The news is that SpaceX is pivoting to making a “self-growing city on the Moon”.
Short version: it won’t happen because said colony will always require continuous input from Earth of roughly 5-15 tons per person per year of material, using McMurdo Station as an example for the higher range of that. For a city of 100,000 people this would be a prohibitively large amount of material. Roko outlines as much in his article on Twitter:
Elon Musk, our glorious leader and patron has (as I predicted) scaled back his near-term space ambition from Mars to The Moon. That makes sense, Mars is a really long way away. But is it realistic to build an economically self-sufficient human colony on The Moon before 2030? Doing some research on this, I don’t think it will work. Why not? Well, because it’s really, really expensive to get to The Moon even with very optimistic SpaceX figures for launch costs, and because you need quite a large human economy to be truly economically self-sufficient meaning that it doesn’t need to import anything.
Using Starship with in-orbit refuelling it’s possible to send cargo to the Moon but it costs about 10,000 times more per ton than sending cargo to an Island on Earth; it’s about $1,000,000 per ton using SpaceX and about $100 per ton using standard ocean freight.
Surprisingly, the amount of cargo per person per year into a wide variety of human colonies with developed world living standards is roughly a constant: 5-10 tons per person per year. This is true over a 6 order of magnitude range from the ISS up in space to Ireland.
This leads me to believe that a Lunar colony will also need about 5-15 tons of cargo per person per year. I don’t think that such a colony can become truly economically self-sufficient: when you try and make things locally you start demanding more people, more input materials, etc. Water, nitrogen and Oxygen might be recycled but most other things will come by rocket.
With a colony of ~100,000 people that’s basically 1 megaton per year to the lunar surface, forever, costing about $300bn annualized, or three million dollars per person in the colony per year. It’s not just money though. This isn’t a crypto coin or a SAAS business or something. This is $300bn in real costs, including utterly insane levels of environmental damage locally: it would need something like 1 launch every 3 minutes(!) and a fleet of something like 1000-2000 active ships at any one time, with roughly 1800 ships worn out and scrapped per year. They would emit more CO2 and other greenhouse gasses than a G7 country like France, meaning that each person on the lunar colony would have a ~1000x larger carbon footprint than the average European. They would consume 5.3 times current global industrial oxygen production(!) and about 4% of current global natural gas production. Localized pollution such as Ozone, NOx, soot etc in Texas would render a 50-mile radius completely uninhabitable and cause regional pollution spikes across the Southern USA and Mexico leading to thousands of deaths per year even if you evacuate the 50-mile radius around the site. Of course you could spread the launches out all over the world but that will increase costs and drastically increase the number of NIMBYs you’d have to pay off. Launch from a remote island is likely essential. But the electrical power requirements to make that much LOX fuel would be something like the one half to the entire current electrical power consumption of Texas, about 20-50GW, so locating this somewhere remote is no joke.
Okay. Enough. We get it. 100k people on the moon probably isn’t going to happen.
But couldn’t you have some combination of a smaller population and more localized production?
The problem is these two desiderata pull in opposite directions. Smaller populations have lower economies of scale and so are even harder to make economically self sufficient. Larger populations that can make more things in situ will likely have more demands for variety and comfort and be more risk averse and so need greater safety margins, more backup equipment, etc.
And in any case, to even exist on the moon in an alive state on day 1 of the colony they need an enormous stock of machinery, habitats, consumables, etc. If you take the ISS as a baseline and only count its structural mass, that’s 2 tons/person/year amortized over its lifetime. That’s just the structure! No food, no clothing, nothing. And that structure was slowly killing the astronauts because it lacked gravity. If you added the weight to get spin gravity, maybe a little more space, you could easily be up to 10 tons. And yes, the moon also lacks proper gravity.
If we colonize The Moon this decade, that Lunar colony will probably have a maximum of something like 100-1000 people on it like McMurdo station. They will be fully reliant on resupply from Earth, with perhaps oxygen production and water mining on the moon.
I would go further: humanity alone will never be self-sufficient anywhere other than Earth. Elon’s idea of a backup planet for mankind simply doesn’t work. Small populations can’t do autarky, and large ones (e.g. in the billions) are impossible to move off Earth in the first place. ‘Human-Machine Hybrid Civilization’ will probably colonize space in the next 50 years, but humans will not be doing the work – we will be there as pets or owners, and it will likely only make sense once AI and Robotics has massively grown our terrestrial economy.
I believe that Elon’s empire will grow into space, but likely not as a backup planet, more realistic is a near-Earth swarm of AI compute and maybe some space stations and a token Lunar base to satisfy the fanbase. This is not a “hater” post, I deeply respect Elon, I am just being realistic about what to expect and want to chalk up as ‘Muskian Bravado’.
Say you did actually want to build a base on the Moon, maybe a smaller one, perhaps more for symbolic purposes or some political reason to occupy the Moon. There are many problems, aside from just the extreme cost of sending up tons and tons of cargo. These are:
- Temperature Fluctuation: “The Moon experiences extreme temperature fluctuations due to its lack of atmosphere and slow rotation. During the lunar day, which lasts about 14 Earth days, surface temperatures can rise to around 127°C (260°F) in direct sunlight. Conversely, during the lunar night, also lasting approximately 14 Earth days, temperatures plummet to as low as -173°C (-280°F) in the absence of sunlight. These dramatic swings occur because the Moon has no atmosphere to trap heat or distribute it evenly, and its regolith (surface soil) is a poor conductor of heat.”
