Living off Mars, literally.

Every kilogram you put on the surface of Mars costs a fortune and a launch window. The two planets only line up for an economical crossing about once every 26 months, and even with the largest rockets ever built, the mass you can land is a thin trickle compared with what a living settlement burns through. A single person needs roughly their own body weight in water, oxygen and food every few weeks. Ship all of it from Earth and the colony is a permanently-resupplied outpost, hostage to the next window. It never becomes a home.

So the entire plan for settling Mars rests on a single idea with an ungainly name: in-situ resource utilization, or ISRU — making what you need on site, out of the local dirt, ice and air, instead of hauling it across the solar system. ISRU is not a luxury or a later phase. It is the difference between visiting Mars and staying there. The first crewed missions are designed around it from the first hour, because the same buried ice that gives you drinking water also gives you the rocket fuel to come home.

What makes ISRU tractable is that Mars, for all its hostility, is unusually well stocked. The air is a ready supply of carbon and oxygen. The ground holds enormous quantities of frozen water. The regolith — the loose, rusty soil that covers the planet — is dense with iron, silicon, aluminium and the oxygen bound up inside its minerals. Mars is not a barren rock; it is an unprocessed warehouse. The work is all in the processing, and the processing runs in a strict order.

Power is the foundation of the stack, so it is the first thing to arrive. There are three real ways to make electricity on Mars, and a serious settlement will use all three for different jobs.

Fission: the workhorse

A small nuclear reactor is the front-runner for primary base power, for one decisive reason: it runs day and night, through dust storms and the long Martian winter, for years on end, in a box you can land in one piece. NASA has been building toward exactly this. The Kilopower programme's ground test — KRUSTY, in 2018 — proved the reactor physics for a compact 1-kilowatt-class flight unit. The follow-on Fission Surface Power project scales that up to a ~40-kilowatt class system (with newer designs pushing toward ~100 kW), engineered to operate for at least a decade and intended to be flight-ready around 2030. A handful of those units, landed and switched on, is enough to run a first base — habitats, the water plant, the fuel plant, the refinery.

Getting a reactor to Mars is a logistics and safety problem more than a technical one. It launches cold and unfuelled or freshly fuelled but inert — a reactor is only dangerously radioactive once it has been running — so a launch failure does not scatter a hot core. It rides an uncrewed cargo ship on an early window, lands, and is set up by robots or by the first crew, deployed a safe distance from the habitat with the ground itself used as shielding.

Solar: cheap, abundant, and treacherous

Mars gets about 43% of the sunlight Earth does — a mean of roughly 590 watts per square metre at the top of its atmosphere, against Earth's ~1,361. That is plenty to run on; solar has powered nearly every Mars lander and rover. But sunlight on Mars carries two traps. The first is the night: with a 24-hour-37-minute day, you spend half of it in darkness and cold, so solar always needs heavy batteries to carry the load through to morning. The second is dust. A global dust storm can blanket the sky for weeks to months and cut the power reaching the surface by more than 90%. That is not theoretical — a 2018 planet-wide storm is what finally killed the Opportunity rover, and accumulating dust on its panels is what ended the InSight lander. Solar is excellent for spreading power across a wide settlement, but you cannot bet a crew's survival on a clear sky.

Radioisotope units: small, eternal, limited

The third option is the one already working on Mars right now. A radioisotope thermoelectric generator (RTG) makes electricity from the heat of decaying plutonium-238 — no sun, no moving parts, no refuelling. The MMRTG aboard Curiosity and Perseverance produces about 110 watts from roughly 4.8 kg of plutonium dioxide, and keeps doing it for well over a decade. The catch is in that number: 110 watts runs a rover, not a town. Plutonium-238 is also rare and extraordinarily expensive to produce. RTGs are perfect for keeping a critical system alive anywhere, forever — and useless as the main power supply for anything large.

Water is the second layer of the stack and, in a real sense, the keystone of the whole settlement: it is drinking water, it is the air you breathe once you split it, it is the fuel that flies you home, and it is the working fluid of nearly every industrial process. The single most important fact about Martian real estate is that the water is already there — you do not have to bring it, you have to reach it.

