So like I don't get to use my degree except to tutor kids and answer physics questions and the like, so indulge me for a second because this doesn't make much sense to me. So like the endgame is to create a new Pryor, OK facility in space, right? It doesn't make sense if we're not shipping something that's about the same as what you can build on Earth.
Epoch.AI says that this facility has 100,000 TPU chips, eats 370 megawatts of electricity, but designed to get rid of up to 480 megawatts of heat. Sun power is 1.36 kW/m^2, which is something like 7 megawatts per football field (assuming ~5000 m^2), but you have to put up something like 3-5x that because solar panels are only 20-30% efficient, right? Assume that you can arrange it so that the solar panels are one side, the radiators on the other, you can maybe get away with only 300-400 football fields maybe? So like 1.5 million m^2 or 1.2 km wide, right? At LEO, 650 km, you have (1.2km)/(650 km) * 180 degrees/pi is 0.11 degrees or 6.6 arcminutes of visual size. The moon is only like 30 arcminutes. And the paper talks about how they're going to not do one big monolithic construction but an oval of fridge-sized objects separated out 2-3 times this size -- so like I don't see how you get another Pryor, OK size data center in LEO without basically having it look like a second, smaller moon flying across the sky 10 times per day.
And if this were a wildly successful idea are we talking about having like 5-10 of them, a few for each big frontier lab? This just sounds like we're talking about the most profound shift to our night skies since we started having to deal with light pollution in our cities.
There's a lot of different designs besides this one.
So far as I can work out, literally all of the plans are bad. The "why" varies, but they're all bad.
I'm too tired to double check your maths, so I will assume correct: one likely difference even for this plan is a terminator following sun-synchronous orbit, which means you'll only see it twice a day despite the orbital period being about 90-100 minutes, and when you see it will be specifically at sunrise and sunset.
Visibility is also a question of reflection, not just size. Terminator following orbits are worse than normal satellites, because one of the tricks for reducing e.g. Starlink visibility is to tilt them as they cross the terminator and you can't do that if they're always on the terminator.
The SpaceX plans (a million small ones) becomes a glitter band in some parts of the sky and will appear visually contiguous in other parts, though I need to double check my maths and assumptions about visibility given this happens during sunrise and sunset so the sky itself is pretty bright.
I think there are at least a few problems with the math’s assumptions:
1) Don’t the radiators need to have more surface area than the solar? (Unless the chips run very hot.)
2) The datacenter will have the pesky earth between it and the sun some fraction of the time, and have to either shut down or run off batteries. ~50%, assuming LEO, right? If you leave LEO, then the latency sucks, so they’re training-only clusters. At 50% the solar doubles and you need 370 megawatt hours per hour of darkness, or you run the machines 50% of the time, rebooting for each orbit. If you make the orbit shorter (so you can have smaller batteries), then they wear out faster. The batteries also emit heat. Plus, you need to double the solar so they charge while the workload is running.
3) How do they cope with cosmic rays? The standard approach is still to duplicate or triplicate all computation, or use larger/slower processes, right?
The obvious answer to each question makes the engineering design at least twice as dumb, and they stack. There are many other problems like these.
>1) Don’t the radiators need to have more surface area than the solar? (Unless the chips run very hot.)
to generate 1KW you need 5m2 of solar panels. And black body radiation of 1.5m2 at 70 C is 1KW.
A unit with 1 GPU, 2m2 radiator and 5m2 solar panels is say 20kg. At promised Starship price well under $100/kg, that is less than $2000 to put that unit into orbit. That is much cheaper than $15000 per 1KW of a ground-based datacenter, especially when additionally factoring in [expensive and climate change causing] ground-based electricity vs. free electricity in space once you launched the unit with its solar panels.
Add political opposition on Earth, Iranian drones hitting datacenters, various laws (i.e. costs) that your ground-based datacenter is subject too ... the space starts to look like a very cozy place for a datacenter :)
at current prices it's above $1000/kg to orbit, and we're just conveniently ignoring all of the operational constraints of a completely unmanned datacenter upon which you can perform no maintenance.
What do you do when a micrometeorite causes pinprick leaks in your radiator loop?
is LEO sufficiently shielded from cosmic and solar radiation?
nobody is building real datacenters in space at $1000/kg - at that price the cost is comparable to ground-based and thus no big point. The explosion will start at $100/kg when as i mentioned the base cost would beat ground-based almost 10x, and thus would allow for the additional costs (with total cost still beating ground-based) of all the additional issues/concern you and the others usually mention in this context.
4 years ago we had the ChatGPT moment and notice how civilizational change has been accelerating since then. In a few years we're gong to have the seemingly profound Starship moment. Back at the time it was obvious, i may be said it is even here, that Starship low price will cause explosion of amount of launched payloads and a technological revolution as a result, yet it wasn't clear what kind of payloads it would really be. And now we have the AI - Starship and AI are basically ideally dove-tail each other. Two revolutionary technologies accelerating each other - Strarship lowering deployment cost and removing other obstacles for AI while AI providing guaranteed massive scale launch market for Starship - that will be a thing to watch (or participate if you're lucky :)
As per another of my comments, Alphabet, applying learning curves arbitrarily far into the future, recon it will take SpaceX launching 370,000 tons to LEO to make the costs come down enough to be worth it: <a href="https://arxiv.org/pdf/2511.19468" rel="nofollow">https://arxiv.org/pdf/2511.19468
That's not even $100/kg, that's $200/kg.