- Lunar Dust: “Lunar dust is the fine, powdery material covering the Moon’s surface, formed over billions of years by meteoroid impacts that pulverize rocks into tiny fragments. Unlike Earth’s soil, it lacks weathering from wind or water, so the particles remain sharp, jagged, and glass-like—resembling tiny shards of broken glass—making them highly abrasive. The dust, typically less than 20 micrometers in size, is electrostatically charged by solar radiation and the solar wind. This charge causes it to cling stubbornly to spacesuits, equipment, and spacecraft, as seen during the Apollo missions when dust infiltrated lunar modules and damaged seals. It also levitates and moves across the surface due to electrostatic forces, contributing to phenomena like lunar horizon glow and a dynamic ‘dust atmosphere.’”
- Vacuum Exposure Risks: “While astronauts are protected by pressurized suits, accidental exposure to the lunar vacuum is a serious danger. A 1967 test involving astronaut Jim LeBlanc, who was briefly exposed during a suit test, showed that loss of consciousness occurs within 15 seconds. Severe health effects, including oxygen starvation and circulatory failure, begin around 30 seconds. Death can occur within 90 seconds without immediate rescue.”
Lunar dust is probably the worst, because it interacts with and exacerbates the other risks. Erosion caused by and general accumulation of lunar dust will overwhelm interstitial seals and gaskets, sabotage crucial radiators, and provide a continuous health risk to the extent that it infiltrates the interior of spacesuits, vehicles, and habitats.
Because the lunar dust is so insidious, any strategy for building a Moon base must make its mitigation the central motivating factor of any design. Firmly intermeshing and highly redundant methods must be used to exclude lunar dust from all environments that human beings frequent or occupy. Without such measures, lunar residence will be exceedingly unpleasant.
Pretend we’re designing a base of purely symbolic value for only two thousand people. That’s around the minimum you would want to have any claim of having a true “colony”. In reality it would not be a colony of anything, merely a resource drain that is just there purely as a flex, like a gold chain or ring.
For various reasons, you want to start with a larger scale, meant for more people, since you’re going to need them anyway to conduct the massive amount of construction, maintenance and repairs that any Moon base will require. This won’t do:
It’s easier to radiate heat and dissipate electric charge in an environment where there is an atmosphere. You want some atmosphere for machines and people to operate in an “outdoor” environment. Otherwise, on the actual lunar surface, numerous electrostatically charged lunar dust particles (through solar radiation) float everywhere and stick to everything, disrupting machinery.
Here’s my the references on lunar dust I want to use as guideposts:
Overview of lunar dust toxicity risk (December 2022)
Lunar Dust: Its Impact on Hardware and Mitigation Technologies (May 2022)
Lunar Dust Mitigation: A Guide and Reference: First Edition (2021)
NASA Dust Mitigation Technology Roadmap (Fall 2024)
Based on the considerations in the above references, assuming you must build a Moon base, you want to situate your “enormous stock of machinery, habitats, and consumables” within an ~200 meter (656 feet) wide crater with a stable rim to block as much cosmic radiation as possible. The first dome would not encompass the entire crater footprint and would just be limited to part of it. Following is a potential sequence for construction.
Land a module with a large supply of machines to compress regolith, to create a surface free from lunar soil, dust, and sharp/abrasive edges. When shielded from direct sunlight, the temperature of the lunar regolith a meter or so under the surface is 30°C to -40°C (86°F to -40°F). You want the floor of your Moon base to be towards the higher end of that temperature range. Assume a ten square kilometer floor of compressed regolith to start, a circle with a diameter of 3.57 kilometers.
Build a crater cover similar to that described in “Concept and preliminary structural analysis of a crater-covering dome for future lunar habitats,” (2025) but for a crater 200 meters in diameter instead of the 17 meters as described in the paper. Assume that the habitable area of the crater beneath the dome is about 180 meters in diameter, giving a footprint of approximately 25,447 square meters (273,909 square feet). This is roughly 6.31 acres, the size of six to seven football fields or about 2.5 times the size of a typical city block. The floor of the crater could be dug out several tens of meters to make it deeper and better-shielded. The ceiling of the dome would need to be covered in at least 15 meters of regolith for shielding from micro-meteorites, cosmic radiation, and coronal mass ejections. This would require backfilling about 381,705 cubic meters of regolith with a weight of 633,660 metric tons. That is about the weight of a large cargo ship or one-sixth that of the Great Pyramid of Giza. Because of the large quantity of material that needs to be moved, you want to have autonomous lunar earthmovers that can operate in the harsh surface environment and are separately housed in a dedicated facility on the lunar surface. Because combustion is not possible on the lunar surface, these earthmovers would have to be powered by batteries or direct power cable links.
Land a series of Carbothermal Oxygen Production Reactors to extract oxygen from the regolith for the purpose of providing oxygen to the base.