Mining the Moon for Helium-3: How It Would Work, What It Would Cost, and Whether the Math Closes

In May 2025, a startup called Interlune unveiled a full-scale prototype excavator designed to ingest 100 metric tons of Moon dirt per hour and return it to the surface in a continuous motion. The target: helium-3, an isotope so rare on Earth that global terrestrial production runs around 22,000 to 30,000 liters per year, and so valuable that purified grades exceed $10,000 per liter. Within months, the company had signed offtake agreements with Bluefors (more than $300 million), Maybell Quantum, and the U.S. Department of Energy.

The contracts are real. The scarcity they address is real. Whether mining the Moon to solve it makes physical and economic sense is a harder question than the press releases suggest, and the answer depends on variables that have nothing to do with space technology. I explain why in this companion to my comprehensive helium-3 explainer, where I cover wh=y helium-3 matters, who needs it, and why it cannot be swapped for helium-4.

Why Helium-3 Is on the Moon (and Almost Nowhere on Earth)

The asymmetry comes from the Sun. Solar wind carries helium-3 particles outward through the solar system at roughly 400 kilometers per second. Earth’s magnetosphere deflects them. The Moon has no magnetosphere. For roughly 4.5 billion years, solar-wind helium-3 has been hitting the lunar surface, embedding itself in the top few meters of regolith (the loose, rocky dust that blankets the Moon), and staying put.

Retention is not uniform. The dark patches visible from Earth, the mare basalt regions like the Sea of Tranquility and the Ocean of Storms, are rich in ilmenite (an iron-titanium oxide) whose crystal structure traps helium-3 more efficiently than the lighter highland soils. Concentrations vary from a few parts per billion in the highlands to roughly 20 to 50 parts per billion in high-titanium mare basalts, based on Apollo, Luna, and Chang’e sample analyses plus orbital spectroscopy from the 1994 Clementine mission. NASA-cited estimates put the total lunar inventory at roughly one million tonnes, though this figure is extrapolated from a limited sample set and should be treated as an order of magnitude, not a measured reserve.

Parts per billion. That number deserves emphasis, because it defines the entire extraction challenge. At 30 ppb, every gram of helium-3 is dispersed across roughly 33,000 tonnes of regolith. The gas is there; getting it out at a cost and rate that close the economics is the engineering problem that Interlune and everyone before it has to solve.

How You’d Get It Out

Extraction has been studied since the 1980s, when Harrison “Jack” Schmitt (the Apollo 17 geologist who is now an Interlune co-founder) and the University of Wisconsin’s Fusion Technology Institute first proposed the concept. The basic process has three stages.

First, excavation. You scoop or agitate regolith from the surface to a depth of roughly one to three meters, where the solar-wind-implanted gases are concentrated. Interlune’s excavator prototype is designed to process 100 metric tons per hour in continuous operation, with the regolith returned to the surface after processing.

Second, thermal extraction. Heating regolith to approximately 600 to 700 degrees Celsius releases the embedded volatiles. About half the loosely bound gases (including helium-3) can be freed at lower temperatures through mechanical agitation alone, according to the early Wisconsin studies, but full extraction requires heating. The volatile mix that comes off includes not only helium-3 but also helium-4, argon, xenon, krypton, hydrogen, and other solar-wind-implanted species.

Third, separation. The released gas mixture is processed to isolate helium-3 from the other gases. This is where Interlune’s DOE-funded research on novel separation technology comes in. The company has stated its harvester is smaller, lighter, and less power-hungry than earlier academic concepts, though detailed specifications have not been published.

The numbers define the scale. At 30 ppb and 100 metric tons per hour, the excavator processes 2,400 metric tons of regolith per day. From that volume, a best-case extraction yield delivers roughly 70 to 80 milligrams of helium-3 per day, or about 25 to 30 grams per year of continuous operation. In liters at standard pressure, that is roughly 190 to 225 liters per year from a single excavator. The Bluefors contract calls for up to 10,000 liters per year, which at this rate would require dozens of simultaneously operating units, each with its own power supply, thermal processing module, and gas-separation system, all on the lunar surface. Interlune has not published fleet-scale architecture details.

