessed that Washington accelerated its schedule from fear that once a Chinese-Russian reactor sits on the surface, an effective no-go zone follows. Whoever's reactor reaches the best water ice first could declare a "safety zone"; not a land claim, but one that shuts others out anyway, because nobody sets up next to a reactor. In the new space race, power becomes presence and presence becomes precedence. A reactor is never just a power plant; it is infrastructure, and space is ceasing to be a project and becoming infrastructure.

How a reactor becomes territory without a flag

Legally, no one owns the moon. The 1967 Outer Space Treaty bans countries from claiming it, though they own the equipment they send there; and the treaty does not ban nuclear power in space. Antonino Salmeri, director of the nonprofit Lunar Policy Platform, told Al Jazeera a reactor can be lawful if it follows the treaty's principles: peaceful use and avoiding harmful contamination. Norbert Schorghofer of the Planetary Science Institute added that all signatories have the right to mine the surface for peaceful purposes.

But the treaty never imagined permanent hazardous infrastructure. A declared safety zone creates no territorial rights; yet a reactor shuts others out by its nature. The first arrival sets the practical terms of access. Meanwhile two rival rule systems are forming: the US-led Artemis Accords and the China-Russia International Lunar Research Station, which has drawn in partners across the Global South including Egypt, Pakistan, Venezuela, Thailand and South Africa.

The military backdrop is hardening. In September 2026 the US announced it had deployed weapons into orbit to defend its forces, likely space-based communications jammers, experts said, raising the stakes of every civilian program above the atmosphere, from reactors to the satellites both sides already watch with suspicion.

What breaks first if someone wins

The liability vacuum is the most concrete gap. A Lunar Policy Platform report released in September 2026 found existing liability rules "have not been designed to regulate the use of long-term hazardous equipment on another celestial body", no one has answered who is responsible if something goes wrong. The precedent is earthly: in 1978, the Soviet satellite Kosmos 954 scattered radioactive debris across nearly 124,300 square kilometres of northern Canada. On the moon, Middleburgh notes, contamination would be "extremely localised", but the legal fallout would not be.

Then the fuel politics: HALEU is in short supply, and Russia holds the world's largest stockpile of low-enriched uranium; the practical edge in a race where the fuel, not just the reactor, is a chokepoint. And the propulsion spillover: the same fission technology feeds nuclear-electric spacecraft, so the moon race directly accelerates deep-space capability on both sides. Each deadline justifies the other's; each launch proves the machinery for the next one.

Three dates that will decide the race

December 2028: Space Reactor-1 Freedom's nuclear-electric flight to Mars; the technology proving ground. December 2030: NASA's launch-ready deadline for Lunar Reactor-1, industry-led and already facing the hardest engineering problems. 2035–2036: China's ILRS nuclear plant, included in the program's December 2025 presentation, and Russia's Selena switching on for a decade of autonomous operation.

Until then, watch the lander programs, the launch licensing, and whether 2030 slips from hardware to paperwork. History counsels skepticism: the Soviet Union flew more than 30 fission reactors in Earth orbit between the late 1960s and the 1980s, while the US has spent $20 billion on advanced space nuclear programs over 70 years and never launched a reactor since a 1965 experiment that shut down after 43 days. The moon has a deadline now. It has had deadlines before.