Helium-3 has quietly become one of the most valuable substances on Earth, and the industry that needs it most is racing to find more of it anywhere it can, including 239,000 miles away.
Gold trades around $4,067 an ounce this week, or roughly $130,700 a kilogram. Helium-3, the isotope that keeps quantum computers cold enough to function, sells for something closer to $20 million a kilogram, according to Interlune co-founder and CEO Rob Meyerson. That gap did not open gradually. Before 2001, the isotope traded at a steady $40 to $85 a liter through federal auctions. Today bulk buyers pay up to $20,000 a liter, and the gas keeping the fridges of Google, IBM, and Amazon running is bought as a byproduct of a Cold War stockpile that stopped growing decades ago. Into that shortfall has stepped a Seattle startup with a plan that reads like the plot of a science fiction film: extract the isotope from the Moon.
Superconducting qubits, the fragile units of information at the heart of machines built by Google, IBM, and Amazon, only hold their quantum state near absolute zero, below 15 millikelvin. The only equipment that reaches and holds that temperature continuously is a dilution refrigerator, a plumbing system of nested copper chambers that mixes helium-3 and helium-4 to strip away the last traces of heat. Bluefors, the Finnish company that has delivered more than 1,500 of these machines and dominates the market, calls a single unit an investment north of $600,000. Helium-3 is not an optional input to that machine. It is the fluid that does the cooling, consumed in dozens of liters per system today and, as processors scale toward thousands of qubits, in the hundreds or thousands of liters per system tomorrow.
The isotope does not occur in extractable quantities anywhere on Earth. Virtually the entire commercial supply comes from tritium, a radioactive hydrogen isotope with a 12.3-year half-life, decaying inside America's nuclear weapons stockpile. Global terrestrial production runs at an estimated 22,000 to 30,000 liters a year, a figure that has not meaningfully grown since the U.S. tightened control of the material after a 2008-2009 shortage burned through more than half the country's reserves in a single year and left Congress ordering a full review of how the isotope was managed. Russian supply, once a secondary source, has been effectively cut off from Western markets since 2022.
Corban Tillemann-Dick, founder and CEO of dilution-refrigerator maker Maybell Quantum, put the arithmetic bluntly when his company signed its own helium-3 supply deal: the industry will go "from a few hundred quantum computers worldwide to thousands, then tens of thousands, and they all need to get cold." Consultancies disagree sharply on how fast that happens, with estimates for 2030 ranging from roughly 2,000 installed systems to more than 10,000, but every projection points the same direction, and none of them assume helium-3 supply grows to match.
The U.S. government has taken the shortage seriously enough to look past helium-3 entirely. On January 27, DARPA issued an urgent call for proposals to develop modular, sub-kelvin cooling systems that eliminate the isotope altogether, warning that reliance on it constrains both defense and commercial quantum programs. Competing approaches are emerging: Munich-based Kiutra has demonstrated continuous sub-30 millikelvin cooling using magnetic refrigeration with no helium-3 at all, and Chinese researchers published a paper on a rare-earth cooling alloy just two weeks after DARPA's announcement, reaching 106 millikelvin without a drop of the isotope. None of these alternatives yet match a dilution refrigerator's cooling power at industrial scale, which is why the companies that actually build the machines keep signing contracts for more helium-3 rather than betting the transition happens in time.
Interlune was founded in Seattle in 2020 by Rob Meyerson, former president of Blue Origin, alongside former Blue Origin chief architect Gary Lai and Harrison Schmitt, the Apollo 17 geologist and the only living person who has both walked on the Moon and trained as a scientist. The pitch is straightforward: helium-3 has been accumulating in lunar regolith for four billion years, deposited by solar wind that Earth's magnetic field deflects away. Interlune estimates more than a million metric tons of the isotope sit in the Moon's surface layer, dwarfing anything left in America's tritium stockpile.
The company has spent the past year converting that pitch into signed contracts. In May 2025, Maybell Quantum became Interlune's first commercial customer, agreeing to buy thousands of liters annually from 2029 to 2035 for its sub-10-millikelvin "Big Fridge" systems, the same week the U.S. Department of Energy's Isotope Program committed to purchasing three liters by April 2029. Four months later, Bluefors signed the largest of the three: an agreement worth more than $300 million for up to 10,000 liters a year between 2028 and 2037, the biggest commercial commitment to an off-world resource to date. Interlune has raised $18 million in venture funding to build toward those deliveries, backed by an Air Force SBIR contract, a NASA TechFlights grant, and a Texas Space Commission award, and has partnered with rover company Astrolab and equipment maker Vermeer to build the hardware that will do the digging.
The near-term business is not actually on the Moon. Interlune's harvesting process, excavate, sort, extract, separate, was designed to work on lunar regolith, but the company discovered the same cryogenic separation technology can pull helium-3 out of terrestrial natural gas supplies today, funded in part by a Department of Energy grant and a $1.25 million Air Force AFWERX contract. That earthbound business is what will keep the company solvent while it builds toward a lunar demonstration mission.
The Moon itself remains the harder problem. A U.S. Geological Survey astrogeologist has compared the scale of the undertaking to operating a full copper mine, since Apollo samples put helium-3 concentrations in regolith at only 10 to 50 parts per billion, meaning industrial quantities require processing millions of tons of lunar soil. Interlune's answer is a full-scale excavator, built with Vermeer, designed to ingest 100 metric tons of regolith an hour, tested on Earth using auxiliary components ahead of an eventual lunar deployment with Astrolab's rover.
Legal ambiguity hangs over all of it. The 2015 SPACE Act grants American companies rights to resources they extract from celestial bodies, but the Outer Space Treaty still forbids any nation from claiming sovereignty over the Moon itself, leaving commercial harvesting in a gray zone that has not been tested by an actual delivery. Interlune has structured its contracts as private sales rather than territorial claims, betting that a signed agreement with Bluefors matters more than an unresolved legal theory.
Whichever path wins, a solid-state alloy that needs no isotope at all, a terrestrial extraction process good enough to stretch the existing stockpile, or an actual shipment of gas mined from four billion years of lunar accumulation, the answer will determine who controls the coldest, and now priciest, ingredient in the machines expected to reshape drug discovery, cryptography, and computing itself. For a company that did not exist six years ago, being first to answer that question is worth more than the gold its product has already outpriced.