01 — Why here
Bimodal, and that is the whole point.
Almost every argument for the Moon is made about the Moon in general. It should be made about roughly forty square kilometres of it.
At the lunar south pole the axial tilt is small enough that certain rim points stay in near-continuous sunlight while crater floors a short distance away have been in permanent shadow for on the order of two billion years. That is not a gradient. It is two environments — one of the most reliable solar sites in the inner solar system and one of the coldest surfaces yet measured anywhere — separated by a distance a vehicle can cross in an afternoon.
Nowhere else in the solar system offers that adjacency. Mars has seasons and dust; the belt has neither sunlight worth the name nor a gravity well to hold anything down; Mercury has the illumination but sits behind a Δv wall higher than Saturn's. Shackleton is the only place where the power problem and the volatile problem are solved by geography rather than by logistics.
Everything downstream — the mass driver, the cryogenic compute case, the water export economics, the legal fight over who gets the rim — follows from that one fact of adjacency.
02 — Constraint topology
Envelope
What bounds capability here. None of it changes with time on station — these are the numbers you design against on day one and on day ten thousand.
| Parameter | Value | Consequence | Rated |
|---|---|---|---|
| Insolation, rim | ~1,360 W/m² at grazing incidence; ~90% annual capacity factor at favoured points | Solar is viable without storage sized for a 29.5-day night — the single largest architectural saving on the Moon | Observed |
| Insolation, floor | 0 | Every watt on the floor is imported down the corridor or generated by fission | Observed |
| Thermal sink, floor | ≈40 K | Radiative rejection is nearly free; cryogenic compute and long-term volatile storage become cheap | Observed |
| Surface gravity | 1.62 m/s² · 0.165 g | Excavation reaction forces must come from mass or anchoring, not weight. Escape velocity 2.38 km/s puts electromagnetic launch in range | Observed |
| Light time to Earth | 1.28 s one way | Teleoperation is workable. Compute assists rather than replaces the crew — the last destination where that is true | Observed |
| Rim–floor separation | Order 10–20 km, several km of relief | Sets the corridor: power transmission, cryogenic transfer, and vehicle cycle time all scale off this number | Observed |
| Earth line of sight | Near-permanent from selected rim points; none from the floor | Direct-to-Earth comms from the rim; the floor needs a relay. Earth hangs at the horizon and barely moves | Observed |
03 — Hazard topology
Hazards
What damages, at what rate, with what return period. Every entry here is a rate or a probability — that is what separates this table from the one above, and it is why the two are kept apart.
| Hazard | Magnitude / rate | Defence | Rated |
|---|---|---|---|
| Galactic cosmic rays | Chronic, unshielded surface dose on the order of tenths of a mSv per day; no atmosphere, no field | Regolith overburden. Mass is the only answer and it is locally free | Observed |
| Solar particle events | Episodic, hours of warning, potentially lethal unshielded | Storm shelter — hydrogen-rich mass, which means water or polyethylene rather than regolith | Observed |
| Thermal cycling | Polar amplitude far smaller than equatorial, but the terminator crossing is near-instantaneous | Rate, not amplitude, drives the fatigue spec. Design for the shadow edge, not the average | Building |
| Cryogenic embrittlement | Floor at ≈40 K — outside the qualified range of most terrestrial lubricants, elastomers and seals | Materials selection, dry lubrication, heated mechanism enclosures. The 40 K asset and the 40 K hazard are the same number | Projected |
| Regolith dust | Sharp, sub-100 µm, electrostatically charged, non-weathered | Abrasion of seals, optics and radiators; adhesion to everything. Mitigation is unsolved at scale | Observed |
| Terminator dust levitation | Charge-driven, observed since Surveyor; worst at the lit/dark boundary | The corridor runs through exactly the region where this is worst | Projected |
| Micrometeoroids | Continuous, tens of km/s, no atmosphere to stop anything | Overburden and redundancy; pressure vessels and radiators are the exposed surfaces | Observed |
| Shallow moonquakes | Recorded up to roughly M5.5, with unusually long ring-down — no water in the crust to damp them | Structural spec closer to seismic design than most lunar architectures assume | Observed |
| Partial-gravity physiology | 0.165 g, effects over multi-year exposure genuinely unknown | No countermeasure validated at this level. The largest open question for permanent habitation | Modeled |
Two of these deserve emphasis because they are routinely underweighted. Dust is the mitigation problem nobody has solved at industrial scale — Apollo encountered it over days, not decades, and a mass driver moving tonnage across a dust-laden corridor is a different exposure regime entirely. And 0.165 g is a complete unknown: every human partial-gravity datapoint we hold is either microgravity or one g, with nothing in between.
