In light of SpaceX's recently winning the contract to go to the moon for NASA, I started thinking (again) about what avenues sending a rocket that size might open in our near-future. So for a short story I'm writing, I'm building a first base on the moon.

pretty big

SpaceX's vehicle (do I need to say this) is the one on the left. Yes, that thing is big, ten story building big, and with (multiple rounds of) in-space refueling, and expendable launchers (use them to build your first habs on the moon), we can send about 100 tons to the moon each time. SpaceX is making sounds (through their CEO) that suggest a build/launch marginal cost (i.e. ignoring initial R&D) of about USD \$5-10M. It's still not "dirt-cheap", mind you but at that cost, NASA, SpaceX or whomever hires them to do so can afford to send a lot of mass. (For scale relative to the enormous US economy, US defense budget in 2021 was USD \$733 billion, so a fleet of 100 starships at the price mentioned above would be less than 0.2% of that).

Now, to set up my first smelter/factory, how much mass would I need? Let's put a very low bar in place and say it needs to be able to process 1 ton of material per day. For comparison, a rather primitive ore smelter with late 1800's technology could process 20,000 tons per day. But you know, we're in space. Stuff is a bit hardercitation needed there. So we'll make allowances by reducing production several thousand-fold. Still if you think about it, even being able to produce 1 ton of stuff, say aluminum rods or iron bars, per day, would dramatically reduce habitat construction costs on the moon, since each ton you produce locally is a ton you don't have to schlep around expensively from Earth, and especially for bulk materials, that seems to make a lot of sense to get going as soon as humanly feasible.

So, how small can this kind of setup be, in terms of mass?

  • 1
    $\begingroup$ 100t is more than enough to prime production on the moon, it tigthly connected to bootstraping problem. But the problem is - do it right - because there is more than one way to do so, and u typical not that smart approach may cost more than military budget and yeild nothing substancial. Thinking about habs in that early stage is basically slashing any good option before it matures, wait for 1GW power production before u move in that direction, relay on automation teleoperation before u reach that 1GW in energy production, which can be a base for starting some habs and stuff. $\endgroup$
    – MolbOrg
    Commented Apr 21, 2021 at 15:35
  • 1
    $\begingroup$ switched to science-based tag $\endgroup$ Commented Apr 21, 2021 at 15:46
  • 2
    $\begingroup$ Aren't most of the materials already on the moon? So you only need the machinery to mine/dig, and enough food/water to last until the next re-supply/crew shift change(?) Then the thing grows exponentially. $\endgroup$
    – Len
    Commented Apr 21, 2021 at 16:42
  • 2
    $\begingroup$ @Len - traditionally, "machinery to mine-dig" is very mass-intensive. As is smelting. And heat for mining-related activities is a little harder to come by on the moon. Also, there's a lot of regolith to get through before you get to useful minerals. Also, while the materials may be there, they're scattered in different locations all over the moon, just like on Earth. $\endgroup$
    – jdunlop
    Commented Apr 21, 2021 at 17:32
  • 1
    $\begingroup$ @jdunlop: heat is free for half the month if you’ve got enough mirrors. No convection means your bigger problem is cooling things down!! $\endgroup$
    – Joe Bloggs
    Commented Apr 21, 2021 at 17:53

2 Answers 2


Let's look at what is needed for smelting any amount of ore. First off what are you smelting? Different items melt at different temperatures (IE Iron is 1500C, while aluminum is approx 650c - https://www.metalsupermarkets.com/melting-points-of-metals/). You need a container to smelt the items in which can withstand the melting temperature required. For many of the materials required you can actually hollow out a bowl in the regolith and use that for melting provided a viable heat source and a means to bleed off the metals. It melts between 1350 and 1600 Kelvin, which is approx 1000-1300 Celsius.

As to a heat source there are many varieties of molten salt reactors which are relatively lightweight and can product the requisite heat (http://fhr.nuc.berkeley.edu/wp-content/uploads/2014/09/AHTR.Nuclear.Technology.Article.May20.2003.pdf) provided you can bleed off the excess heat.

Next you need the requisite mining equipment. A large excavator weighs approx 50 metric tons (https://www.gregorypoole.com/new-equipment/machines/excavators/352f-l-hydraulic-excavator/) and the requisite batteries would weight another 2-3 metric tons. One should suffice to start with. Along with the digging you would need a hauler. As it must be an offroad vehicle, a 42 metric ton ton weighs approx 8 tons (https://www.gregorypoole.com/new-equipment/machines/off-highway-trucks/770g-off-highway-truck/). Again you are looking at another 2-3 tons for batteries.

Next you need charging infrastructure. This would require a means of converting reactor heat to electricity. Seeing as how steam is not an effective medium in a 0 pressure environment you are likely looking at thermovoltaics. The many, if not all, US space missions used this method for power in their launch vehicles (https://en.wikipedia.org/wiki/Radioisotope_thermoelectric_generator). This is a lightweight solution (Potonium gives approx 140W/g) to recharging batteries along with providing power as required to workers on site.

Finally there is habitation for workers. Bigelor Aerospace has their B330 Lunar module in planning right now with a launch weight of 23 tons and provides 113 square meters of space, 330 cubic meters total volume. 2 or 3 of these would provide ample accommodation for workers.