It comes in three forms, and a good claim site has at least one within reach:

How you get it out

The most elegant method for thick buried ice borrows a trick from Antarctic research stations: the Rodwell, or Rodriguez well. Instead of digging the ice out, you drill a hole down to it and melt it in place with a heated probe, letting a pool of liquid water collect in a cavity underground, then simply pump it to the surface. No excavation, no hauling tonnes of frozen dirt — the ice never leaves the ground until it is already water. NASA has funded exactly this concept for Mars under the name RedWater. For shallower or dirtier ice, the alternative is to excavate icy regolith and bake it in a sealed oven, capturing the vapour as it sublimates; and for the dry equatorial sites with no ice at all, you heat the hydrated soil to release its bound water. Every one of these is, fundamentally, the power plant from Part Two doing work as heat.

Getting the water plant there

This is the part that reorders a Mars mission. The water-and-fuel plant does not wait for the astronauts — it flies ahead of them, uncrewed, on an earlier launch window. Robots land it, deploy the reactor and the drill, and the machine spends a year or more quietly making water and stockpiling propellant, so that when the first crew finally touches down they arrive to a full tank and a fuelled return ship, not an empty drill and a prayer. No serious crewed Mars architecture sends people first; the resources are pre-positioned, because a crew that lands to find the water plant broken is a crew that does not come home.

Here is why water is the keystone of the entire stack: once you have liquid water and electricity, two simple, century-old chemical reactions hand you breathable air and the rocket fuel to leave — out of nothing but water and the Martian sky.

The first is electrolysis: run a current through water and it splits into its two elements.

The second is the Sabatier reaction, which takes that hydrogen and combines it with carbon dioxide pulled straight from the atmosphere — and remember, the Martian air is about 95% CO₂, an effectively unlimited supply.

Read those two together and the magic becomes clear. Water and Martian air, driven by the reactor, yield oxygen to breathe, methane to burn, and recovered water that recycles into the loop. Methane and oxygen — "methalox" — happen to be exactly the propellant that SpaceX's Starship runs on, which is no accident: the vehicle was designed around the fact that its return fuel can be manufactured on Mars. This is why a one-way fuel budget is enough to reach the surface. The ride home is made there.

And the air half of this has already been demonstrated on Mars, not just on paper. MOXIE, a toaster-sized instrument aboard Perseverance, pulled oxygen directly out of the Martian atmosphere — producing up to 12 grams an hour and about 122 grams in total across 16 runs before it was retired in 2023. It is a small number on purpose; it was a proof of concept. But it was the first time human beings manufactured a usable resource on the surface of another world, and it worked exactly as the chemistry promised.

Mars has a serious radiation problem, and it comes from two things it lacks. It has no global magnetic field to deflect charged particles — only weak, patchy remnants frozen in the crust — and its atmosphere is barely 1% as thick as Earth's. Together, the natural shield over your head on Mars is on the order of one-fortieth of what protects you at sea level on Earth.

Two kinds of radiation get through. Galactic cosmic rays are a constant, faint drizzle of extremely high-energy particles from across the galaxy — always on, coming from every direction, and very hard to stop. Solar particle events are sudden storms of particles flung out by the Sun — occasional, sometimes intense enough to be acutely dangerous, but far easier to block. Curiosity's onboard detector, RAD, has actually measured the result on the ground: about 0.7 millisieverts per day on the Martian surface — well over 200 mSv a year, dozens of times the natural rate on Earth. The journey out is worse, around 1.8 mSv/day in deep space, so a full round-trip mission accumulates on the order of one sievert — enough to meaningfully raise lifetime cancer risk. This is a problem you manage, not one you ignore.

The shield is local

The decisive insight is that you never launch radiation shielding from Earth — it is far too heavy. You make it on Mars, out of the cheapest material on the planet: the ground itself. The standard plan is to bury the habitat, berm it over with regolith, or stack sandbags of Martian soil on the roof — about a metre makes a real difference, and a long-term settlement wants several. Water and ice make excellent shielding too, and there is a subtle reason to prefer them: hydrogen-rich materials stop radiation more effectively per kilogram than metal, and they produce fewer of the dangerous secondary particles that get knocked loose when cosmic rays slam into heavy atoms. Best of all is to skip construction entirely and move in underground — Mars is riddled with lava tubes, natural caverns under tens of metres of rock that shield as well as any engineered bunker, for free.

