8 ms·
> The basic math is that launching a million tons per year of satellites generating 100 kW of compute power per ton would add 100 gigawatts of AI compute capaci
by rybosworld 8mo ago
> The basic math is that launching a million tons per year of satellites generating 100 kW of compute power per ton would add 100 gigawatts of AI compute capacity annually, with no ongoing operational or maintenance needs. Ultimately, there is a path to launching 1 TW/year from Earth.
> My estimate is that within 2 to 3 years, the lowest cost way to generate AI compute will be in space.
This is so obviously false. For one thing, in what fantasy world would the ongoing operational and maintenance needs be 0?
- moomoo11 8mo agoThey'll just be decommissioned and burn up in space. nVidia will make space-grade GPUs on a 2-3 year cycle.
- pantalaimon 8mo agoThey don't need to be space grade, consumer hardware will do just fine. For AI a random bit flip doesn't matter much.
- ohyoutravel 8mo agoRandom bit flips might even improve output.
- adastra22 8mo agoSingle upset events in a modern GPU are not bitflips. They destroy the surrounding circuitry and usually disable the whole unit.
- pantalaimon 8mo agoIf that happens you disable that CUDA core. If you GPU is too damaged, you deorbit the satellite.
- adastra22 8mo agoYeas, and this will happen within weeks of launch with the orbits under consideration.
- q3k 8mo agoOnly if that bitflip happens somewhere in your actual data, vs. some GPU pipeline register that then locks up the entire system until a power cycle. Or causes a wrong address to be fetched. Or causes other nasty silent errors. Or... Try doing fault injection on a chip some time. You'll see it's significantly easier to cause a crash / reset / hang than to just flip data bits. 'rad-triggered bit flips don't matter with AI' is a lie spoken by people who have obviously never done any digital design in their life.
- gbriel 8mo agoAs long as they stay below Van Allen belts and deal with weaker magnetic shielding in sun synchronous orbit (high latitudes). I would say they probably something a little beefier than consumer hardware and just deal with lots of failures and bit flips. But cooling is a bigger issue probably?
- adastra22 8mo agoYou do realize that “space-grade” involves process changes that intrinsically incur orders of magnitude efficiency losses? Larger process sizes, worse performing materials. It’s not just a design thing you can throw money at.
- moomoo11 8mo agoPre-requisites: Ketamine
- jopsen 8mo ago> in what fantasy world would the ongoing operational and maintenance needs be 0? Well, if you can't get there, you can't do maintenance, so there is zero maintenance :)
- adastra22 8mo agoSatellites have large operational costs. Satellite fleets even more so.
- wmf 8mo agoPet satellites or cattle satellites?
- fanf2 8mo agoKessler satellites
- torginus 8mo agoI remember reading somewhere that satellites are extra expensive for 2 reasons: - launch costs are so high that doing exotic bespoke engineering might be worth it if it can shave off a few pounds - once again because launches are expensive and rare, you cannot afford to make mistakes, so everything has to work perfectly If you are willing to launch to lower orbits, and your launch vehicle is cheap, you are building in bulk, then you can compromise on engineering and accept a few broken sats Undergrads afaik even high schoolers have built cubesats out of aluminum extrusions, hobbyist solar panels, and a tape measure as an antenna. These things probably dont do that much, but they are up there and they do work.
- adastra22 8mo agoThey are also expensive because there are unique challenges to making reliable spacecraft. E.g. cosmic rays and microgravity absolutely wreck electronics. Those undergrad cube sats are lucky to last more than a few months in the relative calm of low inclination, low altitude earth orbit. They would die on their first pass in a sun synchronous polar orbit.
- afavour 8mo agoAs soon as a statement contains a timeframe estimate by Musk you know to disregard it entirely.
- Culonavirus 8mo agoThe thing is: at the end of the day, SpaceX takes the "impossible" and makes it "late". People are going to Tory Bruno the space datacenters until one day their Claude agent swarm's gonna run in space and they'll be wondering "how did we get here"?
- haritha-j 8mo agoOr takes the impossible and puts a half baked version of it behind a $99/ month paywall.
- phkahler 8mo agoPay wallet? Starlink was never gonna be free.
- smw 8mo agonapkin math says sq kms of radiators to cool 100MW, it's just patently ridiculous
- p1mrx 8mo agoWhat if they use heat pumps to raise the temperature? Heat rejection is proportional to T^4.
- afavour 8mo agoThe thing is: at the end of the day, making absolute statements about the inevitability of future success is a fool’s errand. Musk has a documented history of failing to deliver on promises, timescale or no. So it’s best to engage in some actual critical thinking about the claims he is making.
- padjo 8mo agoIt's always 2-3 years with this guy
- bckr 8mo agoConveniently, about the amount of time it takes for the average person to forget and/or rematerialize in a new parallel dimension
- lotsofpulp 8mo agoThe average person is not aware or does not care about Elon Musk’s claims and whether or not they come true.
- bckr 8mo agoThe average person sees some headlines and gets a vague awareness that the guy is some kind of super genius who is single handedly changing the world. The average person who has an opinion on musk has roughly the same long term memory consolidation pattern as the average person in general.
- darth_avocado 8mo agoYou’re just not going at the speed of light as this guy’s brain is, time dilation is a thing
- wongarsu 8mo agoYou operate them like Microsoft's submerged data center project: you don't do maintenance, whatever fails fails. You start with enough redundancy in critical components like power and networking and accept that compute resources will slowly decrease as nodes fail No operational needs is obviously ... simplified. You still need to manage downlink capacity, station keeping, collision avoidance, etc. But for a large constellation the per-satellite cost of that would be pretty small.
- fred_is_fred 8mo agoBeing under the ocean in a metal box you don't get too many micro-meteors or cosmic rays though.
- gehsty 8mo agoJust like you don’t get much water in space.
- manquer 8mo agoMy understanding was that access to very large body of cold water was a core feature for the project. The water was to be used for cooling relatively efficiently or cheaply.
- gehsty 8mo agoMy point was that they are both quite hostile environments for different reasons. In the same way space has abundant power supply, subsea has an abundant heat sink.
- samsk 8mo ago...and costs pennies compared to putting anything up there, so it can even enjoy those cosmic goodies.
- willis936 8mo agoHow do you make a small fortune? Start with a big one. The thing being called obvious here is that the maintenance you have to do on earth is vastly cheaper than the overspeccing you need to do in space (otherwise we would overspec on earth). That's before even considering the harsh radiation environment and the incredible cost to put even a single pound into low earth orbit.
- tokyobreakfast 8mo ago> For one thing, in what fantasy world would the ongoing operational and maintenance needs be 0? Do you not understand how satellites work? They don't send repair people into space. This has been a solved problem for decades before the AI gold rush assumed they have some new otherworldly knowledge to teach the rest of the world.
- wat10000 8mo agoThe problem is solved by making satellites extremely expensive.
- verzali 8mo agoI fly satellites. None of them have a zero operational cost. None. Even the most automated cost money to keep running.
- rybosworld 8mo ago> Do you not understand how satellites work? Not trying to be rude - but it's you who doesn't understand how satellites work. The U.S. has 31 GPS satellites in orbit right now. The operational cost of running those is $2 million/day. Not to mention the scale of these satellites would be on the order of 10x-100x the size of the ISS, which we do send people to perform maintenance.
- schmidtleonard 8mo agoThe problem is not solved and the techniques they use to deal with it run directly contrary to maximizing compute, because that's not historically something they have remotely cared about.
- hwillis 8mo ago> They don't send repair people into space. There were five separate flights to service the Hubble telescope. It was designed from the beginning to be repaired and upgraded. https://en.wikipedia.org/wiki/STS-125 https://en.wikipedia.org/wiki/STS-125
- paxys 8mo agoThe famous Musk timeline. "By next year, 2 year tops".
- don_neufeld 8mo agoBy his original timeline we should have landed on Mars… 4 years ago. https://www.planetary.org/articles/20170929-spacex-updated-colonization-plans#:~:text=SpaceX's%20previous%20plan%20called%20for,not%20appear%20to%20have%20diminished. https://www.planetary.org/articles/20170929-spacex-updated-c...
- chihuahua 8mo agoand "Full Self-Driving" was a solved problem in 2016. Ooh, happy 10th anniversary, FSD?
- gostsamo 8mo agoYou lost me on million tons.
- vessenes 8mo ago[flagged]
- rybosworld 8mo ago> I am super intrigued that Elon thinks this. He has a habit of saying things that ultimately are just hype building. I do not believe that he really believes in space data-clusters.
- adastra22 8mo agoI believe that he believe the hype will be good for SpaceX IPO.
- vessenes 8mo agoThat's the most measured Elon critique I've read today. :) I've been told by SpaceX folk that Elon's job is to keep a 20 year view in the future and essentially get folks to work backward from that. I think I might kind of be sold on data-clusters in space in 20 years. I can understand if I had lift that cost 1/10 what everyone else in the world paid for it, I'd be even more sold on them. That said, this newfound enthusiasm of his certainly makes a commercially reasonable path forward to turn xAI stock into spacex stock. Elon takes care of his investors, generally speaking.
- dsr_ 8mo agoElon doesn't think this. Elon says whatever bullshit thing comes into his head without regard for technical, economic or physical plausibility. As long as it raises the stock price! The market has had almost a hundred years of being well-regulated, so when a sociopath lies through their teeth, we're inclined to give them the benefit of the doubt. But in the last few years, that regulation has been worn down to nothing, and the result is and was entirely predictable: fraud.
- phkahler 8mo agoBut spacex isnt a publicly traded company.
- wat10000 8mo agoNever mind operational and maintenance costs. In what fantasy world is it cheaper to put a computer in orbit than in a building on the ground? I don't care how reusable and maintenance-free Starship gets, there's no way even absurdly cheap launches are cheaper than a building. The whole thing makes no sense. What's the advantage of putting AI compute in space? What's even one advantage? There are none. Cooling is harder. Power is harder. Radiation is worse. Maintenance is impossible. The only reason you'd ever put anything in orbit, aside from rare cases where you need zero-gee, is because you need it to be high up for some reason. Maybe you need it to be above the atmosphere (telescopes), or maybe you need a wide view of the earth (communications satellites), but it's all about the position, and you put up with a lot of downsides for it. I feel like either I'm taking crazy pills, or all these people talking about AI in space are taking crazy pills. And I don't think it's me.
- nhq1298 8mo agoThe most generous interpretation is that the "AI in space" nonsense is a cover for putting limited AI in space for StarShield (military version of StarLink), which is essentially the "Golden Dome". It might be possible to scam the Pentagon with some talk about AI and killer satellites that take down ICBMs.
- quesera 8mo ago> It might be possible to scam the Pentagon with some talk about AI and killer satellites that take down ICBMs. Honestly that story sounds right up Pete Hegseth's alley.
- wat10000 8mo agoWhy would they need a cover, though? They can just say “we’re putting AI in space so we can shoot down missiles” and that would be fine.
- hristov 8mo agoCurrently, just a cursory google search shows $1500-3000 per kilogram to put something into low earth orbit. Lets take the low bound because of efficiencies of scale. So $1500. A million tons will cost $1500x1000x1000000= 1,500,000,000,000. That is one and a half TRILLION dollars per year. That is only the lift costs, it does not take into account the cost of manufacturing the actual space data centers. Who is going to pay this?
- pantalaimon 8mo agoThat's the price before Starship which would be the prerequisite for this whole project.
- vel0city 8mo agoYes, and as we know Starship will be doing regular commercial launches starting in 2020, maybe 2021. We're getting close to having the time for Starship's delays to be the same as the actual time for the Saturn 5 to go from plans to manned launches (Jan 1962-Dec 1968).
- fooker 8mo agoAre you trying to say it'll be delayed or that it'll never work? One is obviously true, and the other is very likely false.