For context, one square kilometer of mare regolith mined to three meters depth at 30 ppb contains approximately 33 kilograms of helium-3, or roughly 250,000 liters of gas at standard conditions. The resource is there. Accessing it at scale is a manufacturing problem on a body with no atmosphere, no infrastructure, and a 1.3-second communications delay.

The Economics: What the Contracts Actually Say

Interlune was founded in 2020 by former Blue Origin executives Rob Meyerson and Gary Lai, alongside the Apollo 17 geologist Harrison Schmitt. The company has raised $18 million in seed funding and holds grants from the DOE, NASA TechFlights, and the NSF.

Three contracts anchor the near-term business case. The Bluefors agreement covers up to 10,000 liters per year for delivery from 2028 to 2037, reported at roughly $300 million over the term. The Maybell Quantum deal covers thousands of liters per year between 2029 and 2035. The DOE Isotope Program has agreed to purchase three liters of lunar helium-3, deliverable by April 2029 at approximately today’s commercial market price, in the first DOE purchase of a non-terrestrial resource.

The mission roadmap lays out several steps this decade. Crescent Moon, a hyperspectral camera ridesharing to the lunar south pole, was scheduled for late 2025 to map helium-3 concentrations. Prospect Moon sends a lander to a selected high-concentration site. A full process demonstration around 2027 is intended to test the complete extraction-and-return chain. A pilot plant is planned for the surface by 2029. Interlune also received a NASA SBIR Phase III contract worth $6.9 million (awarded May 2026) for a regolith-heating extraction payload on a 2028 lander mission.

Meyerson cites a commercial price of about $20 million per kilogram for helium-3, and frames it as “the only resource in the universe that is priced high enough to warrant going to space and bringing it back to Earth.” At $20 million per kilogram and 7,500 liters per kilogram, that works out to roughly $2,700 per liter at standard conditions. Against current bulk market pricing of $1,900 to $2,600 per liter (pre-purification), the margin exists if and only if the total cost of lunar extraction, processing, and return falls below the terrestrial price, which is far from certain given the mission infrastructure required.

Why It Might Work

The arithmetic of units is the counterargument to the “you’d need to process unimaginable amounts of regolith” dismissal. A kilogram of helium-3 is nearly 7,500 liters of gas, and a dilution refrigerator needs only about 5 grams (40 liters). So a single kilogram of lunar helium-3 could charge approximately 200 dilution refrigerators. The resource does not need to be abundant in mass terms; it needs to be extractable in liter terms, because the application consumes liters, not kilograms. A few tens of kilograms per year covers the cryogenics market.

This unit insight is what makes the economics plausible. Combined with four structural tailwinds, the case is not trivially dismissible: the scarcity is real and worsening (a 20,000-to-40,000-liter annual deficit), the price is high and rising, launch costs are falling (SpaceX’s Starship changes the mass-to-lunar-surface economics), and one square kilometer of mare regolith holds enough helium-3 for decades of cryogenics demand.

The resource itself is functionally renewable over centuries: solar wind continues to implant helium-3 in the exposed regolith, replenishing it on a timescale of roughly a hundred years for any given surface patch. And Interlune is hedging its own technology risk by developing terrestrial helium-3 separation capabilities under a DOE research grant, which means the company can generate revenue from Earth-side separation while lunar operations scale up.

What Could Change the Economics Underneath It

Three developments, none of them from space, could shrink the market Interlune is selling into before lunar supply arrives.

The first is magnetic cooling. Continuous adiabatic demagnetization refrigeration (cADR) is the only technology that can reach millikelvin temperatures without helium-3. The German company kiutra has commercialized cADR and demonstrated a research platform operating continuously below 30 millikelvin with wiring for a five-qubit superconducting processor. If cADR reaches production scale for large arrays before lunar helium-3 comes online in volume, the cryogenics market shrinks under Interlune’s feet. I cover the current state, the numbers, and what remains to be proved in my assessment of ADR versus the dilution refrigerator.

The second is terrestrial breeding. Helion, the leading deuterium-helium-3 fusion company, plans to produce its own helium-3 from deuterium-deuterium fusion rather than source it externally. If D-D fusion at commercial scale becomes feasible, it would create a terrestrial He-3 source decoupled from tritium decay. Helion’s Polaris prototype reached 150 million degrees Celsius with D-T fuel in January 2026, but the D-He-3 operating regime requires roughly five times that temperature, and sustained D-D production is undemonstrated. This path is real but remote.

