> In the interests of clarity, I am a former NASA engineer/scientist with a PhD in space electronics. I also worked at Google for 10 years, in various parts of the company including YouTube and the bit of Cloud responsible for deploying AI capacity, so I'm quite well placed to have an opinion here.
Sorry bud, we're past the stage where expertise means anything or will convince anyone when there are memes and lobbyists.
Putting failure-prone (this doesn't seem to get as much press, but I've been told from people in a position to know that those $$$$$$$$$$$ GPUs have a fairly high failure rate), heat sensitive equipment into what is in effect a universe sized thermos seems less than optimal. Yes there are ways to cool, but they are specialized (i.e. costs more money), don't work as well as options readily available on Earth (i.e. costs even more money) and would have to get into orbit along with the equipment (yet even more money).
Where is the cost savings? Where is the efficiency?
Ahh I see my mistake. I didn't realize that solar energy can only be utilized in orbit. I would imagine even if someone invented a way to do it on Earth, since it gets dark and we have no way of storing energy for low consumption periods would be a problem.
I also didn't realize that stuff in space just spontaneously builds itself. No time needed to build any of the equipment, all this specialized equipment needed to operate in harsh conditions it just manifests itself when someone says the magic words. And it's so much cheaper to launch heavy equipment into space vs. shipping things around on Earth, that when things break, we can just say the magic words to make more equipment appear and launch it. No need to pay FedEx to ship a replacement part and then pay someone to swap it out. That could cost hundreds of dollars! Why pay hundreds of dollars when we can pay millions to replace it in space.
They are designing the satellites to use a sun-synchronous orbit, no batteries needed.
It’s not about price, they have billions flowing, they need scale fast, and solar on land uses lots of-land.
They have plenty of GPU’s ready to go, but nowhere to install them. And the technology gets obsolete in 5 years anyway, no reason to bother replacing anything
I mean no one is actually planning to build data centers in space though. They’re saying it to artificially boost their stock price because they know investors are idiots.
NASA spent most of a decade telling SpaceX that reusable rockets weren't feasible and used their retarded space shuttle as an example. They did nothing to help, just watched, made jokes, and sneared at us while we built up launchpad 39a. They were adamant that we use their "preferred vendor" for parachutes for dragon until our 2 year study showed how much better SpaceX's were than NASA's and all of their design flaws.
Then suddenly every NASA asshole wanted to walk thru Hawthorne with their freinds acting as if NASA not only wasn't a huge burden on everything we did over all those years but pretending they helped while being a huge burden because we had to do the dog and pony show for their fucking tour
The mistake is thinking of this from an engineering or economic point of view. If you think of it from the point of view of “where do the data centers need to go when we are doing stuff that is illegal in most jurisdictions”, the case for space becomes obvious.
This is being deliberately obtuse. A data center in the context of this discussion is a commercially viable operation to generate deepfake porn for Twitter blue checks. We're not talking about a computer in space, but thousands of GPUs operating for profit.
This is actually an even dumber idea. You need just as much radiator, but now thousands times more of every other component, including the lasers. Data centers are made massive because they benefit from economies of scale.
Also, Kessler Syndrome. Do you want Kessler Syndrome, because this is how you get Kessler Syndrome.
It’s not about cost, they have no shortage of money, it’s about scaling fast and sooner than everyone else.
Kessler syndrome is an overblown fear, look at Starlink which only keeps growing. Even if it did happen the LEO satellite would fall out of the sky within a decade. And we’re not clogging up the higher orbits at anywhere near the same scale
I'm aware that the economics are about burning through as much VC cash as possible in the shortest possible time then leaving it to future generations to clean up the mess. That's part of why I call it a stupid idea.
I understand your cynicism but I think you’re out of your depth, this is not a field you can just armchair quarterback. There is real math being done to figure out these solvable problems
The trend of Data Centers in Space is the latest Corporate Puffery to keep TSLA P/E at around 200:1. That's it.
