Why it is different
The Tonnage Is Easy; the Kilograms Are Hard
Off Earth, industry meets average rock instead of ore, and the scarce elements are the ones used in the smallest quantities.
Terrestrial industry runs on ore: rock in which water, heat and plate tectonics concentrated an element by orders of magnitude. Most other worlds never had that processing. Measured against ore grade instead of average crust, the solar system splits cleanly. The structural elements - oxygen, silicon, iron, aluminum, magnesium, calcium, titanium - are common almost everywhere. The elements used in small amounts for chemistry - copper, lithium, fluorine, chlorine, platinum-group metals, uranium, helium - can sit hundreds or thousands of times below terrestrial ore grade. The Moon is abundant in structure and starved of chemistry. Modeled
| Class | Elements | Treatment | Rated |
|---|---|---|---|
| Abundant | Oxygen, silicon, iron, aluminum, magnesium, calcium, titanium | Assume local extraction on rocky bodies | Modeled |
| Recoverable at a penalty | Water, carbon and nitrogen where present, chlorine on Mars, nickel and cobalt in metal-rich asteroids | Local, but carry the beneficiation, energy and plant cost | Modeled |
| Import | Copper on the Moon, lithium, fluorine, platinum-group metals, uranium, helium, specialty alloying elements | Ship until a substitution or a better recovery process changes the answer | Modeled |
Risk runs opposite to mass. Water and shielding dominate the manifest and have several routes to closure. Catalysts, greases, seals, specialty gases, halogens, pharmaceuticals and fissile fuel may total a few hundred kilograms and stop a plant when the last container empties. A pump is hardware, covered by a spare; helium is inventory, covered by reserve policy and recovery.
Substitution
Change the Process Before Building a Better Mine
Many technologies Earth abandoned were displaced by convenience, not failure; off Earth, supply becomes part of performance.
| Earth default | Off-Earth substitute | Penalty paid | Rated |
|---|---|---|---|
| Copper conductor | Aluminum conductor | Larger cross-section; lighter for equal conductance | Modeled |
| Lithium-ion storage | Sodium-ion for stationary storage | Lower energy density, which a parked battery bank can carry | Modeled |
| PEM electrolysis on iridium | Alkaline electrolysis on nickel | A larger plant, with no platinum-group dependency | Modeled |
| Disposable carbide tools | Resharpenable high-speed steel | Slower cutting; no tungsten import | Modeled |
| Imported cement | Sintered regolith | Works where the environment allows | Modeled |
| Fluoropolymer seals | Metal seals in some services | Higher clamping loads | Modeled |
| Helium pressurization | Autogenous pressurization | Reduces helium demand; cannot remove it everywhere | Modeled |
| Carbon-fiber composite | Basalt-class mineral fiber | About 10 points of structural mass saving | Modeled |
| Platinum-rhodium fiber bushing | Molybdenum, graphite or ceramic bushing | Thin service-life data | Modeled |
The halogen wall
Silicon Underfoot Does Not Make a Chip
Semiconductor fabrication is chlorine and fluorine chemistry from purification to etch.
Lunar chlorine and fluorine run at tens of parts per million, so the Moon can supply the bulk atom of a chip and not the process chemistry. Mars soil carries chlorine at concentrations hundreds of times higher, much of it as perchlorate - a toxicity problem for crews and a feedstock for a chemical plant. If advanced devices are ever made off Earth, Mars has a chemical case the Moon does not. The lunar limit is partly a recovery specification: a line that recycles 99% of its halogen inventory needs a tenth of the makeup that a 90% line does - see Ship or Make. Modeled
Structure
Mineral Fiber Has Metal's Energy and Composite's Strength
The usual comparison is metal against carbon fiber; the third route, fiber drawn from melted rock, is the answer for most of the solar system.
| Route | Process energy | Chemistry needed | Installed power | Shipped plant | Rated |
|---|---|---|---|---|---|
| Metal, pre-reduced nickel-iron | 0.53 kWh/kg | None if pre-reduced; carbothermic reduction at 1,800 K if not | - | - | Modeled |
| Mineral fiber, basalt-class | 0.55 ideal · 0.84 realized kWh/kg | None: one melt, one draw at ~1,700 K | 10-15 kW | 2.5-5 t, refractory locally castable | Modeled |
| Carbon fiber, pitch route | 15-50 kWh/kg | Precursor synthesis, stabilization, 1,500 K inert carbonization; two catalysts on the import list | 40-80 kW | 15-40 t | Modeled |
| Polyethylene matrix | 5-8 kWh/kg | Ethane cracking, polymerization; no reagents | 15-25 kW | ~1.3 t | Modeled |
| Material | Specific strength | Against 304L stainless | Rated |
|---|---|---|---|
| 304L stainless steel | 73 kN·m/kg | 1.0× | Observed |
| 6061-T6 aluminum | 115 kN·m/kg | 1.6× | Observed |
| Ti-6Al-4V titanium | 214 kN·m/kg | 2.9× | Observed |
| Basalt-fiber laminate, quasi-isotropic | ~190 kN·m/kg | 2.6× | Modeled |
| Carbon-fiber laminate, quasi-isotropic | ~375 kN·m/kg | 5.1× | Modeled |
| UHMWPE composite | ~412 kN·m/kg | 5.6× in tension; poor in compression | Modeled |
Basalt-class fiber is drawn from silicate melt at about 1,700 K through a bushing and wound, with no reagent, inert gas or organic precursor. The carbon route does not fail on energy per kilogram - every local route beats imported metal by two to four orders of magnitude - but on plant mass and installed power: a carbonization furnace held at 1,500 K continuously takes 40-80 kW of a 100 kWe site. Carbon fiber is still the ideal import: highest specific strength per shipped kilogram, inert on a spool, no shelf life. Importing tow and making the matrix locally cuts delivered structural mass about 4.2 times.
The matrix
Shelf Life Rules Out Thermosets Past Mars
Thermoset resin, hardener and prepreg expire in one to two years; the spares queue to Callisto is 3.82 years.
Every thermoset consumable arrives out of date on its first delivery to the outer system. Thermoplastic tape has no shelf life and consolidates on demand under heat and pressure. Epoxy also consumes chlorine: 0.208 kg per kilogram of resin at stoichiometry, or 109-146 kg per ton of composite structure in real processes. Among thermoplastics, every high-performance engineering polymer - PEEK, PPS, PVDF, PTFE - is made through a halogen; the polyolefins are the only halogen-free family. Polyethylene suits the outer system better than Earth: its creep is frozen out at 134 K on Callisto and 94 K on Titan, it is the best structural radiation shield per kilogram, and UHMWPE fiber matches carbon fiber in specific tensile strength. Its weaknesses are compression, bolted joints and low-temperature toughness. Modeled
| Matrix family | Callisto | Titan | Europa or Io | Rated |
|---|---|---|---|---|
| Thermoset epoxy | 0.8 | 0.8 | ~0 | Modeled |
| Polyethylene thermoplastic | 0.7 toughness · 1.0 strength | 0.7; 0.4 in warm wetted service | ~0 | Modeled |
| Mineral fiber in a mineral matrix | 0.95 | 0.95 | 0.9 | Modeled |
At Callisto, 30 years of cosmic rays deposit about 12.6 Gy in a polymer, four orders of magnitude below any damage threshold. At Europa a polymer absorbs about 2 million gray a year and fails within 12 months. Mineral fiber is indifferent to both.
The shop
A Composite Bay Turns the Machine Shop Into a Two-Material Facility
Forming and welding thermoplastic covers 64% of failures against 77% for a metal shop; the gap is fasteners and springs.
| Element | Function | Mass | Rated |
|---|---|---|---|
| Single-screw extruder | Stock, tube, profile, gaskets, plain-bearing blanks | ~200 kg | Modeled |
| Heated platen press, 450 K, vacuum bagging | Consolidating thermoplastic tape | ~300 kg | Modeled |
| Induction or resistance welder | Structural repair and tube fusion - the tool that replaces the lathe | ~50 kg | Modeled |
| Filament and tape winder | Pressure vessels, tanks, tubes | ~300 kg | Modeled |
| Oven, 450 K | Annealing and drying | ~200 kg | Modeled |
| Phase 1 total | Forming and repair | ~1.05 t, against 3.09 t for the metal shop | Modeled |
| Bushing furnace and drawing head, 1,700 K | Phase 2: mineral fiber from local silicate | 2.5-5 t | Modeled |
Presses, winders and welders react their own loads through a closed frame, so a composite bay is the one fabrication capability that does not wait for anchoring to be solved - see Ship or Make. Fasteners, springs, bearing journals and cutting tools stay metal, about 100-200 kg of five-year inventory. Molten polymer is a worse fire load than metal swarf, which strengthens the case for running the shop as a separate, robot-operated pressure zone. The metal side is on The Machine Shop MVP.
By destination
Each World Makes a Different Structure
Local chemistry sets what can be made; the destinations specialize.
| Destination | Fiber | Matrix and route | Rated |
|---|---|---|---|
| The Moon | Mineral fiber from regolith | No polymer matrix - carbon is ~150 ppm. Sintered or geopolymer binder, or salvaged hull metal | Modeled |
| Mars | Mineral fiber | Everything: carbon from the atmosphere, chlorine at weight-percent levels, basalt underfoot | Modeled |
| The Belt | Not needed | Pre-reduced nickel-iron at 0.53 kWh/kg; the best site for an off-Earth carbon-fiber plant if one is built | Modeled |
| Callisto | Mineral fiber, if the surface lag draws | Polyethylene from imported carbon or mineral binder; carbon-fiber tow imported for tension parts | Modeled |
| Titan | UHMWPE from ethane | Fiber, matrix and spinning solvent from one molecule; behind a 7.1-year spares queue | Modeled |
| Europa and Io | Mineral only | Polymers are destroyed in under a year at ~2 million gray per year; metal or mineral regime | Modeled |
Gravity vector
Some Processes Need a Down; Others Need None
The surface and orbit are consecutive stations in one production chain, not competitors.
