What the Moon can supply for power—fission, fusion, solar feedstock, hydrogen, storage, and the chemical and isotopic niches—and what the sourcing literature actually says about each.
The organizing principle
Lunar energy resources are not a single inventory. They divide along the same line that organized the nuclear-fuel compendium: consumable fuels versus capital infrastructure. A fusion fuel, a fission fuel, and a hydrogen feedstock are consumed and must be replenished. A photovoltaic array, a heat-storage block, or a betavoltaic cell is built once and depreciates. The Moon’s resource base is asymmetric between these categories, and the asymmetry is the argument of this part.
The sourcing base is equally asymmetric. Solar, storage, and fission have been studied to the point of pilot-plant design. Helium-3 fusion has moved, within the last several years, from a physics conjecture to a set of named machines with construction starts and fuel-supply agreements—though none has yet demonstrated net energy from deuterium–helium-3. Betavoltaics and isotopic power occupy a narrow, well-defined niche. The sections are ordered from the most mature to the most speculative, and each distinguishes what is demonstrated from what is modeled.
1. Solar energy and the lunar photovoltaic feedstock
1.1 The resource
The lunar surface receives approximately 1,361 W m⁻² at normal incidence, with no atmospheric attenuation and no weather. A 25% efficient array with a 45% duty cycle yields roughly 1,350 W m⁻² of average surface insolation at the illuminated fraction. The lunar day is 14 Earth days and the lunar night is 14 Earth days, so a surface installation at low or mid latitudes must either store two weeks of energy or shut down. The polar regions change the problem: certain ridges near the south pole experience near-continuous illumination, while adjacent permanently shadowed regions never see the sun.
1.2 The feedstock
Silicon is the obvious photovoltaic material because it is abundant in lunar materials, and ESA studies frame in-situ silicon production as the path to PV arrays manufactured on the Moon rather than launched from Earth. The industrial pathway to that silicon is treated in Part Two; the energy point is that the Moon can supply the semiconductor feedstock for its own power infrastructure.
1.3 System architectures
NASA’s lunar surface technology portfolio describes a layered approach: vertical solar array technology for polar illumination, regenerative fuel cells for night survival, and fission surface power for baseload. Combining regenerative fuel cells with vertical arrays enables 10 kW of supply and recharge to permanently shadowed regions. ESA’s ISRU pilot plant architecture assumes power from a European Charging Station for the Moon, the surface power node the plant draws on. The solar resource is therefore not merely an insolation number; it is a distributed architecture of arrays, storage, and a charging station.
2. Fission energy
Fission is the only lunar power source that provides continuous, illumination-independent baseload without a two-week storage mass penalty. NASA’s Kilopower project produced the 10-kilowatt-class Fission Surface Power system, and Artemis-era planning places a ~40-kW-class unit as the baseload component of an integrated lunar energy architecture alongside solar PV, radioisotope power, batteries and regenerative fuel cells, and regolith thermal energy storage.
The critical resource question is whether the fuel can be lunar. As a primary product, it cannot. Uranium on the Moon averages 27–30 ppb, with KREEP-rich materials reaching a few ppm. Read as a stand-alone ore, that concentration never closes a processing mass budget against the reactor it would supply: the plant mass required per unit of fuel exceeds the reactor it feeds, and the verdict holds even at the upper end of the KREEP range.
There is one qualification, and it matters for architecture. The mass-budget verdict applies to uranium as a primary product. It does not necessarily apply to uranium as a sidestream of a primary mineral operation. If a plant is already excavating, comminuting, and beneficiating regolith for titanium, iron, rare-earth elements, phosphorus, or zirconium, then the bulk of the mass and energy cost is charged to the primary product, and the actinides report to a small accessory-mineral concentrate that can be further processed at modest marginal cost.
The accessory minerals that host uranium and thorium on the Moon are known from returned samples. Ion microprobe analyses of Apollo 11, 12, and 14 materials show that U, Th, Pb, and the rare-earth elements are concentrated in apatite, whitlockite, zircon, baddeleyite, zirkelite (zirconolite), and tranquillityite (Andersen & Hinthorne, 1976; Andersen et al., 1972). Later work identifies monazite, thorite, and Zr–Ti-rich phases as additional U- and Th-bearing hosts (Crawford, 2019). These are trace phases, typically micrometre-scale inclusions within silicate or oxide grains, but they are precisely the phases that report to the incompatible trace-element concentrate when the host rock is processed.
