Lunar soil is not fuel, but it contains the ingredient that makes fuel logistics brutal: oxygen. Most lunar regolith is about 40 to 45 percent oxygen by mass, locked in silicates, oxides, and glass. A plant that can free that oxygen could support life support, metal production, and the oxidizer side of lunar ascent propellant. The Moon is rich in oxygen locked inside minerals. Turning that oxygen into breathable gas or rocket oxidizer is a reactor, power, mining, storage, and maintenance problem. AI-generated image AI-generated editorial image for this explainer. Key Stats 40-45% Oxygen by mass in typical regolith 1,600°C+ Molten electrolysis temperature 20%+ Potential oxygen yield target LOX Main propellant product How It Works On Earth, oxygen is cheap because air is free and industry is mature. On the Moon, every kilogram launched from Earth carries a launch cost, a lander cost, and a scheduling risk. Oxygen dominates the mass of many chemical propulsion combinations. Liquid oxygen is about 86 percent of the mass in a methane oxygen propellant pair at common mixture ratios, and it is also the consumable astronauts need every day. That is why oxygen is usually the first serious target for in situ resource utilization. The catch is chemistry. Lunar regolith is not a pile of loose oxygen atoms. It is a mix of minerals such as anorthite, pyroxene, olivine, ilmenite, and glassy impact products. The oxygen is bonded to silicon, aluminum, iron, magnesium, calcium, and titanium. Extraction means breaking strong chemical bonds at high temperature, then separating, cooling, purifying, compressing, and storing the product. Engineers usually discuss three broad families of oxygen extraction from lunar soil. Hydrogen reduction targets iron oxides, especially ilmenite, and produces water vapor that can be electrolyzed into oxygen and recycled hydrogen. Carbothermal reduction uses carbon-bearing gases to pull oxygen out of oxides at high temperature. Molten regolith electrolysis melts the soil and drives oxygen ions through an electrical process while leaving metals or metal alloys behind. Hydrogen reduction is attractive because the process temperature is lower than full melting. Typical concepts heat ilmenite-rich soil to roughly 900 to 1,100 degrees Celsius and flow hydrogen through the reactor. The reaction makes water, then electrolysis splits that water. Hydrogen returns to the reactor, while oxygen is collected. The limitation is feedstock. The process works best where titanium and iron content are high, which is why older studies focused heavily on mare soils. Carbothermal reduction pushes temperature higher and can work across more mineral types. Methane or carbon monoxide participates in reactions that remove oxygen from oxides, often producing carbon monoxide or carbon dioxide that is recycled through additional steps. The hardware has to survive high heat, abrasive dust, carbon chemistry, seals, valves, and heat exchangers in vacuum. The process can produce good yields, but it is not a simple furnace with a tank attached. AI-generated image The engineering value is in the full process chain, not only the headline component. The Engineering Tradeoffs Molten regolith electrolysis is the most direct and most ambitious option. It heats bulk regolith above roughly 1,600 degrees Celsius until it melts, then uses electrodes to release oxygen at the anode and reduced metals at the cathode. Because it does not need imported hydrogen or carbon as a consumable, it has a strong long-term appeal. The hard parts are electrode lifetime, corrosion, containment, power electronics, insulation, startup energy, and slag handling. Yield matters. Studies of oxygen extraction from regolith often cite potential yields around 20 percent by mass for some solid state or partially molten processes and up to the high 20 percent range for more aggressive molten reduction approaches. Those numbers are not a promise of field production. They are chemistry and reactor targets. A real plant loses time to excavation, sieving, heating, cooling, maintenance, rejects, dust control, and storage. Mining is half the plant. A reactor that can make oxygen from one kilogram of carefully prepared simulant is not enough. A useful lunar plant needs excavation equipment, conveyors or sealed hoppers, particle sizing, dust isolation, feed metering, waste handling, and ways to keep abrasive grains out of bearings and seals. Lunar dust is sharp, clingy, and electrostatically troublesome. It can damage radiators, optical sensors, joints, gaskets, and thermal surfaces. Power is the next gate. Heating rock to 1,000 degrees Celsius or melting it above 1,600 degrees Celsius takes large amounts of energy, then the plant needs electrolysis, pumps, compressors, controls, avionics, thermal management, lighting, communications, and mobility support. A small demonstration might run from solar arrays and batteries during lunar day. A propellant-scale plant probably needs large solar farms with storage, power beaming, or fission power to run through interruptions. The storage problem is easy to underestimate. Oxygen leaves the reactor hot and chemically messy. It has to be purified, dried, compressed or liquefied, transferred into tanks, and kept within temperature and pressure limits. Liquid oxygen boils at about minus 183 degrees Celsius at one atmosphere, so cryogenic storage on the Moon is both helped and hurt by the environment. Shadowed cold can reduce boiloff, but plumbing, valves, insulation, and transfer lines still face thermal cycling and dust. A lunar oxygen plant also has to close its loops. Imported hydrogen, carbon, catalysts, seals, electrodes, and filters are expensive if they are consumed quickly. The more a process depends on recycled reactants, the more leaks and side reactions matter. A one percent loss in a lab may be fine. A one percent loss every production cycle in a remote plant can become a resupply bill. Approach Strength Main constraint Lower complexity Easier to test and maintain Often lower peak performance Higher performance Better output and flexibility More power, heat, controls, and cost Integrated system Can become real infrastructure Needs supply chain, standards, and field service Why It Matters Commercially The first customers will not need thousands of tons. Early demand could be oxygen for life support recharge, emergency reserves, metal oxide processing experiments, welding, surface construction, and small ascent or hopper demonstrations. The first plant that makes kilograms per day reliably will be more valuable than a paper design for tons per year. The Moon rewards reliability before scale. NASA and commercial teams have already treated ISRU as a stepwise program. Demonstrations such as oxygen extraction experiments, polar volatile prospecting, regolith handling tests, and surface power contracts are all pieces of a larger industrial chain. Companies working on lunar excavation, landing pads, power towers, communications, and cryogenic transfer matter because oxygen production depends on all of them. Location decides strategy. A polar site with water ice nearby may favor water mining and electrolysis if the ice is accessible, clean enough, and politically available. A site without reliable ice may favor regolith oxygen because the feedstock is everywhere. Mare regions with higher ilmenite content may favor hydrogen reduction. Highlands may push designs toward molten processes that can handle broad mineral mixtures. The business case is tied to cadence. If only a few landers arrive per year, shipped oxygen may remain simpler. If Artemis, commercial landers, cargo hoppers, pressurized rovers, and surface habitats create regular demand, local oxygen becomes infrastructure. The inflection point is not one reactor. It is traffic. Waste is not necessarily waste. Oxygen extraction can leave reduced metals, glass, ceramics, or slags that may become feedstock for construction,