The Mars Industrial Regime

The organizing claim

Closure Is an Output of the Power Buildout, Not an Input to It

What Mars can make in a given year is decided by how much reactor has landed, and by nothing else.

Sort everything a Martian base might manufacture by the specific energy of the transformation, and the build order falls out without anyone choosing it. Moving regolith costs a thousandth of a kilowatt-hour per kilogram. Making a semiconductor costs between a thousand and a hundred thousand. That is seven orders of magnitude across one site, and each rung waits for the power plant to grow roughly tenfold before it becomes affordable.

This runs the opposite way from the usual account, where industrial capability is described as a technology roadmap that unlocks over time. It is not a technology question. Every process on the ladder is understood, and several have flown. The schedule is a power schedule, which makes Mars Power the upstream page for this one and makes the reactor decision the single largest determinant of what the settlement can do in its first decade.

Two consequences worth holding onto before the detail. A base does not become self-sufficient by acquiring capabilities in the order of their importance — it acquires them in the order of their energy cost, which is a different order. And the capabilities that matter most for survival, the ones that keep hardware running, sit at the top of the ladder rather than the bottom.

The ladder

The Closure Ladder, Sorted by Specific Energy

Seven orders of magnitude from the bottom rung to the top.

Two entries deserve their own treatment because they break the pattern in opposite directions. Salvage sits off the curve entirely, at a ratio so favorable that it changes what the early base is. And metals reduction, which the popular account treats as the foundation of off-Earth industry, turns out to be an export business that a young settlement should not attempt.

Arrives complete

The Site Arrives Complete and Then Decays

Closure is replacement, not construction — which is a different problem with different answers.

A minimum viable outpost is roughly 230 t of hardware, and a cargo vehicle lands about 100 t. The entire site therefore arrives in two or three landings, complete. Nothing has to be built up from raw material to exist; the base is fully formed on day one and begins degrading immediately.

What it loses splits into two streams that behave nothing alike. Hardware fails: at roughly 240 failures a year averaging about 19 kg each, the site sheds around 4.6 t of components annually, which is about 2% of itself. Consumables deplete outright at roughly 12 t a year. The upper rungs of the ladder serve the failing stream and the lower rungs serve the depleting one — and only the first has urgency, because a stockpile can cover consumables while nothing covers a part type that has no spare.

That reframes what industrial capability is for. A Martian base is not a construction project that becomes self-sufficient by building more of itself. It is a maintenance operation whose survival depends on replacing about 4.6 t a year of specific, awkward, high-value components — and those components sit at the expensive end of the ladder, which is the tension the rest of this page works through.

Salvage

Salvage Sits Off the Ladder, at 300 to 1

Earth already paid the beneficiation cost. The launch vehicle is the delivery mechanism.

Liberating material from a landed hull costs about 0.04 kWh/kg and yields material carrying 10–15 kWh/kg of embodied energy — a return of roughly 300 to 1 that no other rung approaches. The stream also self-scales: scrap arrives with traffic, and traffic grows with the site.

The usual framing counts the steel, and the steel is the least interesting part. A hull holds roughly 66 t of stainless against a site that consumes about 400 kg of structural metal a year — 165 years of supply, which is not a constraint anybody needs to solve. What binds is components, and a landed vehicle carries about 9 t of exactly the classes the shop cannot make: engines, fluid systems, valves, avionics, harness, motors, bearings. Salvage does not close the import rungs; it defers them by 3 to 17 years per hull, depending on how much of the vehicle was designed to come apart.

Two Mars-specific advantages apply, both covered in more detail on Martian Regolith, Halogens & Materials. A vehicle that lands here arrives with ceramic thermal protection tiles — furnace refractory and cryogenic insulation — and with four aerodynamic flaps carrying high-power electric actuators. A stripped lunar lander carries neither.

And the first tranche of salvage is free, because it is a byproduct of outfitting rather than a separate activity: the tank internals, baffles, downcomers and feed lines have to come out anyway before a hull becomes habitable. What that costs is crew time, not energy — on the order of 500 to 1,000 crew-hours per hull for clearing alone, which is half a tour to a full one and which the 0.04 kWh/kg figure does not capture at all.

The bottom rung rests on somebody else’s business decision. If cargo vehicles become reliably reusable and fly home, rung zero disappears and takes the metal supply, the pad material and the pressure volume with it. That is the largest external dependency in the entire closure argument, and it is a vehicle-company economics call rather than a Mars program decision.

Earthmoving

Every Load-Bearing Dependency Runs Out of Earthmoving

One root, and Mars is the only destination in the series that clears it.

Trace the prerequisites of the industrial ladder and they converge on a single capability. Habitable volume needs a trench, because a standing hull cannot be shielded and a horizontal one needs 2,900–4,200 m³ of material moved. The landing pad needs grading. Below about 1.67 m/s² every machine tool’s poured foundation needs excavation and anchoring. Three separate chains, one machine at the bottom of all of them.

