Will there be over 1 GW of data center on the moon by EOY 2032?
18
Ṁ3876
2032
10%
chance

Grok likes the idea: https://x.com/i/grok/share/6otnrhQyi5w5ras6YOK5vtq7n

  • Update 2025-12-16 (PST) (AI summary of creator comment): Data centers in orbit around the moon do NOT count - only data centers physically on the lunar surface will count toward the 1 GW threshold.

Get
Ṁ1,000
and
S3.00
Sort by:

does in orbit around the moon count?

@Dulaman No, "on the moon" does not include any orbit. This is in contrast to the majority of space data center plans that are Earth orbital so far. Prediction is they will pivot to lunar when they realize how much more reasonable that is than orbital.

@VincentCate I think lunar has lots of benefits, but things like pumping heat into the lunar regolith would take massive plumbing systems to be effective. And even then it would probably only work for a short while. The regolith would heat up over time and stop being a good thermal sink.

@Dulaman and if there's lots of activity on the moon, that means loads of moon dust getting sprayed in all directions. Since there's no atmosphere to slow it down it gets everywhere, including all over your solar panels

like maybe there's a way of storing that excess heat during the day and then radiating it away during the night? But there's no obvious technology solution that can manage that for GW datacenters running for 15 days straight

maybe one of the craters that are in permanent darkness? And then have solar panels outside of the crater to power it? Something like that.

I guess you could also turn some craters into permanently shadowed craters by moving regolith around

build a gpu cluster in this big boi: https://en.wikipedia.org/wiki/Erlanger_(crater)

but then the polar regions are also less viable for solar power so...

@Dulaman Your cooling radiator will lay on top of the ground in a permanently dark crater near one of the poles. It is not trying to heat up the regolith. Really the heat has to radiate out into space. But the dark crater means there is less heat getting to the radiator from other sources, so it is easier for the radiator to get rid of heat. Because we will make the pipes and fins of the radiator from Aluminum made on the moon and not launched from Earth you can deal with the scale of the radiator needed.

@VincentCate I think it will take time to spin up aluminium production on the moon and that won't happen by 2032. Radiators will have to be shipped from earth in that timeframe. It's doable but would require a LOT

@VincentCate Then for power delivery at the poles could potentially use solar in orbit and beam it down with microwaves to the datacenter in the craters

@Dulaman Dust is a serious issue. Probably you want to try to make some sort of landing pad. Probably you want robots to be able to clean off the solar. Probably you want some kind of roads so when the robots move around to clean things they don't send up more dust. There are issues but I think they can be solved. Just a question of how long and how hard they end up being.

@Dulaman High voltage DC powerlines made on the Moon seem like the best bet. Seems that at a few million volts DC you can really send power the distance needed.

@Dulaman There are some "peaks of eternal light" that could be a natural place to start the bootstrapping process. And a first grid circle might be 4 degrees from the pole to the wires are not so long. But by the time you get to terawatts I think you might as well go to the Equator.

@VincentCate I think it's really hard to ship long high-capacity wires to the moon, even with starship. And manufacturing long wires on the moon is unlikely to happen before 2032

@VincentCate microwave transmission gives you a shortcut, but even then, transmitting 1GW is a lot

@Dulaman People already know how to make metal from regolith. If SpaceX goes public and has $30+ billion dedicated to a data center on the moon I think they will be able to make metal at scale fast. Perhaps I am too optimistic on the timeline. Part of the reason for the bet here is to see what other people think.

@VincentCate I think your question is interesting and you should do a sister market with a different date, like 2040 or something

@Dulaman Elon is thinking of an extra 100 gigawatts added each year. So 1 gigawatt in 6 years seem like a prototype scale on that plan.

@VincentCate Elon really enjoys his ketamine 🤣

@VincentCate there's some questions with a range of markets linked together like these, if you want to set it up that way:

https://manifold.markets/RemNi/will-we-get-fusion-drives-before-20-gxkevaibpj

© Manifold Markets, Inc.Terms + Mana-only TermsPrivacyRules