Sheet and ribbon want controlled sag, casting and sedimentation expect gravity, and large melt pools are easier to handle when down exists - those processes favor a surface. Containerless melting, buoyancy-free crystal growth and alloys whose components would otherwise separate favor free fall. The Moon's 2.46-times-longer capillary length should let float-zone silicon boules grow toward 490 mm, beyond the ~200 mm limit on Earth, while orbit offers true microgravity and fast return to Earth. Surface industry supplies feedstock and gravity-dependent steps, orbital industry the steps that need its absence, and the product goes to the customer. Modeled
Flying today
Return Capsules Have Turned Orbit Into a Production Line
Pharmaceutical crystals, optical fiber and semiconductor crystal growth are flying in Western capsules and on the ISS.
| Company or program | Product | Status | Rated |
|---|---|---|---|
| Varda Space Industries | Pharmaceutical crystals; reentry testing | W-1 returned ritonavir crystals to Utah in February 2024; five capsules landed through W-6 in May 2026; W-8 and W-9 launched 1 October 2026; $250M Series D at $1.6B. Revenue about 70% government | Observed |
| Flawless Photonics | ZBLAN optical fiber | Drew 11.9 km on the ISS in early 2024 | Observed |
| Redwire SpaceMD | Drug crystallization | 54 PIL-BOX units flown since November 2023, 45 compounds; 32 units booked on Starfall for 2028 | Observed |
| SpaceX Starfall | Return capsule, up to 1,000 kg | First flight recovered in the Pacific 23 June 2026 | Observed |
| Space Forge ForgeStar-1 | Compound-semiconductor crystal growth | Furnace reached 1,000 °C and generated plasma in orbit, December 2025 | Observed |
| ESA metal 3D printer · Airbus | Stainless parts | First metal part printed on the ISS August 2024; parts returned for testing | Observed |
| ATMOS, Inversion, Outpost | Return capsules | Phoenix 2 early 2027; Arc orbital flight 2026; Outpost service launched September 2026 | Building |
China
China Has Drawn Fiber From Lunar-Soil Simulant and Flown It to Tiangong
China's lunar program is testing the structural rung - bricks and fiber from regolith - before any Western mission.
| Program · organization | Product | Status | Rated |
|---|---|---|---|
| Lunar-simulant fiber · Donghua University | Fiber drawn from lunar-soil simulant | Flown to Tiangong on Tianzhou-10, 13 May 2026, for long-duration exposure; the team also drew fiber from Chang'e-5 soil | Observed |
| Containerless materials cabinet · Tiangong | Refractory alloys | Heated tungsten alloy above 3,100 °C, reported August 2025 | Observed |
| Regolith-simulant bricks · Huazhong University | Construction material | 74 bricks on Tiangong; first batch returned November 2025 | Observed |
| Continuous-fiber composite printer · CAST | Carbon-fiber composite parts | Printed two parts in orbit on the Long March 5B test flight, May 2020 - a world first | Observed |
| Lihong-1 · CAS Space | Suborbital microgravity capsule | First Chinese metal 3D printing in space, 12 January 2026; orbital reusable version planned | Observed |
| Reusable experimental spacecraft | Undisclosed | Fourth mission launched February 2026; no manufacturing role disclosed | Building |
Contested
Deliberately Not Claimed
The load-bearing uncertainties, tagged with the confidence they actually earn.
Fiber drawing from asteroid-like silicate is untested. Terrestrial basalt fiber comes from 45-55% silica melts; carbonaceous material runs 33-35%. A crucible test on meteorite-analog compositions would settle it cheaply.
Callisto's structural loop rests on one measurement. Whether the surface lag deposit falls inside the fiber-drawing window decides whether structure closes at the anchor. The lander that measures ice grade can measure it.
Orbital manufacturing revenue is mostly government. Varda reports about 70% government revenue, and no drug made in orbit has yet reached a patient.
Cold weld allowables are unmeasured. Thermoplastic composite welding is mature at room temperature; nothing has been measured at 134 K.
Record
Provenance and Data Sources
- Confidence
- Observed for flown, fielded or measured - Building for hardware and programs in development - Projected for announced plans - Modeled for original analysis
- Status date
- Q4 2026
- Flight record
- Varda, Space Forge, Redwire, Flawless Photonics, SpaceX Starfall, ESA and Airbus announcements; CAS, CAST, Donghua University and Tiangong reports
- Weakest link
- Fiber drawing from non-basaltic silicate; cold thermoplastic weld properties; Callisto lag composition