That threshold is where Mars separates from everything else. A 20 t machine developing 20 kN of breakout force against a friction coefficient of 0.6 needs 1.67 m/s² to react its own working loads through weight alone. The Moon sits at 1.62, just under. Mars at 3.71 clears it with margin, so terrestrially proportioned equipment works without ballast, anchors or reaction mass. Callisto, Titan, Ceres and the small bodies do not clear it at all.

The consequence for the ladder is that on Mars the graph is mostly flat. With the root cleared, the rungs are reorderable by paying for them — closure sequencing becomes a budgeting problem rather than a critical path. That is a materially different situation from the lunar case, where the same root is a genuine gate, and it is arguably the strongest single industrial advantage Mars holds.

Low-gravity excavation has no flight heritage anywhere and is nobody’s program. Mars needs it least of all the destinations, which is precisely why nobody has been forced to solve it, and why every architecture past Mars inherits an unsolved problem.

The shop

The Machine Shop, the Spares Seam, and Why Commonality Beats Both

A 3 t shop addresses 69% of what breaks — and the mass argument for it does not close.

A minimum machine shop — lathe, mill, welder, plasma cutter, sintering head, measurement — runs about 3.09 t and can address roughly 69% of the failure population by part count. The obvious justification is mass: make parts instead of shipping them. That argument was tested and withdrawn, because the shop plus its power chain outweighs the spares it displaces over any realistic horizon.

Three justifications survive, and they are better ones. Availability: a site with a shop loses roughly 3.2 times fewer function-months to waiting for parts. Latency: without a shop, a broken part type means waiting for the next window, so the iteration loop runs at 790 days; with one it runs at a day. Self-expansion: once a base can cast locally, a new machine ships as about 17% of its terrestrial mass, with only the precision surfaces and the components arriving from Earth.

Underneath sits the seam that organizes the whole ship-or-make question. Every rung below spares passes a mass test — salvage, oxygen, water, forming, cast bases all return more than they cost. Spares fails the mass test everywhere. Below the seam, closure pays in mass; above it, closure pays in schedule. And the two curves run in opposite directions with distance, crossing between the Moon and Mars — which is exactly why Mars makes things the Moon would ship, and the reason is the 25.6-month window rather than the distance.

The cheapest intervention is not a machine at all. Reducing a site from 400 distinct part types to 120 removes about 70% of the spares inventory at constant risk, roughly 18 t at outpost scale. It costs paperwork — a common fastener set, a common bus voltage, common bearing and seal sizes, a shared connector standard — and it has to be agreed before the supplier base commits, after which it cannot be had at any price.

Dispatch

The Rungs Do Not Sum. They Time-Share.

One bus, several plants, and not enough power to run them at once.

Adding up the unlock costs of the ladder gives 76–102 t of energy-chain mass against an energy layer of about 55 t in a minimum outpost. The arithmetic does not close, and the reason it does not matter is that the rungs never run simultaneously.

On a 150 kWe bus, oxygen production at 100 t/yr draws about 80 kW and water production at the same rate draws 57 kW at good grade or 114 kW at 1%. At low grade the pair does not fit at all; at high grade there are 13 kW to spare, which is less than the machine shop’s installed load. Closure sequencing is therefore a dispatch problem, and the whole ladder is interruptible except whichever rung is in production at the moment.

That gives the minimum viable outpost a clean definition worth stating plainly: the largest capability set that fits on one bus at once. Not the largest set of capabilities landed, and not the set anyone would like to have — the set that can be powered simultaneously. Everything beyond it waits for the next reactor.

What stays imported

The Permanent Import List Is Short, Light and Unavoidable

Never the bulk materials. Always the trace ones, plus the components nobody can make.

Earth’s ore bodies exist because water, heat and tectonics spent billions of years concentrating a well-mixed crust. Mars ran those processes weakly and stopped, so most elements sit near cosmic abundance and the scarce ones stay scarce. The general rule across this series is that substitution beats beneficiation — redesign the process around abundant elements rather than building a concentration plant.

Two substitutions outside that table are free wins and should be taken early. Aluminum replaces copper at 2.1 times better conductance per unit mass, and aluminum is 9 wt% of the regolith against copper at parts per million. Sodium-ion replaces lithium in stationary storage, which matters because lithium sits far below ore grade everywhere off Earth — and it moves the battery-versus-fuel-cell crossover, making the storage decision partly a sourcing decision rather than a purely physical one.

The whole permanent list weighs a few hundred kilograms a year at outpost scale. It is not a mass problem. It is a dependency problem, and it is the reason a Mars settlement of any plausible size in this century would survive the loss of Earth’s markets and not the loss of Earth’s industrial base.

What would overturn this

The Findings That Would Change the Industrial Picture

Two of these are answerable on Earth, cheaply, today.

Record

Provenance and Data Sources