So for total weights we have

  • 170t (salt reactor)
  • 50t (excavator)
  • 8t (dump truck)
  • 10t (lithium ion batteries - generous estimate)
  • 23t (habitation)
  • 1t (power)
  • 20t (oxygen, water, nitrogen(truck tires) and other sundries)

I think that is all you would need if you can make regolith work - ceramics might work to contain heat within the regolith to avoid melting from excess heat - add another 10-20t for that.

To sum, then, we are looking at approx 302 metric tons. Assuming a lift payload of 16.8t (Falcon heavy to trans mars injection payload to allow for lunar landing and return) we are estimating approximately 18 lifts. If we call it 20 lifts for flexibility you are looking at 100M to 200M USD to begin the process.

For further consideration you will also want to look at which materials will be required on your base as it's being built to protect against radiation and also to prevent air leakage as many of the smelted materials will be somewhat porous in nature. You will likely need some form of plastic or ceramic lining which can be sprayed on the interior, as well as numerous airlocks in case of a breach (asteroids will be far more common than on earth). It may be wiser to use your mining equipment to hollow out tunnels several hundred feet below ground for a protective layer and line them with melted regolith. But that is just my thoughts on the matter :)

Good luck on this.

  • $\begingroup$ Could you use mirrors (either on the sunny side or many more in orbit) to avoid having to have a salt reactor? $\endgroup$ Commented Apr 22, 2021 at 17:20
  • 1
    $\begingroup$ Wouldn't you still want to use steam, or some kind of working fluid, in a closed system, rather than thermovoltaics, which are very inefficient? The reason for RTG thermovoltaics in space probes is no moving parts for longevity, but as a result they only get about 100 watts out of them. Seems like a manned site would need much more power, and could justify some extra maintenance on a compact high-power reactor. $\endgroup$
    – Brianorca
    Commented Apr 22, 2021 at 21:56
  • $\begingroup$ @SerbanTanasa Too much chance for mirrors being damaged or coated by dust. Micro meteorites will over time cause damage to the reflectivity if not the integrity of mirrors, and dust caused by mining will not settle as quickly as on earth, and will also continue in roughly the same direction as it started moving (low gravity, no atmosphere). There is also the weight consideration: How many mirrors are needed? What infrastructure to support those mirrors? What protection if using sunlight directly? $\endgroup$
    – Geoff
    Commented Apr 23, 2021 at 19:48
  • $\begingroup$ @Brianorca The problem with steam is that it is used to turn a turbine to create power. This means that everything must be enclosed in a pressurized container. As steam derives from water you have the weight now of the water, the turbine and the casing. As water boils at a lower temperature according to air pressure, what effect will the pressure on the moon have? Hence you will need the sealed, pressurized container. We are looking at significant weight differences. $\endgroup$
    – Geoff
    Commented Apr 23, 2021 at 19:53
  • $\begingroup$ @Brianorca Sorry, ran out of space :) As I showed above, while RTG's are low power they are also extremely lightweight.( 140W/g of fuel) which allows significant capacity to adjust power as required. Also redundancy is extremely important. 1 turbine. 1000 RTG's. Losing one turbine is catastrophic for all workers. Losing 100 RTG's is an inconvenience. Again we are looking at the total picture of efficiency, weight, safety and effectiveness. A turbine is an ideal for efficiency and effectiveness, but sadly is a little lacking in the other categories. $\endgroup$
    – Geoff
    Commented Apr 23, 2021 at 19:56

Only a few kilograms!

The concept of convergent assembly (http://www.zyvex.com/nanotech/convergent.html) has been around for decades but has long been considered a solution in search of a problem. In my opinion space colonization will be its killer app.

In short, our modern notion of what a factory is and looks like is a relic of the scale of pre-20th century technology. Also in earthly industries where time is money there's an understandable impulse to minimize the manufacturing times of desired goods by essentially trading off energy and mass for time. When it comes to space colonization though this equation is reversed. Energy and mass become very precious while time is for all practical purposes unlimited.

With that in mind, we do not have to send full-scale factories to other worlds to begin the process of industrialization. We could instead send clusters of automated milli-scale (1/1000) factories that will independently collaborate in-situ to produce centi-scale factories and tools. Those then collaborate to produce deci-scale factories and tools, all ultimately converging to produce full-scale factories and tools. Resource and energy requirements would grow gradually and proportionally with scale which would help to maximize efficiencies over time. Each scale generation would also have the option of recycling the previous generation back into raw resources.

Interestingly, bootstrapping it would require a period of the process in reverse. We would have to use full-scale factories and tools to make the deci-scale factories and tools, to make the centi-scale factories and tools, etc. That process would require inventions and technologies that would have any number of other marketable uses so it could financially bootstrap itself as well.

Conceptually it is achievable with modern technology. It would just require lots of something that humans hate to conceed: time.

  • $\begingroup$ We do not have deci-scale production capacities, or it lower limit for some cases. Deci meaning something below 10cm, as 10-100 we do ahve some stuff, but it's mass not few kg's. But nice answer concept. $\endgroup$
    – MolbOrg
    Commented May 27, 2021 at 20:42

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .