zlacker

[parent] [thread] 20 comments
1. rob74+(OP)[view] [source] 2026-02-03 08:40:16
It kinda does make sense if you consider that solar panels in space have been used for a very long time (to power satellites). However, getting the electricity they generate down to Earth is very complicated, so you end up having to use it in space, and one of few things that would make sense for that is indeed data centers, because getting the data to Earth is easier (and Elon already handily has a solution for that).

However I'm curious how many solar panels you would need to power a typical data center. Are we talking something like a large satellite, or rather a huge satellite with ISS-size solar arrays bolted on? Getting rid of the copious amounts of heat that data centers generate might also be a challenge (https://en.wikipedia.org/wiki/Spacecraft_thermal_control)...

replies(4): >>spider+v3 >>mike_h+U3 >>trhway+84 >>kergon+xe
2. spider+v3[view] [source] 2026-02-03 09:07:05
>>rob74+(OP)
A 10MW data center would require square kilometers of solar arrays, even in space.

It’s just as real as the 25k Model 3.

replies(1): >>tim333+IV
3. mike_h+U3[view] [source] 2026-02-03 09:10:48
>>rob74+(OP)
The plan seems to be for lots and lots of smaller satellites.

For inferencing it can work well. One satellite could contain a handful of CPUs and do batch inferencing of even very large models, perhaps in the beginning at low speeds. Currently most AI workloads are interactive but I can't see that staying true for long, as things improve and they can be trusted to work independently for longer it makes more sense to just queue stuff up and not worry about exactly how high your TTFT is.

For training I don't see it today. In future maybe. But then, most AI workloads in future should be inferencing not training anyway.

replies(1): >>KoolKa+yn1
4. trhway+84[view] [source] 2026-02-03 09:12:35
>>rob74+(OP)
>Getting rid of the copious amounts of heat that data centers generate might also be a challenge

at 70 Celsius - normal for GPU - 1.5m2 radiates something like 1KWt (which requires 4m2 of panels to collect), so doesn't look to a be an issue. (some look to ISS which is a bad example - the ISS needs 20 Celsius, and black body radiation is T^4)

replies(1): >>rocqua+Iq
5. kergon+xe[view] [source] 2026-02-03 10:30:20
>>rob74+(OP)
> It kinda does make sense if you consider that solar panels in space have been used for a very long time (to power satellites).

It stops making sense the second you ask how you’d dissipate the heat any GPU would create. Sure, you could have vapour chambers. To where? Would this need square kilometers of radiators on top of square kilometers of solar panels? All this just to have Grok in space?

replies(4): >>missin+Tr >>geertj+XT >>tim333+7V >>bigmm+6h2
◧◩
6. rocqua+Iq[view] [source] [discussion] 2026-02-03 12:03:30
>>trhway+84
So for the ISS at 20c you'd get 481 W/m^2 so you'd only need 2.3m2. So comparing the ISS at 20c to space datacenters at 70c you get an improvement of 63%. Nice, but doesn't feel game-changing.

The power radiated is T^4, but 70c is only about 17.1% warmer than 20c because you need to compare in kelvin.

replies(1): >>trhway+sM2
◧◩
7. missin+Tr[view] [source] [discussion] 2026-02-03 12:12:13
>>kergon+xe
But space is very cold, so no problem there /sarcasm
◧◩
8. geertj+XT[view] [source] [discussion] 2026-02-03 14:55:57
>>kergon+xe
> It stops making sense the second you ask how you’d dissipate the heat any GPU would create.

The answer, as you surmised, is indeed radiators.