- vel0city 8mo ago"it'll never work" is quite black and white while "failure" is a lot more of a grey area. Will it actually launch? Sure, we've seen it. Will it actually hit the reliability as sold? Will it have as fast of turnaround time to reach launch timing goals? Can it actually launch as much payload as promised? Will the economics actually shake out as intended? Did the Cybertruck "never work"? Obviously not, they're on the streets. Was it a <$40k truck with >250mi range? No. Did FSD "never work"? Obviously not, tons of people drive many, many miles without touching the wheel. Does Tesla feel confident in it enough to not require safety operators to follow it on robotaxi trips? No. Does Tesla trust it enough to operate in the Las Vegas Loop? No. Has Tesla managed to get any state to allow it to operate truly autonomously? No. Look, I hope Starship does work as advertised. Its cool stuff. But I don't see it as a given that it will. And given by the track record of the guy who promised it, it gives even less confidence. I'm sad there's less competition in this space. We have so many billionaires out there and yet so few out there actually willing push envelopes.
- CGMthrowaway 8mo agoThere's clearly rhetorical hyperbole happening there. But assuming that thermal rejection is good in space, & launch costs continue falling, as earth-based data centers become power/grid-constrained, there is a viable path for space power gen. The craziest part of those statements is "100 kW per ton." IDK what math he is doing there or future assumptions, but today we can't even sniff at 10 kW per ton. iROSA [1] on the ISS is about 0.150 kW per ton. [1]https://en.wikipedia.org/wiki/Roll_Out_Solar_Array https://en.wikipedia.org/wiki/Roll_Out_Solar_Array edit: iROSA = 33 kW per ton, thanks friends
- __alexs 8mo agoJust put a slightly larger solar array on the same equipment on earth?
- JumpCrisscross 8mo ago> put a slightly larger solar array on the same equipment on earth? Land and permitting. I’m not saying the math works. Just that there are envelopes for it to.
- dangus 8mo agoThe math literally works. The US mandates by law that we grow a fuck ton of corn to mix 10% ethanol into gasoline. If you replaced just those cornfields with solar/wind, they would power the entire USA and a 100% electric vehicle fleet. That includes the fact that they are in the corn belt with less than ideal sun conditions. We aren’t even talking about any farmland that produces actual food or necessary goods, just ethanol as a farm subsidy program. The US is already horrendously bad at land use. There’s plenty of land. There’s plenty of ability to build more grid capacity.
- bobtheborg 8mo agoHello fello Technology Connections watcher?
- b00ty4breakfast 8mo agoThis is par for the course for an Elon-associated endeavor but it's been leaking out into the broader tech sector; make ludicrous claims and promises and somehow investors just throw money at you. FSD has been around the corner for over a decade, martian colonization will be here by the end of the decade for the past 20 years and General SuperAI will be here in a few years for the past few years.
- consumer451 8mo agoHere is my main question: Musk is on record as being concerned about runaway "evil AI." I used to write that off as sci-fi thinking. For one thing, just unplug it. So, let's accept that Musk's concern of evil runaway AI is a real problem. In that case, is there anything more concerning than a distributed solar powered orbital platform for AI inference? Elon Musk appears to be his own nemesis.
- circuit10 8mo agoAside from anything about Elon Musk, here’s an interesting video response to the “just unplug it” argument on the Compuerphile channel: https://youtu.be/3TYT1QfdfsM https://youtu.be/3TYT1QfdfsM
- consumer451 8mo agoHa, I figured that might be the video prior to clicking it. I am a long time fan. Agreed, when I wrote "just unplug it," this counterargument was present in my mind, but nobody likes a wall of text. However, my original point was that a distributed solar powered orbital inference platform is even worse! Think about how hard it would be to practically take out Starlink... it's really hard. Now.. >1M nodes of a neural net in the sky? Why would someone who lives as a god, the richest man in the world, the only person capable of doing this thanks to his control of SpaceX... do the literal worst thing possible?
- defrost 8mo agoThat'd easily take a few LEO detonated fragmentation bombs to trigger a cascading LEO shrapnel field.
- consumer451 8mo agoIt's a lot harder than taking out some terrestrial power lines.
- 8mo ago
- slg 8mo ago>This is so obviously false. One of the biggest but most pointless questions I have about our current moment in history is whether the people in power actually believe the stuff they say or are lying. Ultimately I don't think the answer really matters, their actions are their actions, but there is just so much that is said by people like Musk that strains credulity to the point that it indicates either they're total idiots or they think the rest of us are total idiots and I'm genuinely curious which of those is more true.