A recent experimental demonstration makes the byproduct logic concrete. Li et al. (2023) used photovoltaic-grade silicon as a collector to extract ilmenite, rare-earth elements, and nuclear elements from KREEP-rich mare basalt at 1,873 K. Titanium and iron from ilmenite solidified into Si-based alloys, while the incompatible trace elements—including thorium and uranium—concentrated into a separate ITEs phase, where REEs and the nuclear elements together accounted for 82.61 wt% of that phase. The primary products were Ti and Fe alloys; the actinides were a sidestream concentrate.
The resource formulation is therefore bifurcated. Fission baseload on the Moon remains an imported-fuel, lunar-sited energy source, because no lunar deposit is ore-grade on its own. But a byproduct uranium and thorium stream is plausible at a KREEP-rich or ilmenite-processing plant, where the actinides are already being concentrated by the primary mineral operation. The fuel would not be mined for its own sake; it would be recovered because the plant is there.
3. Fusion energy and lunar helium-3
3.1 The inventory
The lunar regolith contains helium-3 implanted by the solar wind, with concentrations up to approximately 20 ppb by weight in titanium-rich mare soils, and the global inventory has been estimated at up to 10⁹ kg. The release pathway is thermal: heating soil to 673–873 K in vacuum liberates the implanted gas. The energy multiplication for mining, separating, and transporting helium-3 back to Earth has been estimated at roughly 250 for terrestrial fusion supply (Wittenberg et al., 1986).
3.2 The reactor programs
The lunar helium-3 inventory is only a resource if something can burn it. As of 2025–2026, several programs are explicitly designing, building, or testing deuterium–helium-3 machines, and their existence changes the assessment of lunar fusion fuel from a physics conjecture to an engineering schedule.
Helion Energy. Helion’s fuel of choice is deuterium–helium-3. Its seventh-generation prototype, Polaris, began initial operations in 2024 and achieved 150 million °C and deuterium–tritium fusion in 2025. In July 2025 the company began civil construction on the Orion fusion power plant in Malaga, Washington, following a State Environmental Policy Act mitigated determination of non-significance, with the plant intended to use deuterium and helium-3 fuel and to begin operations by 2028 under a power purchase agreement with Microsoft. Helion has also published a closed-loop D–³He fuel cycle that breeds helium-3 as a byproduct of deuterium–deuterium reactions—the supply-side complement to a lunar import strategy (Helion Energy, 2025).
TAE Technologies. TAE has spent nearly three decades developing field-reversed configuration fusion and has built five machines, with its current-generation device, Norm, advancing the technology toward a commercial plant. Its planned first power plant, Da Vinci, is targeted for 2031 at 50 MWe, with future plants scaling to 350–500 MWe. In 2026 TAE signed an agreement with Black Moon Energy Corporation, a lunar-development firm, to secure a prospective helium-3 fuel supply and to collaborate on commercial development. Black Moon plans a robotic lunar mission within five years to test the feasibility of commercial-scale helium-3 production (TAE Technologies, 2026).
Kronos Fusion Energy. Kronos has released a 2026 design study for a compact spherical-tokamak tritium/helium-3 breeder and a deuterium–helium-3 tandem-mirror generator, with the burner design closing at engineering gain QE=1.317±0.14 and a neutron fraction of 5.44% at a frozen reference point of 4,298.5 MW fusion power. The company’s readiness timeline targets a test burner around 2032 and a commercial unit around 2036, gated on specific physics checkpoints. The design is simulation-based and the company states explicitly that nothing in the study is a construction commitment, but it is a published, benchmarked D–³He design with a defined experimental pathway (Kronos Fusion Energy, 2026a, 2026b).
Wendelstein 7-X. The Max Planck Institute’s stellarator generated high-energy helium-3 ions for the first time in 2025 using ion cyclotron resonance heating, a world first in stellarator research. The experiment was not a D–³He burn; it used helium-3 ions to simulate the behavior of fusion-born alpha particles in a reactor-relevant plasma. Its significance for the lunar resource case is diagnostic—it demonstrates that helium-3 ion physics can be controlled in a magnetic confinement device, a necessary step toward any D–³He reactor (Max Planck Institute for Plasma Physics, 2025).