◧◩
9. tim333+7V[view] [source] [discussion] 2026-02-03 15:01:20
>>kergon+xe
You have a dark radiating side on the back of the solar panels. You can spread the GPUs around the solar panels. All the energy in comes from the sun so the temperature should be much the same as any dark panel like object floating in sunlight in space.
replies(2): >>tim333+nM1 >>squibo+1J2
◧◩
10. tim333+IV[view] [source] [discussion] 2026-02-03 15:04:50
>>spider+v3
0.2 sq km approx.
◧◩
11. KoolKa+yn1[view] [source] [discussion] 2026-02-03 16:55:55
>>mike_h+U3
Latency means this still makes no sense to me. Perhaps some batch background processing job such as research or something but that's stretching.
replies(1): >>mike_h+IU3
◧◩◪
12. tim333+nM1[view] [source] [discussion] 2026-02-03 18:33:18
>>tim333+7V
Or something like that - the temperature goes hot and cold as the things go into light and shadow so they have insulation.
replies(1): >>kergon+533
◧◩
13. bigmm+6h2[view] [source] [discussion] 2026-02-03 20:45:15
>>kergon+xe
Elon already answered this type of question before, albeit quite sarcastically iirc, tho I can't find the tweet right now
replies(2): >>squibo+bJ2 >>kergon+h33
◧◩◪
14. squibo+1J2[view] [source] [discussion] 2026-02-03 23:11:25
>>tim333+7V
Random objects floating in space do not have GPUs on them which generate heat. You need to move the heat from GPUs to a radiator, so you are describing the actual solution of radiators in a roundabout way. Radiators weigh an amount and cost money. The consequence of factoring this in with optimistic assumptions is that it's about 1/4 as efficient to build space compute as earth compute. It's hype bullshit.
◧◩◪
15. squibo+bJ2[view] [source] [discussion] 2026-02-03 23:12:48
>>bigmm+6h2
Elon is not a good person to ask on technical matters like this given both his history of saying really silly things about space-related technologies and his enormous incentive to lie to attract investors.
◧◩◪
16. trhway+sM2[view] [source] [discussion] 2026-02-03 23:32:58
>>rocqua+Iq
>The power radiated is T^4, but 70c is only about 17.1% warmer than 20c because you need to compare in kelvin.

17% in T^4 is almost 2x - plugging 293 (in Kelvin of course) in the calculator i get 417 W/m2 vs. 784W/m2 that i got earlier for the 343 (Kelvin for the 70 Celsius).

The ISS targets rejecting 70KW and has something like 140m2 of radiators. These radiators are attached to the ISS and use a lot of plumbing to carry the cooling liquid.

Where is GPUs and everything can be attached directly to the radiators and solar panels. So 70KW - 70 GPUs - can be placed right onto the 10m by 10m radiator panel. In front of those GPUs sitting on that radiator - a 15m by 20m solar panels assembly. Whole thing is less than 1 ton. Between $10K and $100K on Starship.

◧◩◪◨
17. kergon+533[view] [source] [discussion] 2026-02-04 01:09:21
>>tim333+nM1
No, temperature does not decrease significantly when objects are in the shadows, unless hey stay there for a long time. Even when they don’t get energy from solar radiation, they still dissipate it by radiative transfer, which is very inefficient. So they cool down slowly.
replies(1): >>somena+bQ3
◧◩◪
18. kergon+h33[view] [source] [discussion] 2026-02-04 01:10:39
>>bigmm+6h2
Elon cannot change the laws of Physics, is not a serious person, and has no particular engineering skills. He is not authoritative on almost anything. He’s just cosplaying.
◧◩◪◨⬒
19. somena+bQ3[view] [source] [discussion] 2026-02-04 08:23:04
>>kergon+533
I assume the idea is that they'll be in a solar orbit, so there will be a perpetually sun facing side and a perpetually shaded side. The exact physics behind radiating the heat out in this setup are unclear to me, but it seems difficult to imagine that it would pose significant, let alone insurmountable, difficulties.
◧◩◪
20. mike_h+IU3[view] [source] [discussion] 2026-02-04 08:57:06
>>KoolKa+yn1
I think the most providers all give high latency batch APIs significant discounts. A lot of AI workloads feel batch-oriented to me, or could be once they move beyond the prototype and testing phases. Chat will end up being a small fraction of load in the long term.
replies(1): >>KoolKa+Tw6
◧◩◪◨
21. KoolKa+Tw6[view] [source] [discussion] 2026-02-04 23:22:32
>>mike_h+IU3
That would imply there's still capacity here on earth for this type of traffic.
[go to top]