- codechicago277 8mo agoWe’re at a point where propaganda is so much more powerful than reality that the people in power literally can’t tell the difference. When your source of ethics is the stock price, little details like physical impossibility stop seeming relevant.
- Rzor 8mo agoYou put it so succinctly and perfectly that I'll have to favorite your comment. Totally agree. The physical world has become little more than noise for people like Musk. I wonder whether the correction will be a slow market dip, a full collapse, or somehow whether he makes it out like a bandit. Baudrillard is, once again, uncomfortably accurate in his diagnosis.
- fooker 8mo ago> in what fantasy world It is already more expensive to performance maintenance on SOCs than it is to replace them. Remember, these machines are not for serving a database, there are practically no storage needs (and storage is the component that fails most often.) Given that, the main challenge is cooling, I assume that will be figured out before yeeting 100 billion $ of computers into space. Plenty of smart people work at these companies.
- bobthecowboy 8mo agoDo smart people work at Boring Company? Do smart people work on FSD at Tesla? What about the HyperLoop? It is possible for smart people to make technical achievements without the overall project being particularly successful.
- fooker 8mo agoYou're right. I meant it specifically for figuring out cooling computers in space. I am pretty sure this is going to be a solvable problem if this is the bottleneck to achieve data centers in space, given that newer chips are much more tolerant to high temperatures. https://www.marketplace.org/story/2026/01/07/new-ai-chips-will-need-less-cooling-says-nvidia-ceo https://www.marketplace.org/story/2026/01/07/new-ai-chips-wi...
- haritha-j 8mo agoAny estimate by Elon musk, you need to add or substract a zero to/from the end. Here, I'll fix it for you. > The basic math is that launching a 100,000 tons per year of satellites generating 10 kW of compute power per ton would add 1 gigawatt of AI compute capacity annually, with no ongoing operational or maintenance needs. Ultimately, there is a path to launching 0.01 TW/year from Earth. > My estimate is that within 20 to 30 years, the lowest cost way to generate AI compute will be in space.
- MagicMoonlight 8mo agoBut more importantly, there is no heat dissipation in space. There’s no atmosphere to cool you, no water you can put heat into. Just an empty void. You can radiate a little, but the sun alone is enough to cook you, without you having a rack of GPUs inside your satellite. It’s completely delusional to think you could operate a data centre in a void with nowhere to put the heat.
- thegrim000 8mo agoDo you honestly believe that nobody involved has ever considered that?
- askl 8mo agoApparently not, otherwise this silly idea wouldn't exist. Naysayers probably get fired fast.
- uplifter 8mo agoThe ISS’s solar arrays each weigh a metric ton and generate 35 KW a piece[0], and that’s just for the power collection. They’d need incredible leaps in efficiency for an orbiting ton collecting and performing 100 KW of compute. [0] https://en.wikipedia.org/wiki/Electrical_system_of_the_International_Space_Station https://en.wikipedia.org/wiki/Electrical_system_of_the_Inter...
- 5ersi 8mo agoLaunching alone consumes about 75-150kWh per tonne of energy for fuels only (as per ChatGPT). Planned lifespan of Starlink satellites is 5years.
- tzs 8mo agoA million tons a year would be over 18 Starship launches per day.
- sandos 8mo agoIn my world, sending up new satellites is the new "maintenance". And it is certainly NOT for free!
- perlgeek 8mo agoAnother aspect: GPUs depreciate very fast. There's not much use case for building GPU satellites and expect them to last for 10-20 years. So let's say you expect them to do useful work for you for maybe 2 or 3 years? You have to amortize the launch cost and the build-it-for-space premium in a relatively short time frame. And then what? Reentry? With all the pollution that comes with it? Also, what orbit do you use? Low-earth orbit is already getting pretty full, with starlink and similar constellations taking up quite some space and increasing collision risk. The higher you go, the more your launch costs go up, and the higher your latency. In higher orbits, atmospheric drag doesn't de-orbit failed satellites quickly, increasing risk of Kessler syndrome. All in all, I don't buy it.