Even my bull case puts that ($200/kg) 10 years off, which is so far away it lacks relevance just because compute and AI models move so much faster than that timescale; my bear case says that's about 45 years off.
Always account for how over-optimistic Musk's public timelines are. Starship has only just a few days ago managed its first circular orbit, but when he first talked about it:
The basic game plan is like we're going to send a mission to Mars with every Mars opportunity from 2018 onwards. So and they occur approximately every 26 months. So you know, we're establishing cargo flights to Mars that people can count on for cargo.
crdrost · · focus · HN ↗
Epoch.AI says that this facility has 100,000 TPU chips, eats 370 megawatts of electricity, but designed to get rid of up to 480 megawatts of heat. Sun power is 1.36 kW/m^2, which is something like 7 megawatts per football field (assuming ~5000 m^2), but you have to put up something like 3-5x that because solar panels are only 20-30% efficient, right? Assume that you can arrange it so that the solar panels are one side, the radiators on the other, you can maybe get away with only 300-400 football fields maybe? So like 1.5 million m^2 or 1.2 km wide, right? At LEO, 650 km, you have (1.2km)/(650 km) * 180 degrees/pi is 0.11 degrees or 6.6 arcminutes of visual size. The moon is only like 30 arcminutes. And the paper talks about how they're going to not do one big monolithic construction but an oval of fridge-sized objects separated out 2-3 times this size -- so like I don't see how you get another Pryor, OK size data center in LEO without basically having it look like a second, smaller moon flying across the sky 10 times per day.
And if this were a wildly successful idea are we talking about having like 5-10 of them, a few for each big frontier lab? This just sounds like we're talking about the most profound shift to our night skies since we started having to deal with light pollution in our cities.
ben_w · · focus · HN ↗
So far as I can work out, literally all of the plans are bad. The "why" varies, but they're all bad.
I'm too tired to double check your maths, so I will assume correct: one likely difference even for this plan is a terminator following sun-synchronous orbit, which means you'll only see it twice a day despite the orbital period being about 90-100 minutes, and when you see it will be specifically at sunrise and sunset.
Visibility is also a question of reflection, not just size. Terminator following orbits are worse than normal satellites, because one of the tricks for reducing e.g. Starlink visibility is to tilt them as they cross the terminator and you can't do that if they're always on the terminator.
The SpaceX plans (a million small ones) becomes a glitter band in some parts of the sky and will appear visually contiguous in other parts, though I need to double check my maths and assumptions about visibility given this happens during sunrise and sunset so the sky itself is pretty bright.
hedora · · focus · HN ↗
1) Don’t the radiators need to have more surface area than the solar? (Unless the chips run very hot.)
2) The datacenter will have the pesky earth between it and the sun some fraction of the time, and have to either shut down or run off batteries. ~50%, assuming LEO, right? If you leave LEO, then the latency sucks, so they’re training-only clusters. At 50% the solar doubles and you need 370 megawatt hours per hour of darkness, or you run the machines 50% of the time, rebooting for each orbit. If you make the orbit shorter (so you can have smaller batteries), then they wear out faster. The batteries also emit heat. Plus, you need to double the solar so they charge while the workload is running.
3) How do they cope with cosmic rays? The standard approach is still to duplicate or triplicate all computation, or use larger/slower processes, right?
The obvious answer to each question makes the engineering design at least twice as dumb, and they stack. There are many other problems like these.
trhway · · focus · HN ↗
to generate 1KW you need 5m2 of solar panels. And black body radiation of 1.5m2 at 70 C is 1KW.
A unit with 1 GPU, 2m2 radiator and 5m2 solar panels is say 20kg. At promised Starship price well under $100/kg, that is less than $2000 to put that unit into orbit. That is much cheaper than $15000 per 1KW of a ground-based datacenter, especially when additionally factoring in [expensive and climate change causing] ground-based electricity vs. free electricity in space once you launched the unit with its solar panels.
Add political opposition on Earth, Iranian drones hitting datacenters, various laws (i.e. costs) that your ground-based datacenter is subject too ... the space starts to look like a very cozy place for a datacenter :)
eutropia · · focus · HN ↗
What do you do when a micrometeorite causes pinprick leaks in your radiator loop?
is LEO sufficiently shielded from cosmic and solar radiation?
trhway · · focus · HN ↗
4 years ago we had the ChatGPT moment and notice how civilizational change has been accelerating since then. In a few years we're gong to have the seemingly profound Starship moment. Back at the time it was obvious, i may be said it is even here, that Starship low price will cause explosion of amount of launched payloads and a technological revolution as a result, yet it wasn't clear what kind of payloads it would really be. And now we have the AI - Starship and AI are basically ideally dove-tail each other. Two revolutionary technologies accelerating each other - Strarship lowering deployment cost and removing other obstacles for AI while AI providing guaranteed massive scale launch market for Starship - that will be a thing to watch (or participate if you're lucky :)
ben_w · · focus · HN ↗
That's not even $100/kg, that's $200/kg.
Even my bull case puts that ($200/kg) 10 years off, which is so far away it lacks relevance just because compute and AI models move so much faster than that timescale; my bear case says that's about 45 years off.
Always account for how over-optimistic Musk's public timelines are. Starship has only just a few days ago managed its first circular orbit, but when he first talked about it:
- Musk, 2016: <a href="https://elonmuskarchive.org/fr/video/code-conference-2016-06-01" rel="nofollow">https://elonmuskarchive.org/fr/video/code-conference-2016-06...