04 — The site
Shackleton, drawn.
05 — Four readings
The same place, four questions.
Cheap to reach, expensive to leave
5.9 km/s from LEO with no atmosphere to brake against, so arrival is entirely propulsive. Departure is the interesting direction: escape velocity of 2.38 km/s is low enough that electromagnetic launch replaces chemical propellant for bulk cargo, which is what makes export economics close at all.
The first exporter
Not a terminus but a source node. Water lifted to cislunar depots is the Moon's first real export, and it is worth more in orbit than on the surface. The corridor from floor to rim, and the mass driver from rim to orbit, are the same supply chain seen at two scales.
Geography solves it
Power and volatiles are usually a logistics problem traded against each other. Here they are adjacent, which converts a continuous optimisation into a siting decision. That is why the argument is about specific rim points rather than about the Moon.
Dust and an unknown
Radiation is severe but the defence is free — regolith is everywhere and mass is the answer. The unsolved items are dust at industrial scale and the physiology of 0.165 g over years, neither of which yields to more shielding.
06 — Matter
What is actually here.
| Source | Yields | Notes | Rated |
|---|---|---|---|
| Polar water ice | H₂, O₂, consumables, shielding | The keystone. LCROSS measured several weight percent at Cabeus; Shackleton concentrations are inferred and less certain, and grade matters more than presence for any extraction economics | Projected |
| Regolith, bulk | Shielding mass, sintered structure, pads, roads | Free, unlimited, and the dominant tonnage in any build. Cheapest hazard mitigation available anywhere | Observed |
| Regolith, oxides | O₂ (~40–45 wt%), Fe, Ti, Al, Si | Oxygen is the single largest mass fraction and it is all chemically bound — extraction is an energy problem, not a prospecting one | Observed |
| Atmosphere | None | No buffer gas, no carbon, no nitrogen. Everything in that category is imported or arrives with the water | Observed |
| Solar flux, rim | Near-continuous electrical power | Functions as a matter source indirectly: it is what makes oxide reduction affordable | Observed |
07 — Corridors
Where it connects.
| Route | Δv | Transit | Carries |
|---|---|---|---|
| LEO → lunar surface | 5.9 km/s | 3–5 d | Crew, high-value cargo, anything not manufacturable locally |
| Lunar surface → EML-1 | 2.5 km/s | 3–4 d | Water and propellant outbound — the export leg that pays for everything |
| Lunar surface → orbit, mass driver | 2.38 km/s, electromagnetic | minutes | Bulk regolith and processed feedstock, no propellant expended |
| Rim ↔ floor | surface, ~10–20 km | hours | Power down, volatiles up. The corridor that defines the site |
08 — Contested
Deliberately Not Claimed
Helium-3 mining does not close. Lunar regolith carries He-3 at parts-per-billion concentrations, meaning roughly a hundred million tonnes of regolith processed per tonne recovered — and the fusion reactors that would consume it do not exist. The excavation energy alone exceeds any plausible return. This appears here because it is the most persistent bad argument for the Moon, and leaving it unaddressed weakens the good ones.
Ice grade is the load-bearing unknown. Water is confirmed present at the poles. Concentration, depth distribution, and physical form at Shackleton specifically are not, and extraction economics are far more sensitive to grade than to total inventory. Any architecture claiming a production rate is claiming a grade it does not have. Ground truth is the single highest-value measurement available on the Moon right now.
Rim points are a scarce, non-fungible resource, and there is no allocation mechanism. The number of locations with both high illumination and short access to shadowed volatiles is small — plausibly single digits. The Outer Space Treaty forbids sovereign appropriation but says nothing usable about priority of use. This is the first genuine property dispute off Earth and it is a decade away, not a century.
09 — Record
Provenance.
- Confidence
- Building — hardware funded and flying, industrial output not yet demonstrated
- Topography
- LRO / LOLA polar dataset
- Volatiles
- LCROSS impact results; LRO neutron and radiometer data; ShadowCam imaging of the permanently shadowed regions
- Thermal
- Diviner radiometer surface temperature record
- Seismic
- Apollo passive seismic experiment package
- Weakest link
- Ice grade at this specific site, and partial-gravity physiology at 0.165 g