So as an investor your decision is - invest in the nonsense and keep blowing the bubble - or take profit and cash out. The latest Fake It Until You Make it schemes are robots, LLM AI Intelligence and Space Frickin' Datacenters.
Again -you can make money off the puff - but know what you're doing - taking from the greater fool.
Anyone who knows a little bit about computers knows they generate a lot of heat and anyone who knows a little about a vacuum knows that it only lets heat escape via radiation which is pretty limiting until you get far hotter than a computer can tolerate.
> The External Active Thermal Control System (EATCS) is the primary permanent active heat
rejection system on ISS. It acquires, transports, and rejects excess heat from all U.S. and
International Partner modules except the Russian modules. The EATCS contains two ammonia
coolant loops, which cool equipment on the S0, S1, and P1 truss segments. Capable of rejecting
up to 70kW, the EATCS provides a substantial upgrade in heat rejection capacity from the 14kW
capability of the Early External Active Thermal Control System (EEATCS).
> Heat collected by the EATCS ammonia loops is radiated to space by two sets of rotating radiator
wings—each composed of three separate radiator ORUs. Each radiator ORU is composed of
eight panels, squib units, squib unit firmware controller, Integrated Motor Controller Assemblies
(IMCAs), instrumentation, and QDs. Each Radiator ORU measures 76.4 feet (23.3 meters) by
11.2 feet (3.4 meters) and weighs 2,475 pounds (1,122.64 kilograms)
So I’m supposed to believe this person after getting basic facts incorrect? from reading the nasa published paper that they link themselves???
Also tell me who do you think has more real world practical experience with satellites in space regarding actual equipment failures due to radiation?
NASA and the ISS? Or the company that has like what 8k to 10k Starlink satellites in orbit right now? How many has NASA EVER had in space?? Tens or hundreds?
The difference between SpaceX and this person is one is actually building engineering feats, and the other is reminiscing about building shit 40 years ago.
NASA has some truly remarkable things too (mars rover, James Webb, Hubble for example), but this person seems out of their depth.
Peach_Muffin | a day ago
> In the interests of clarity, I am a former NASA engineer/scientist with a PhD in space electronics. I also worked at Google for 10 years, in various parts of the company including YouTube and the bit of Cloud responsible for deploying AI capacity, so I'm quite well placed to have an opinion here.
Sorry bud, we're past the stage where expertise means anything or will convince anyone when there are memes and lobbyists.
mion81 | a day ago
> Cooling on Earth is relatively straightforward. Air convection works pretty well – blow air across a surface
So we just blow vacuum across instead, there you go, yeesh /s
jking13 | a day ago
Putting failure-prone (this doesn't seem to get as much press, but I've been told from people in a position to know that those $$$$$$$$$$$ GPUs have a fairly high failure rate), heat sensitive equipment into what is in effect a universe sized thermos seems less than optimal. Yes there are ways to cool, but they are specialized (i.e. costs more money), don't work as well as options readily available on Earth (i.e. costs even more money) and would have to get into orbit along with the equipment (yet even more money).
Where is the cost savings? Where is the efficiency?
mion81 | a day ago
Indeed.
JambaJuice916 | a day ago
It’s not that simple, you get costs savings from the free solar energy up there.
On earth you have to wait for infrastructure to be built, permits to be approved, etc.
In space you can just chuck thousands of them up with less red tape
jking13 | a day ago
Ahh I see my mistake. I didn't realize that solar energy can only be utilized in orbit. I would imagine even if someone invented a way to do it on Earth, since it gets dark and we have no way of storing energy for low consumption periods would be a problem.
I also didn't realize that stuff in space just spontaneously builds itself. No time needed to build any of the equipment, all this specialized equipment needed to operate in harsh conditions it just manifests itself when someone says the magic words. And it's so much cheaper to launch heavy equipment into space vs. shipping things around on Earth, that when things break, we can just say the magic words to make more equipment appear and launch it. No need to pay FedEx to ship a replacement part and then pay someone to swap it out. That could cost hundreds of dollars! Why pay hundreds of dollars when we can pay millions to replace it in space.
