OK, let’s see if we could start with a much smaller habitat, but still big enough to grow most of its food. I assume a two cylinder habitat with 300 metre radius (just big enough to provide 1g artificial gravity without revolving too fast for comfort), and 500 metres long, which would house around 1,000 people. I am assuming that it would be assembled from parts and resources from earth, apart from lunar regolith for soil. I am also assuming that it is placed in low earth orbit, which reduces the need for radiation protection as it is within the Van Allen belts. A thin outer sheet of aluminium would do (as in the ISS) and an inner plastic plate.
Even so, 1.56 million tons need to be transported from Earth, as shown below, and the costs of such transport is very high.
Now the components and assembly of such a habitat, and transporting lunar regolith as soil, would cost hundreds of millions of dollars, but that is minor compared to the cost of transporting such a large tonnage above into even low earth orbit (LEO). Even at $100/kg launch costs, the cost of transporting the parts for such a habitat from Earth would exceed $150 billion, more than my estimate of a large habitat built in long run space conditions which would house 200 times more people.
Space X claim that a fully reusable planned Mars mission system would have launch costs as low as $140/kg, in which case $100/kg or lower would be feasible to LEO. As mentioned before, however, the high risks would still be present for using rockets. It needs to be a long term objective (or even lower) just to get people long term into space, but then potential low cost solutions (such as the spaceplane/skyhook combination suggested in earlier posts) need high volume to be economic, to cover high capital costs, and you won’t get high volume until costs are low … classic “chicken and eg” dilemma. It would cost over $1 billion to develop Skylon, we are probably stuck with incremental improvements to rockets for now (such as recovering first and second stages).
Not only are the above costs and risks too high, this approach does nothing to develop space manufacturing. There is a better route, to be discussed in the next post.
Note that cumulative costs (launch, assembly, maintenance) of the International Space Station is over $100 billion to date – and that only weighs 420 tons, and the living area is barely the size of a suburban house.
There is another problem with LEO as a habitat location - space junk. There are half a million objects circling this planet already, satellites have already been destroyed by collisions, and the problem is getting worse. Even a small habitat would be a much bigger target than a satellite.
The chart below shows the current distribution of objects in space. Source: aer.com
Most are clustered in the 7000-20,000 kms from earth range, within the protection of the outer Van Allen belt, but high enough to have long lived orbits; even so, the higher objects will spiral down over time. There is a smaller cluster at or just below 36,000 kms which is geostationary orbit, but they are exposed to more radiation (within the range of the outer Van Allen belt).
As long as radiation protection can be assured (see previous posts) then habitats are best placed beyond 36,000 kms to avoid the space junk problem, and will need to take measures to ensure that they don’t create their own space junk.
Costs are acceptable, but fresh air isn’t free