Smaller and conceptual programs. NovaFusion’s Claro-II project proposes reconfigurable twin helical accelerators for D–³He aneutronic fusion. A conceptual design integrating metamaterial-enhanced magnets with AI and quantum control has been published as a design study. A 2025 blueprint for a deuterium–helium-3 magneto-inertial fusion rocket proposes a pulsed field-reversed configuration core with direct magnetic nozzle exhaust and a four-phase roadmap from laboratory demonstration to a starship-class system. These are not construction programs, but they establish that the D–³He design space is being explored across multiple architectures rather than within a single company.
3.3 What the programs do and do not establish
Three programs—Helion, TAE, and Kronos—have named machines, target dates, and in two cases construction starts or fuel-supply agreements. None has yet demonstrated net energy from deuterium–helium-3. The Kronos design is explicitly a simulation study. Helion’s achieved fusion to date is deuterium–tritium, not D–³He, and its D–³He operation lies in the Orion build-out.
What the programs establish is a schedule. The first generation of D–³He machines is targeted for the late 2020s and early 2030s, and at least one company has signed a lunar helium-3 supply agreement in anticipation of that schedule. The lunar resource question is therefore no longer “is there a reactor?” It is “can lunar production be scaled to meet the fuel demand of reactors that are already on a construction timeline?” That is a harder question for the mining architecture than for the reactor physics.
The ISRU verdict shifts from deferred to schedule-linked. Helium-3 is a fusion fuel whose reactor technology is not yet commercially proven but whose development pathway is now dated. A lunar helium-3 industry would be built against a demand curve that is beginning to exist, rather than against a hypothetical one.
4. Hydrogen: the coupled energy and feedstock resource
Hydrogen on the Moon is water-derived, and water is concentrated in the permanently shadowed regions of the poles. Estimates of polar water ice deposits reach 600 million metric tons, cemented within regolith at depths exceeding 11 cm in the PSRs. That ice is both a life-support consumable and the feedstock for in-situ propellant production through electrolysis.
The energy significance of hydrogen is that it is a storage medium and a fuel, not a primary source. Electrolysis of lunar water produces hydrogen and oxygen; the pair can be burned in a fuel cell to produce electricity and water, or fed to a propulsion system. NASA’s technology portfolio explicitly pairs regenerative fuel cells with solar arrays so that the hydrogen–oxygen cycle serves as the night-survival storage mechanism. The IAC literature reaches the same conclusion from the architecture side: water ISRU for propellant can enable both chemical propulsion and nuclear thermal propulsion, and ISRU-based commodity production improves the sustainability of the lunar economy by minimizing Earth reliance.
5. Batteries and regenerative fuel cells
Batteries are the short-duration storage solution; regenerative fuel cells are the long-duration one. The division is a mass trade. A battery sized for a 14-day lunar night is enormous; a regenerative fuel cell sized for the same duration stores energy in chemical form and is recharged by the solar array during the day. NASA’s integrated architecture combines both, with batteries handling transient loads and regenerative fuel cells handling the diurnal cycle.
Neither is a lunar resource in the extractive sense. Both are manufactured systems whose feedstock—lithium, nickel, platinum-group catalysts, and the hydrogen and oxygen working fluids—must either be imported or, in the case of the working fluids, produced from lunar water. The resource statement is narrower than it first appears: the Moon supplies the hydrogen–oxygen chemistry and the water feedstock, not the cell hardware.
6. Thermal energy storage in regolith
Regolith is a thermal storage medium. It is cheap, abundant, and requires no processing beyond excavation and containment. NASA’s integrated lunar energy architecture lists regolith thermal energy storage as a distinct storage category alongside batteries and regenerative fuel cells. Thermal behaviour of regolith across extreme temperatures, presented at Space Resources Week 2026, examines how regolith behaves across the temperature range relevant to solar-driven reactors and thermal storage design.
The resource logic is that thermal storage trades mass for insulation. A well-insulated regolith block charged during the lunar day discharges heat through the night with no chemical conversion step and no working fluid. It is the lowest-technology storage option on the Moon, and for that reason it may be the first one deployed.