JambaJuice916 | a day ago
They are designing the satellites to use a sun-synchronous orbit, no batteries needed.
It’s not about price, they have billions flowing, they need scale fast, and solar on land uses lots of-land.
They have plenty of GPU’s ready to go, but nowhere to install them. And the technology gets obsolete in 5 years anyway, no reason to bother replacing anything
Fearless_Swim4080 | 6 hours ago
Ah yes it’s totally less red tape to… checks notes …launch rockets.
JambaJuice916 | 6 hours ago
Ah shit you’re right, somebody should tell them, this random redditor figured out something they missed!
Fearless_Swim4080 | 6 hours ago
I mean no one is actually planning to build data centers in space though. They’re saying it to artificially boost their stock price because they know investors are idiots.
JambaJuice916 | 6 hours ago
Not true, looks at SpaceX which just merged with XAi, they already have a reusable rocket and the majority of satellites in orbit are theirs
Fearless_Swim4080 | 6 hours ago
lol. wait, do you actually think he's serious? because again, lol.
JambaJuice916 | 6 hours ago
They said the same thing about landing falcon 9, building starship, the Starlink constellation, etc
Bill Gates said “don’t underestimate Musk”
I don’t ~like~ him and his fascist tendencies, but he is serious
Justthetip74 | 16 hours ago
Credentialism is the worst form of argument.
NASA spent most of a decade telling SpaceX that reusable rockets weren't feasible and used their retarded space shuttle as an example. They did nothing to help, just watched, made jokes, and sneared at us while we built up launchpad 39a. They were adamant that we use their "preferred vendor" for parachutes for dragon until our 2 year study showed how much better SpaceX's were than NASA's and all of their design flaws.
Then suddenly every NASA asshole wanted to walk thru Hawthorne with their freinds acting as if NASA not only wasn't a huge burden on everything we did over all those years but pretending they helped while being a huge burden because we had to do the dog and pony show for their fucking tour
Romanian_ | 16 hours ago
Hey chatgpt help me larp as a former NASA engineer/scientist
Live_Fall3452 | 13 hours ago
The mistake is thinking of this from an engineering or economic point of view. If you think of it from the point of view of “where do the data centers need to go when we are doing stuff that is illegal in most jurisdictions”, the case for space becomes obvious.
zero0n3 | 12 hours ago
He is already getting basic facts incorrect.
His 14kW heat rejection is incorrect. Those are EEATCS numbers (Early external active thermal control system)
AnthraxCat | a day ago
I've played enough Terra Invicta to know that a datacenter in space is a very stupid idea. Until we get Exotic Spike Radiators this is a pipe dream.
Ambitious-Wind9838 | a day ago
In Terra Invicta, you literally build hundreds of data centers in space at the start of the game. Lol.
AnthraxCat | a day ago
Not to belabour a joke too much, but I think the early labs you can build are far from data centers.
Ambitious-Wind9838 | a day ago
Even T1 laboratories are data processing centers where personnel are stuck.
Data centers can vary greatly in size, from a single H200 board to billions. A data center is about function, not size.
AnthraxCat | a day ago
This is being deliberately obtuse. A data center in the context of this discussion is a commercially viable operation to generate deepfake porn for Twitter blue checks. We're not talking about a computer in space, but thousands of GPUs operating for profit.
JambaJuice916 | a day ago
On earth, in space it could be 1 satellite per GPU connected to each other with laser links
AnthraxCat | a day ago
This is actually an even dumber idea. You need just as much radiator, but now thousands times more of every other component, including the lasers. Data centers are made massive because they benefit from economies of scale.
Also, Kessler Syndrome. Do you want Kessler Syndrome, because this is how you get Kessler Syndrome.
JambaJuice916 | 23 hours ago
It’s not about cost, they have no shortage of money, it’s about scaling fast and sooner than everyone else.
