3.3 million tonnes of lunar water, lifted to lunar orbit in 45-tonne cartridges by a fleet that imports nothing from Earth. The first off-Earth water economy begins as a radiation shield.
Aegis Station is a rotating habitat in a 100 km lunar orbit, built for people who stay for years. Radiation is the reason it is shielded with water rather than metal, and the amount of water is set by geometry and the 1g rotation baseline, not by what is convenient to deliver.
Lifting that mass from Earth is not a plan; it is a number that ends the program. Every tonne of shield water is therefore mined at the lunar south pole and flown up from the surface. Earth-sourced water is excluded in every phase, including commissioning.
Shield thickness is the multiplier on every number on this page: tankers, sorties, power, and years. The dose case for 3 m is on the water-shield evaluation page.
A 10 m × 2.5 m structural cylinder holding a flexible bladder of ~49 m³, filled by vacuum draw at the ISRU node and emptied by inert-gas displacement at the destination. Cargo the tanker carries, not part of it, with one water port and one gas port that are identical at the node, the pad, the depot, and the station. Water is never a cryogen anywhere in the chain. See the cartridge standard.
A single tanker at low cadence lifts the first cartridges to the Lunar Orbital Propellant Depot and commissions end-to-end propellant production. The depot is the fleet's first customer, and a permanent one. Full fleet deployment is not required to start.
Sustained surface-to-orbit delivery to the station's shield reservoirs. This is the requirement that sizes the 45-vehicle fleet. Rings receive water from the start of assembly; fill sequencing preserves mass balance, and spin-up to 1g follows verification of the complete ring.
Depot resupply, surface distribution between ISRU and LUNET nodes, and emergency water to crewed outposts within hours of dispatch. The extraction and delivery chain does not retire when the shield is full; the water economy starts with Aegis, it does not end there.
The Moon has water ice. Electrolysis splits it into liquid oxygen and liquid hydrogen, the two highest-performing chemical propellants there are and the only ones the Moon can make from a single feedstock without importing anything. The tanker burns both from its first flight. There is no methane phase and no import to taper off later.
An earlier revision baselined methalox with methane shipped from Earth. At fleet cadence that was hundreds of tonnes a day landed on the Moon, a supply chain larger than every lunar landing to date, repeated indefinitely. Anything a program has to keep buying from Earth is the thing that ends it, so the fuel that needed an Earth supply chain is gone.
The hydrogen objection is answered the way LUNET answers it: produce near, consume fast. A tanker fills from a liquefier a few metres from the pad and is empty in orbit within hours. The honest cost moves from the launch manifest to the power plant: the tanker pad is a refinery-class facility, not a LUNET fill node, and its size sets the fill rate. The full argument, the vehicle specs, and the refuelling trade are on the Tanker Fleet page and in the dossier below.
Ice-bearing regolith is excavated from permanently shadowed craters near the south pole and heated by microwave or radiant thermal rigs. The ice sublimes, is cold-trapped, condensed, and filtered to a single water grade that serves both cargo and propellant. An ambient surface buffer absorbs production swings and feeds two consumers: cartridge fill for the tanker, and pad-side electrolysis and liquefaction for the tanker's own propellant.
Pads with dust control and blast deflection, a ready apron so no tanker holds a pad while it waits, cartridge fill and staging, and ground inventory of displaced inert gas. Six pads hold the 4.5-year fill only at a 3.2-hour turnaround; eight are recommended for margin.
Electrolysis, LOX and LH₂ liquefaction, and a buffered store of about half a day's propellant. Tankers fill on demand through ground couplings treated as scheduled consumables. This is the plant that fuels the fleet and the plant the depot and the surface network already assume exists.
LUNET nodes for surface mobility, refuelling, and inter-node exchange. Tankers hop between them at 160–940 m/s to redistribute water and LOX, up to six hops per orbital propellant load.
One console for 45 tankers: dispatch by state of charge, idle time, position and wear; six pad conveyors and six station berths with FIFO queues; standby reserve; anomaly response that reassigns a lost sortie and frees its pad. See the Fleet Commander.
Cartridge reception, electrolysis, LOX and LH₂ storage, and top-off for visiting vehicles. Emplaced before the resource flows and commissioned by the first cartridge to arrive. Methane is imported for the Hauler's return burn only and never touches the tanker fleet. See the depot.
Cartridge docking ports and water transfer lines to the shield reservoirs, autonomous transfer monitoring, inventory reconciliation, and fault management. Rings are filled and balanced before they rotate.
$150/kg is a program target, not a bottom-up estimate. The engine pipeline for the fleet, several thousand printed engines over the campaign, is one to two percent of the delivery figure at serial-production prices and is stated in the dossier.
The propellant decision and the refuelling method are fixed. The campaign numbers are stated as a baseline and these are the things they rest on, carried as open rather than hidden. The full list, with the scope assumptions behind the fleet analysis, is Section 11 of the dossier.
One flight per tanker per day for the whole campaign, about 1,600 cycles per airframe. A requirement, not an assumption: the fill time is derived from it. An airliner cadence in an environment gentler than any reusable booster faces, and one no reusable orbital-class vehicle has yet flown. That is what the fleet has to demonstrate, and the engines, pads, and refinery below are what it demands.
About 18 MW of continuous power per daily sortie, for electrolysis and hydrogen liquefaction. At fleet cadence that is a refinery of order 800 MW continuous on the lunar surface. It is the largest single item the fill depends on, and it is not yet sized.
Each engine position sees ~500 hours of burn over an airframe life, so engines are scheduled replacement items: about 3,000–6,000 over the campaign, two to four a day from one printed part number. $3–10 B at serial prices; engine life is the lever.
One launch every 32 minutes. Six pads and six berths hold the fill only if each turnaround is 3.2 hours or less. Refuelling by fluid couplings at the surface, with cartridge swap and transfer overlapped, targets 2.6 hours. Whether that is achievable with the hardware as designed is open.
Shield water plus propellant water is several million tonnes against polar ice estimates that span two orders of magnitude. Deposit assessment is the gate, and it is why the program is waiting on a viable moonbase and industrial-scale lunar water rather than building.