7. Betavoltaics and isotopic power
Betavoltaics convert the decay electrons of a beta-emitting isotope directly into electrical current. They are low-power, long-life, and radiation-hard—suitable for sensors, landers, and distributed instrument networks rather than for base power. The lunar resource connection is indirect: betavoltaic devices do not consume a lunar resource, but they compete with the radioisotope power systems that the lunar energy architecture includes as a generation category.
The isotopic resource base on the Moon is thin. There is no lunar production of the isotopes that betavoltaics and radioisotope generators depend on. Like fission fuel, isotopic power on the Moon is an imported consumable whose value is that it removes a storage requirement, not that it draws on a lunar deposit.
8. Chemical energy: fuel cells and metal combustion
Fuel cells are the chemical-energy workhorse of the lunar architecture. Hydrogen–oxygen fuel cells produce electricity and water, and the regenerative variant closes the loop with electrolysis.
A less conventional chemical pathway is metal combustion. Accelerated Combustion of Metals for Exothermic Heating, presented at IAC 2024, proposes an ISRU-based concept for long-term energy storage on the lunar surface in which lunar-derived metals are burned for heat (Maana Electric, 2024). The resource logic is that metal powders—aluminum and iron among them—are storable at ambient lunar conditions, have high energy density by volume, and can be produced from regolith by the same electrolytic processes that produce oxygen. The aluminum inventory that serves as a propellant feedstock in Part Two therefore also serves as a combustible energy store.
9. The solar-wind collector question
The lunar regolith is the passive default for helium-3, but it is not the only architecture. A magnetic scoop with an electrostatic isotope filter can intercept the solar wind directly, concentrate the charged particles with a field whose effective cross-section greatly exceeds any physical collector, and separate helium-3 from helium-4 by mass-to-charge ratio at the throat. A coil roughly twenty meters across can produce a capture radius of several hundred kilometers, and a 500 km capture radius at 20% efficiency corresponds to roughly 89 MW of collected particle power—useful as a propellant and fuel depot rather than as bulk energy, given that sunlight at 1 AU is roughly 2.4 million times more energy-dense per square meter (Nosuchdevice, 2026).
For the lunar context, the point is architectural. A collector in lunar orbit or at a lunar Lagrange point could supply fusion fuel and propellant without regolith mining. It competes with, rather than complements, the helium-3 extraction industry. No flight design exists; the physics is sound and the engineering is not.
Synthesis for Part One
The lunar energy resource base divides into three tiers of maturity and three tiers of ISRU value.
Tier one—demonstrated and near-term: solar PV, batteries, regenerative fuel cells, and regolith thermal storage. These are manufacturing industries whose lunar resource input is water for the hydrogen–oxygen cycle and, eventually, in-situ silicon for arrays. ESA, NASA, and ESRIC have moved from study to pilot-plant and challenge-scale demonstration.
Tier two—demonstrated but imported-fuel: fission surface power. It solves the night problem and draws on no lunar fuel deposit as a primary product—though a byproduct actinide stream from a KREEP or ilmenite plant is a plausible secondary source. Its ISRU value is architectural first and extractive second.
Tier three—dated reactor schedules, unbuilt mining architecture: helium-3 fusion, with betavoltaics at scale and direct solar-wind collection as adjacent long-horizon options. Unlike a decade ago, this tier now has named machines, construction starts, and a signed lunar fuel-supply agreement. What it does not have is a demonstrated D–³He burn or a lunar production plant. The resource is real, the demand curve is forming, and the mining architecture is the missing link.
The through-line is that the Moon’s energy future is mostly a storage and manufacturing future, not a fuel extraction future—with two exceptions that are both: water-derived hydrogen, which is already coupled to the propellant economy, and helium-3, which is now coupled to a reactor schedule rather than to a hypothesis. Uranium and thorium sit in a third, narrower category: not a primary energy resource, but a possible byproduct of the industrial feedstock economy described in Part Two.