Kessler syndrome is an overblown fear, look at Starlink which only keeps growing. Even if it did happen the LEO satellite would fall out of the sky within a decade. And we’re not clogging up the higher orbits at anywhere near the same scale
AnthraxCat | 23 hours ago
I'm aware that the economics are about burning through as much VC cash as possible in the shortest possible time then leaving it to future generations to clean up the mess. That's part of why I call it a stupid idea.
Yes, Starlink is also part of the problem.
JambaJuice916 | 23 hours ago
I understand your cynicism but I think you’re out of your depth, this is not a field you can just armchair quarterback. There is real math being done to figure out these solvable problems
Afraid_Sample1688 | a day ago
The trend of Data Centers in Space is the latest Corporate Puffery to keep TSLA P/E at around 200:1. That's it.
So as an investor your decision is - invest in the nonsense and keep blowing the bubble - or take profit and cash out. The latest Fake It Until You Make it schemes are robots, LLM AI Intelligence and Space Frickin' Datacenters.
Again -you can make money off the puff - but know what you're doing - taking from the greater fool.
Squidgy-Metal-6969 | 22 hours ago
Anyone who knows a little bit about computers knows they generate a lot of heat and anyone who knows a little about a vacuum knows that it only lets heat escape via radiation which is pretty limiting until you get far hotter than a computer can tolerate.
Designer_Version1449 | 14 hours ago
also speaking of radiation, the other kind this time, theres a butt ton of it in space. you know what hates radiation? computers.
hmmmmmm maybe this isnt the greatest idea
zero0n3 | 12 hours ago
You think SpaceX doesn’t have stats on how impactful that is to their equipment???
They already have thousands of satellites up there.
Designer_Version1449 | 5 hours ago
Simple communications satellite vs gpu cluster.
supified | 18 hours ago
How would they cool down?
zero0n3 | 12 hours ago
Except the ATCS system is built on a physics equation where temperature DELTA matters.
Literally in the paper he links.
Also he gets the heat dissipation wrong. And it’s literally in the PDF he links.
> The EATCS is designed to provide 35 kW of heat rejection per loop for a total capability of 70 kW
GPUs are also hotter than the environment temp you need to maintain for those modules.
This dude is out of his depth it seems, unless the paper NASA published itself is wrong???
zero0n3 | 12 hours ago
https://www.nasa.gov/wp-content/uploads/2021/02/473486main_iss_atcs_overview.pdf
> The External Active Thermal Control System (EATCS) is the primary permanent active heat rejection system on ISS. It acquires, transports, and rejects excess heat from all U.S. and International Partner modules except the Russian modules. The EATCS contains two ammonia coolant loops, which cool equipment on the S0, S1, and P1 truss segments. Capable of rejecting up to 70kW, the EATCS provides a substantial upgrade in heat rejection capacity from the 14kW capability of the Early External Active Thermal Control System (EEATCS).
> Heat collected by the EATCS ammonia loops is radiated to space by two sets of rotating radiator wings—each composed of three separate radiator ORUs. Each radiator ORU is composed of eight panels, squib units, squib unit firmware controller, Integrated Motor Controller Assemblies (IMCAs), instrumentation, and QDs. Each Radiator ORU measures 76.4 feet (23.3 meters) by 11.2 feet (3.4 meters) and weighs 2,475 pounds (1,122.64 kilograms)
So I’m supposed to believe this person after getting basic facts incorrect? from reading the nasa published paper that they link themselves???
Like holy fuck wake up.
zero0n3 | 12 hours ago
Edit:
Also tell me who do you think has more real world practical experience with satellites in space regarding actual equipment failures due to radiation?
NASA and the ISS? Or the company that has like what 8k to 10k Starlink satellites in orbit right now? How many has NASA EVER had in space?? Tens or hundreds?
The difference between SpaceX and this person is one is actually building engineering feats, and the other is reminiscing about building shit 40 years ago.
NASA has some truly remarkable things too (mars rover, James Webb, Hubble for example), but this person seems out of their depth.