The Moon May Support Small Settlements—Not a Million-Person City

The Moon may have enough energy to run a substantial economy, but its water inventory appears to set a much tighter ceiling on permanent settlement. In a 2026 analysis, Martin Elvis and Jonathan McDowell concluded that lunar polar ice could support small, highly recycled communities for centuries, yet is insufficient for a self-sustaining city of one million people over the long term under today’s best demonstrated recycling performance. 6

The key assumption: one billion tonnes of water

The study uses roughly 1 billion tonnes of water in permanently shadowed polar regions as a deliberately generous baseline. That figure is not a confirmed, readily recoverable reserve; it is an upper-end working assumption for testing what large-scale settlement would require. 6

The authors estimate annual water demand at roughly:

• 125 tonnes per person for direct personal needs
• About 500 tonnes per person for food production

That makes agriculture the dominant part of the water challenge. Water used in a closed habitat can be recycled, but every imperfect step in a settlement’s life-support and food system creates losses that must be replaced from the Moon’s finite ice stock. 6

What settlement size can lunar water support?

The outcome changes sharply with population size and recycling efficiency.

Settlement scale	Result under the study’s assumptions
Around 1,000 people	A village could be sustained for centuries with 98% recycling. 6
Up to around 10,000 people	Town-scale settlements could also remain viable over centuries with ISS-level recycling. 6
1 million people	A city would not be sustainable long term from the assumed indigenous water inventory at 98% recycling. 6

For a million residents, using the full water demand without recycling would consume the 1-billion-tonne baseline in roughly 2.4 years. With 98% recycling—comparable to the International Space Station—the estimate improves dramatically, but only to a little over a century. 6

That is a major improvement in operational terms, but it is not the multi-century or open-ended lifetime usually implied by a self-sustaining city.

Why the real situation could be worse

The one-billion-tonne figure is already optimistic. The paper notes that likely accessible water could be substantially lower—potentially around 30 times below that generous estimate. 56

The uncertainty matters because the settlement calculation is fundamentally a stock-and-loss problem: less recoverable ice means the same population reaches depletion much sooner. A million-person settlement that is marginal even under an optimistic inventory becomes far less plausible if accessible ice is limited.

Why 99.9% recycling changes the calculation

The authors find that raising effective recycling from 98% to 99.9% could extend the million-person scenario to about 2,000 years. 6

The difference sounds small in percentage points, but it represents roughly a tenfold reduction in water losses. Achieving that standard across an entire lunar settlement—including food production rather than only drinking water and hygiene—would be a demanding engineering and operational challenge.

The study’s central conclusion follows from that sensitivity: a large, long-lived lunar population needs either recycling about ten times more effective than the ISS-level case, a comparably larger recoverable water reserve, or a continuing external supply. 6

What could relax the water limit?

Several proposed technologies could improve the outlook, but they address different parts of the problem.

Solar towers can help with energy, not supply

Solar towers could provide near-continuous power to polar regions and shadowed areas, supporting ice extraction, processing and industry. This is why the authors do not treat power as the primary population constraint. But electricity cannot replace water lost from a closed system: it can make extraction and recycling feasible, not create a new local water inventory. 6

Better exploration could find more recoverable ice

Deeper drilling, improved surveys and seismic methods could identify buried ice deposits that remote observations have not fully characterized. If much larger deposits exist and can be extracted, the supply side of the calculation improves. Until those reserves are verified, however, they remain a possibility rather than a solution. 6

Asteroid mining would turn a local settlement into an import-dependent one

Water-rich asteroids could provide an off-Moon source of volatiles. That would change the premise from living on a fixed indigenous inventory to maintaining a settlement through imports. It could support a larger population, but only with a reliable space-mining, transport and delivery system operating at major scale. 6

More efficient food production would cut the largest demand

Because food production accounts for about 500 tonnes of water per person per year in the study’s illustrative budget, better crops, farming systems, dietary choices and food-production technologies could have an outsized effect. Lower agricultural water demand directly reduces the replacement water a settlement must obtain. 6

The practical takeaway for lunar settlement plans

The paper is not an argument that people cannot live on the Moon. Its result is more specific: small settlements may be feasible for centuries with strong recycling, while a million-person, self-sustaining lunar city is not supported by the assumed polar-water inventory at 98% recycling. 6

For lunar planners, the priority is therefore not simply generating more power. It is proving recoverable ice reserves, minimizing water losses throughout life support and agriculture, and determining whether external supply chains can be made dependable enough to close the remaining gap. Until then, lunar development looks more compatible with villages and towns than with permanent metropolises. 6