Living off the land on the lunar surface.

The flagship ISRU pipeline is water-to-propellant. Rovers prospect the permanently shadowed regions near the poles, where temperatures stay cold enough to trap water ice for billions of years. Once a deposit is mapped, thermal mining heats the soil just enough to sublimate the ice into vapor, which is drawn off and condensed into liquid water.

Prospecting is the step most people skip, and it is the one that decides everything downstream. Orbital neutron spectrometry can tell you hydrogen is present within a few tens of kilometres, but not whether it is concentrated ice, thin frost, or hydroxyl bound into mineral grains. Only ground truth, drilling and analysing cores metre by metre, converts a promising map into a mineable reserve. A plant sized for five percent water that lands on one percent ground is a stranded asset.

That water is then electrolyzed, split by electric current into hydrogen and oxygen. Liquefied and stored, those two gases are exactly the cryogenic propellants that power high-performance rocket engines. A lander that arrives empty can leave full, refueled entirely from material it never had to launch.

Storage is quietly the hardest link in the chain. Liquid hydrogen boils at 20 kelvin and leaks through almost anything, so cryocoolers must run continuously to keep boil-off near zero. Many architectures dodge the problem by making only oxygen locally, which is about four-fifths of propellant mass anyway, and importing the comparatively small hydrogen fraction from Earth.

A parallel pipeline works the dry regolith found everywhere on the surface. Molten regolith electrolysis melts the soil and passes current through it, liberating breathable oxygen while leaving behind iron, silicon, and other metals that can be cast or printed into hardware. Nothing is wasted; the leftovers of one process are the feedstock of the next.

Oxygen on the Moon is not scarce, it is bound. Every extraction route is ultimately a way of spending energy to break a chemical bond that has been stable for four billion years. Estimates for molten regolith electrolysis cluster around 20 to 40 kilowatt-hours per kilogram of oxygen, so even a modest plant producing a few tonnes a year implies tens of kilowatts running continuously.

Continuously is the operative word. At most sites the lunar night lasts about 354 hours, and a plant that shuts down every two weeks loses its thermal state, wastes energy reheating, and cycles its hardware through the exact stresses that break it. Two answers exist: place the plant on a polar ridge with near-permanent sunlight, or bring fission power. Surface reactor programmes in the 10 to 100 kilowatt class are being developed for precisely this reason.

The uncomfortable geometry is that the best sunlight sits on high ground and the best ice sits in the dark craters below it. Bridging that gap with power cables, relocatable reactors, or shuttling icy regolith uphill is an unglamorous engineering problem that determines whether an entire architecture closes.

ISRU is a capital-expenditure trade. You launch a plant, a power system, spares, and the mass of everything that installs it, then you produce for free, or rather for the price of maintenance and electricity. Break-even arrives when cumulative production exceeds the launched mass of the system that produces it.

That framing makes the answer obvious in both directions. A two-week sortie should bring its own supplies; no plant pays back in fourteen days. Anything measured in years inverts completely, because Earth resupply scales with time while plant mass does not. The crossover is the single most important number in a lunar business case, and it moves with launch cost, plant mass, and production rate.

There is a second-order effect that matters more than the first. A propellant depot in cislunar space does not just save money on the Moon, it changes what missions are possible anywhere beyond it. Vehicles can launch dry and fill up in space, which collapses the mass they need to lift from Earth. That is the difference between the Moon as a destination and the Moon as infrastructure.

Explural models these trade-offs directly. If you want to see how plant mass, power, and production rate interact, the interactive model is the fastest way in.

ISRU stopped being theoretical in 2021. MOXIE, a toaster-sized unit aboard NASA's Perseverance rover, produced oxygen from Martian carbon dioxide across repeated runs until its campaign concluded in 2023. The quantities were small by design. The point was to prove that a machine on another world can manufacture a consumable, and it did.

Lunar evidence has been accumulating for longer. Lunar Prospector detected excess hydrogen at the poles in the late 1990s; the LCROSS impactor kicked up a plume from Cabeus crater in 2009 that contained water; the Lunar Reconnaissance Orbiter has been mapping surface temperature and hydrogen distribution ever since. Together they establish that the resource exists, without settling how concentrated or how accessible it is.

The near-term programme is prospecting hardware. Drills and mass spectrometers delivered under commercial lunar payload contracts are designed to sample volatiles directly from the subsurface at polar sites. Those results will decide siting for everything that follows, because you cannot design a plant until you know the ore grade.

The gap between prospecting and production is where the real work sits, and it is the gap Explural builds into: autonomous extraction systems that keep running long after the last crew has left.