"Screening" the radiation is really three disciplines working together:

• Monitoring — every settler carries a dosimeter and the base tracks cumulative exposure, exactly as a nuclear or medical worker does on Earth.
• The storm shelter — a small, very heavily shielded room (often the most ice- and regolith-wrapped corner of the base) where the crew rides out a solar particle event, which can flare up with only hours of warning.
• Site selection — choosing ground that is naturally protected. Low-lying basins sit beneath more atmosphere and so take a measurably lower dose; Hellas Planitia, the deepest basin on Mars, carries roughly double the planet's average surface pressure, and that extra air is extra shielding. Where you stake your ground is itself part of the radiation plan.

A settlement that can make power, water, air and fuel still has to build things — structures, tools, spare parts, more solar panels — and it cannot fly every bolt from Earth forever. The last layer of the stack is turning Martian dirt into metal and construction material. The encouraging news is that Mars is, geologically, an ore body lying in the open.

Martian regolith is iron-rich — the ferric oxide in it, ordinary rust, is literally why the planet is red — and alongside the iron it carries silicon, aluminium, magnesium, calcium, sulfur and titanium, almost all of it locked up as oxides. The striking part, once you look at the composition by weight, is how much of the dirt is simply oxygen, chemically bound inside those minerals. The ground is nearly half oxygen by mass — which means the regolith is not only a source of metal but, indirectly, another source of air.

First, the catch: perchlorates

Before Martian soil is useful — or safe — it has to be cleaned of perchlorate salts, which make up roughly 0.5 to 1% of the regolith and are toxic to the human thyroid. They have to be washed or baked out before the dirt can be handled, farmed in, or sintered. The redeeming twist is that perchlorates are themselves a resource — heated, they break down and release oxygen and chlorine — so the step that detoxifies the soil also extracts useful chemicals. A nuisance and a feedstock in the same handful of dirt.

Refining: one process, metal and oxygen

The leading idea for getting metal out of regolith is molten regolith electrolysis: melt the soil to a glowing liquid and pass an electric current through it. The metal ions migrate to one electrode and plate out as pure metal — iron first, then silicon, then aluminium — while oxygen is liberated at the other electrode as a byproduct. It is the same elegance as the water-and-fuel loop: a single, power-hungry process that yields both the structural metal to build with and more breathable air. (This is also where Part Two comes back to bite — smelting is enormously energy-intensive, which is precisely why the reactor has to come first.)

Building without cement

Not everything needs to be metal. The first Martian buildings will mostly be made of lightly-processed Mars itself:

• Sintered regolith — heat the dust until its grains fuse into a hard ceramic brick, no binder required. Microwaves or concentrated sunlight can do it.
• Sulfur concrete — melt Martian sulfur and mix it with regolith aggregate; it sets as it cools and, crucially, needs no water — a real advantage on a planet where water is too precious to pour into pavement.
• Geopolymer "Mars concrete" — chemically binding regolith into a cement-like solid, the workhorse for foundations, radiation berms and roads.

Put the whole chain together and a mature settlement looks less like a base and more like a foundry that eats dirt: it mines regolith, refines out metal and oxygen, casts concrete and brick, and grows its own structures — each new launch window adding capability rather than just replacing supplies. That is the moment a colony stops being an outpost and becomes a place that builds itself.

A homestead has never been about naming a piece of ground — it has been about working it. Every frontier framework in our companion deep dives, from the history of land titles to adverse possession, rewards the same thing: the settler who showed up, improved the ground, and put it to productive use. The Homestead Act asked you to build and farm. Mining law asked you to actually work the claim. Brazilian usucapião asks for real, good-faith possession with the intent of an owner. Across two thousand years, the law's sympathy goes to the person who made the land do something.

In-situ resource utilization is exactly what "improving the land" means on Mars. A parcel's worth is not the pixel on the map — it is the power you can raise on it, the ice you can reach beneath it, the basin air that shields it, and the ore you can refine from it. That is why we map resources parcel by parcel instead of selling random dots. A claim that says "I picked a spot" is a novelty. A claim backed by a real resource-and-development memo for that specific ground — ice depth, sun, shielding, ore — is a documented, good-faith, improvement-minded claim, the kind the entire history of homesteading takes seriously.

We have to be just as precise about what this does not mean. Knowing how to work Martian ground does not make that ground yours. No one can hold legal title to Mars today — the 1967 Outer Space Treaty bars any nation from appropriating it, and no court or registry has the jurisdiction to grant or enforce ownership off-world. Spaceclaims conveys no title and guarantees no recognition. What understanding the resources does do is let you stake your parcel on credible ground and document, honestly and early, the good-faith intent to develop it — building the strongest forward claim a person can make while the law waits to catch up.