References
Andersen, C. A., & Hinthorne, J. R. (1976). Ion microprobe mass analysis of lunar samples (NBS Special Publication 427). National Bureau of Standards. https://nvlpubs.nist.gov/nistpubs/Legacy/SP/nbsspecialpublication427.pdf
Andersen, C. A., Hinthorne, J. R., & Fredriksson, K. (1972). U, Th, Pb and REE abundances and ²⁰⁷Pb/²⁰⁶Pb ages of individual minerals in returned lunar material by ion microprobe mass analysis. Earth and Planetary Science Letters, 14(2), 195–200. https://doi.org/10.1016/0012-821X(72)90006-4
Crawford, I. A. (2019). Lunar resources: A review. Reviews in Mineralogy and Geochemistry, 85, 829–885. https://www.nhm.uio.no/english/about/organization/research-collections/people/emeriti/rtronnes/1/epmd/a-rev/revmin23-moon-19-crawford-resouces.pdf
European Space Agency. (2024). Exploration mission-pull technology requirements: ISRU pilot plant study. ESA Concurrent Design Facility. https://esamultimedia.esa.int/docs/business_with_esa/Exploration_Mission-Pull_Technology_Requirements_TDE_WP_2024.pdf
European Space Agency. (2025, January 30). ESA launches the Second Space Resources Challenge. ESA Commercialisation Gateway. https://commercialisation.esa.int/2025/01/esa-launches-the-second-space-resources-challenge-src/
European Space Resources Innovation Centre. (n.d.). Research. https://www.esric.lu/research
Helion Energy. (2025). How to engineer a renewable deuterium–helium-3 fusion fuel cycle. https://www.helionenergy.com/
Jolliff, B. L., Gillis, J. J., Haskin, L. A., Korotev, R. L., & Wieczorek, M. A. (2000). Major lunar crustal terranes: Surface expressions and crust–mantle origins. Journal of Geophysical Research: Planets, 105(E2), 4197–4216. https://doi.org/10.1029/1999JE001103
Kronos Fusion Energy. (2026a). Tandem-mirror design-space mapping for a D–³He burner: 0-D power balance benchmarked against WHAM and the Pastukhov-plug operating point. Zenodo. https://zenodo.org/records/22645911
Kronos Fusion Energy. (2026b). The D–³He fuel cycle does not close indigenously: Helium-3 supply, the lunar gate on burner-scale fuel, and the breeder’s strategic co-product (Paper 2.35, The Kronos 2026 Series). https://www.kronosfusionenergy.com/publications/papers/helium-3-fuel-cycle.pdf
Li, C., Wei, K., Li, Y., Ma, W., Lei, Y., Yu, H., & Liu, J. (2023). A novel strategy to extract lunar mare KREEP-rich metal resources using a silicon collector. Journal of Rare Earths, 41(9), 1429–1436. https://doi.org/10.1016/j.jre.2022.07.002
Luxembourg Space Agency. (2025). Space directory 2025. https://space-agency.public.lu/dam-assets/publications/2025/space-directory-2025-screen.pdf
Maana Electric. (2024). Accelerated Combustion of Metals for Exothermic Heating: Surviving the lunar night thanks to ISRU paradigm (Paper IAC-24,D3,IPB,14,x85363). International Astronautical Federation. https://iafastro.directory/iac/archive/browse/IAC-24/D3/IPB/85363/
Max Planck Institute for Plasma Physics. (2025, May 27). World premiere in fusion research: High-energy particles generated by radio waves in Wendelstein 7-X. https://www.ipp.mpg.de
NASA. (2026). Lunar surface technology. https://www.nasa.gov/lunar-surface-technology/
NASA Science Technology Mission Directorate. (2025, May 20). Restructuring development and implementation of ISRU in NASA’s Artemis Program [Presentation]. Luxembourg Space Resources Week. https://ntrs.nasa.gov/api/citations/20250003732/downloads/Restructuring%20STMD%20and%20ISRU-Sanders.pdf
Nosuchdevice. (2026). Solar wind particle fuel collectors: The case for a magnetic scoop 500 kilometers wide. https://nosuchdevice.com/eco-tech/futuristic-fuels/solar-wind-fuel-collectors/
TAE Technologies. (2026, August 5). TAE signs an agreement for future fuel supply options with helium-3. https://tae.com
Wittenberg, L. J., Santarius, J. F., & Kulcinski, G. L. (1986). Lunar source of ³He for commercial fusion power. Fusion Technology, 10(2), 167–178. https://doi.org/10.13182/FST86-A24797