Lunar ISRU is not one technology. It is an industrial chain: excavate material, separate useful compounds, refine them, store products, and use them in life support, construction, power systems, and transportation. The highest-value early products are water, oxygen, hydrogen, metals, shielding mass, and landing-surface material.
Strategic conclusion
- Oxygen from regolith is the most mature high-value lunar ISRU pathway. Lunar soil contains more than 40% oxygen by mass, chemically bound in silicates and oxides.[8]
- Water ice at the poles is potentially the most valuable resource, but its usable concentration, physical form, depth, and accessibility remain incompletely characterized.
- Regolith construction is technically plausible and has been demonstrated with terrestrial simulants, but no lunar system has yet autonomously built and qualified a permanent habitat.
- Hydrogen–oxygen propellant can be produced from water, but complete lunar extraction, purification, liquefaction, storage, and flight-scale use remain unproven.
- Aluminium and titanium production is chemically feasible but industrially immature. Oxygen extraction is easier and should precede large-scale metal refining.
- The major unresolved risks are not basic chemistry; they are excavation rates, abrasive dust, thermal control, power availability, volatile loss, equipment lifetime, autonomous operation, and reliable materials certification.
1. Polar water ice
### Resource location
Water is expected primarily in permanently shadowed regions (PSRs) near the lunar poles. These craters receive little or no direct sunlight, allowing volatile compounds to survive for geological periods. NASA orbital observations using instruments including M3, LOLA, and LAMP identified and mapped candidate polar volatile deposits, but orbital signatures do not yet provide a complete mineable-reserve estimate.[8]
The critical distinction is between:
- Detected or inferred hydrogen/water signatures
- Recoverable ice
- Economically extractable ice at a known concentration and depth
Only the first category has been established at broad regional scale. The latter two require landed prospecting and pilot extraction.
### Extraction architecture
A practical water-mining system would require:
1. Robotic excavation of icy regolith.
2. Sealed transport to prevent volatile loss.
3. Heating in a vacuum or controlled-pressure reactor.
4. Capture and purification of released water vapour.
5. Electrolysis into hydrogen and oxygen.
6. Liquefaction or high-pressure storage.
7. Distribution to life-support and propulsion systems.
A NASA/JAXA-oriented processing architecture has been evaluated for both granular and consolidated icy regolith. The design heats regolith, captures vapour in a cold trap, removes contaminants, and electrolyses the recovered water into hydrogen and oxygen.[7]
NASA’s Lunar Auger Dryer ISRU (LADI) work has targeted continuous icy-regolith processing at approximately 1.78 kg of water per hour with a target extraction efficiency of 75%. The subsystem has reached approximately Technology Readiness Level 4, meaning laboratory validation in a relevant-environment development context, not lunar deployment.[2]
### What is demonstrated
- Remote orbital detection and mapping of polar hydrogen and suspected water-bearing areas.
- Laboratory extraction of water from lunar simulants.
- Ground testing of excavation, transport, heating, vapour capture, and water-processing subsystems.
- Designs for mobile water capture and polar regolith transport have been ground-tested with lunar simulants.[5]
- Electrolysis of recovered water into hydrogen and oxygen is established terrestrial technology.
### What remains unproven
- The actual abundance and distribution of mineable ice at a selected lunar site.
- Whether ice is concentrated in accessible layers or dispersed at low percentages through difficult regolith.
- Excavation of real PSR material under lunar vacuum, cryogenic conditions, and severe dust exposure.
- Continuous operation through long lunar night periods.
- Industrial-scale water production with acceptable energy consumption.
- Long-duration storage of cryogenic hydrogen and oxygen near the lunar surface.
### Ark relevance
Water is both a life-support consumable and a strategic industrial feedstock. It provides:
\[
2H_2O \rightarrow 2H_2 + O_2
\]
By mass, water yields approximately 11.1% hydrogen and 88.9% oxygen. One tonne of water can theoretically produce about 111 kg of hydrogen and 889 kg of oxygen, before process losses. The oxygen is useful for breathing and propulsion; hydrogen is valuable as rocket fuel, chemical feedstock, and energy-storage medium.
Water also provides radiation shielding. A habitat surrounded by several tonnes of water gains shielding while retaining access to a recoverable resource.
2. Oxygen extraction from regolith
### Resource base
Lunar regolith is dominated by oxygen-bearing minerals. Oxygen is not present mainly as free gas; it is chemically bound in silicates, oxides, and other minerals. NASA’s lunar ISRU strategy uses a planning value exceeding 40% oxygen by regolith mass.[8]
At a theoretical 40% yield, processing 1 tonne of regolith could release up to 400 kg of oxygen. Actual plant yield will be lower after incomplete reaction, contamination, feed handling, and energy losses.
### Main process families
#### Molten regolith electrolysis
Molten regolith electrolysis (MRE) melts soil at very high temperature and passes electrical current through it. Oxygen evolves at an electrode while metals accumulate as co-products.
Potential products include:
- Oxygen
- Iron
- Silicon
- Aluminium
- Titanium
- Calcium and magnesium products
NASA’s MRE work has successfully extracted oxygen from lunar regolith simulant in a vacuum reactor. The programme was reported as completed in 2025.[1]
A NASA progress review reported:
- More than 20 g of oxygen per kWh of thermal energy
- More than 20% oxygen yield by regolith mass in tested conditions
- More than 99.7% recovery of carbon in a carbon-assisted melt configuration
- Five consecutive melt operations under simulated lunar vacuum and thermal conditions
- A single melt equivalent to approximately 140 kg of oxygen per year at the tested scale
- Prototype hardware advanced to approximately TRL 5 in relevant-condition testing[2]
These figures demonstrate process viability, not a production-ready