The third is simpler: invest more in terrestrial tritium production. The United States can irradiate more lithium to make more tritium, recovery from existing tritium inventories can improve, and Canada’s Darlington output can grow. The current squeeze reflects deliberate underinvestment, not a hard physical ceiling. Mundane industrial expansion competes with lunar mining every bit as much as exotic technology does.

The Technical Gauntlet

Every step of the lunar helium-3 chain is unproven at the scale the contracts assume.

Interlune’s Crescent Moon is a camera mission, not a mining test. Prospect Moon is a prospecting lander. The 2027 demonstration would be the first full extraction-and-return test, and even if it succeeds, converting a single-mission demonstration into a continuously operating pilot plant by 2029 compresses an enormous amount of engineering into two years: power supply (likely nuclear, given the two-week lunar night), thermal processing, gas separation, storage, and return-vehicle integration, all autonomous and all operating in an environment with temperature swings of more than 250 degrees Celsius, abrasive regolith, vacuum, and no servicing infrastructure.

The return leg is itself an underexamined challenge. Helium-3 at parts-per-billion concentration means the processed volume per gram is enormous but the final product is a very small quantity of gas. Getting that gas from the lunar surface to a customer on Earth requires a return vehicle, a re-entry system or orbital transfer, and terrestrial reception. Interlune has not published return-vehicle architecture.

No lunar helium-3 from non-terrestrial sources should be assumed available before 2029 at the earliest, and commercial volumes before the early 2030s, as my cryogenics infrastructure guide also notes.

The Geopolitics: Why China Is Already There

China is the only country that has collected and returned lunar samples from the mare regions with the highest helium-3 concentrations. Chang’e 5 returned approximately 1.7 kilograms from Oceanus Procellarum in 2020, and Chang’e 6 returned roughly 1.9 kilograms from the far side in 2024, both from basalt-rich areas. Chinese researchers identified a new mineral, Changesite-(Y), in the Chang’e 5 samples, and the analytical work on helium-3 retention in these samples is ongoing.

China’s stated target for a crewed lunar landing is before 2030. Its lunar program is backed by a relay satellite network (Queqiao-2), an ILRS (International Lunar Research Station) plan with partners including Russia, and a track record of executing missions roughly on schedule. For Beijing, lunar helium-3 is not only about cryogenics or fusion. It is about not depending on anyone else’s tritium for a material that could become strategically significant. I trace this pattern across the broader quantum supply chain in Quantum Sovereignty.

The legal framework remains unsettled. The 1967 Outer Space Treaty prohibits national appropriation of celestial bodies but does not explicitly address resource extraction by private or state-sponsored entities. The Artemis Accords, signed by 50-plus nations as of 2025, provide a framework for resource utilization, but China is not a signatory and operates under a parallel interpretation. A United Nations space-resources treaty draft is expected around 2027.

Where This Leaves the Business Case

Lunar helium-3 mining is not absurd and it is not assured. The physics is real (the gas is there, in the quantities estimated, deposited continuously by the Sun), the demand is real (terrestrial supply falls short of consumption by tens of thousands of liters per year), and the economics are plausible at current prices if the extraction and return costs can be held below the market price.

The risks are equally concrete. The technology is unproven at every stage past excavation prototyping. The cryogenics market that anchors the near-term contracts is bounded (modality-dependent, recirculated, scaling with new-machine builds not throughput) and racing against an ADR alternative that is advancing rapidly. The fusion market that would make the numbers truly large does not exist and is not close. And terrestrial supply can expand if the investment case materializes.

The $300 million Bluefors contract is a real vote of confidence from the industry’s most important fridge-maker. It is also structured as “up to” 10,000 liters per year, which gives both sides flexibility. Interlune has raised $18 million against a mission plan that requires substantially more. Everything beyond the Crescent Moon camera ride is contingent on execution that has not been demonstrated.

My read: treat Interlune’s contracts the way my cryogenics infrastructure guide frames them, as long-term strategic hedges, not near-term procurement dependencies. If you are managing a quantum hardware supply chain today, your helium-3 plan is recovery, recycling, and relationship with Laurentis or the DOE Isotope Program. The Moon is a decade-scale option, and options are worth holding. Confusing an option with a delivery date is how supply chains break.