Which hasn't yet been proven to be either technically or economically viable, even on paper. It's a pipe dream.
The cynical viewpoint is that this is Elon capitalizing on current datacenter hype to inflate SpaceX's valuation based on theoretically overcoming tremendous amounts of hard physics problems, over the next 5-10 years. As he did with FSD, Boring Company / Hyperloop, Twitter, etc.
Neither was reuse of rockets and I remember the ex-boss of Arianespace laughing at those bozos of SpaceX who try to pretend that they are a serious space business.
Musk made some bad bets, but also some good ones (Falcon rockets, Starlink) and some at least promising ones (Starship, Neuralink). And Twitter bought him enormous political influence - I wouldn't consider this a failure either, from the realistically-cynical point of view. That cannot be measured by revenue alone.
I still have no idea how that would work. Imagine launching an entire data center building into space, and then imagine also launching a solar array to power it, and then also launching a gigantic radiator to cool it... and the radiator is full of some kind of liquid that can never leak even though it's in a vacuum.
Like, sure, but also, that seems like a lot of work, a lot of extra cost, and a lot of risk, all just to avoid building it in Kansas.
It was also hard for many people to imagine a reusable booster, a belly flopping Starship, catching the largest booster ever built with “chopsticks”, a 10,000 satellite constellation, etc.
Orbital compute is technically very feasible. We’re not talking about a datacenter-sized structure, but a lot of rack-sized satellites connected by laser links. SpaceX has gotten pretty good at building, launching, and managing large constellations.
Economically it obviously it has challenges, but there are some advantages (6x solar production, free real estate, less regulation, arguably simpler cooling) to balance the extra costs (launch, radiators, lack of access for maintenance, limited lifespan, etc).
>It was also hard for many people to imagine a reusable booster, a belly flopping Starship, catching the largest booster ever built with “chopsticks”, a 10,000 satellite constellation, etc.
I don't actually think this would be hard to imagine. I've been a huge fan of Space X since it's launch exactly because these types of things do seem feasible because they save so much of money if they are achievable.
A moon base with a secondary launch site, yes. Mining asteroids for precious metals, definitely. I'm not some Luddite.
My only point here is that you can build a data center on the ground trivially easily. Any data center that can exist in space could much more easily exist on the ground... where you can update it and fix things that go wrong. The only issue is politics. I'm entirely happy to be wrong here. If someone can explain the thesis, I'd be happy to get on board.
Solar power is already one of the cheaper (and cleaner!) forms of power generation. In dawn dusk sun synchronous orbits where satellites are always fully illuminated the panels will produce around 6 times as much electricity as those on the ground. And you don't need batteries to operate 24/7 (and as a bonus, the satellites will follow the work day demand curve through the day, reducing latency, at least for some people).
A fully reusable Starship will drastically change the economics of both initial launch as well as maintenance. I expect it will become more feasible to send vehicles to refuel/repair/replace components and keep satellites flying longer. Especially for orbital compute where there will be relatively few dense orbital planes.
> catching the largest booster ever built with “chopsticks”
I've heard this often. It's not what happens. More correct to say that engineers (not Musk) got a booster to land in a pre-specified spot. The "chopsticks" aren't waving around to catch anything flying by. The booster comes to them.
LMAO, ok, how about “precisely landing the largest flying structure ever built such that it could be grabbed by two mechanical arms attached to the launch pad”?
That will make it clear it’s not actually a giant Mechazilla robot reaching out to grab the booster using literal chopsticks where ever it happens to come down.
That's the cool part, scoofy. You don't need to understand how it will work, you don't need to take any physics classes, round up enough investors, seek out and explain the basic ideas with the people who will make it happen, or invent anything new. Nor do you need to understand political sciences, taxation, jurisdictions, supply chains, or anything else needed to understand the question behind the Data Centers in Space solution.
You aren't even being roped into it with taxes, nor do you have to buy a single share. Other than willingly reading about it on whichever news sources you choose, your observed life will not change a single bit.
You can choose to seek out that info, or you can remain blissfully ignorant. But please don't join the online cacophony of people polluting the threads thinking everyone wants to understand just how ignorant they are.
I get it, I really do. It's a hard task and you don't understand it. But WHY do you feel the need to share that you don't understand? Do you think it makes you look smarter? Do you feel like you fit in more now? If you seek to understand, why aren't you asking questions??
>You aren't even being roped into it with taxes, nor do you have to buy a single share.
Because of the eventual index inclusions, and insane market cap, this affects nearly everyone with a retirement account. Unless you aren't tracking big indexes for some reason.
And some of our compute will run in space, the point is that it's all happening in the background. Most people have no idea how their retirement accounts work beyond knowing it's a bunch of companies pooled together. You don't need to understand how stocks get traded through Alternative Trading Systems, how the companies can decide to take profits vs paying out dividends, etc. A lot of the information is freely available, but you don't need to understand it or take any action.
You should actually stop and try to imagine it, or failing that, read some of the proposals from companies who want to do it.
None involve launching buildings.
You can look at their models for comparisons with Kansas. There is literature and papers you can read about this. Some go back decades if you include space power transmission, which are related.
It's much more likely SpaceX will continue building more ground data centers and using their sat relays to make global connection faster than ground connections can allow.
Do they, really? Because putting data centers in space would mean multiplying the infrastructure cost by a few orders of magnitude, while being far, far away from cheap energy - photovoltaics would work, certainly, but it will take a lot of it, and it's not like you can just slap panels on the roof - easy cooling, and people.
It's a ridiculous idea, and I don't believe it's what they are really pursuing.
Yes but the statement is in fact a milestone to meet in order to vest Class B stock options, specifically SPCX needs to put 100 terawatts of compute [1] in outer space and beam it back to somewhere, my guess is Earth.
There's even more rewards for putting a million people on Mars and reaching a market cap of 7.5T by a certain date. Oh yeah he has to stay employed too.
From the SEC Form 3 filed June 12th:
1) This Form 3 does not include 1,302,072,285 shares of restricted Class B Common Stock issued to and held of record by the Reporting Person, which may be voted by the Reporting Person, and the vesting of which is subject to the satisfaction of certain performance and other conditions. 1,000,000,000 shares of restricted Class B Common Stock vest upon (i) the Issuer's achievement of specified market capitalization milestones across 15 equal tranches ranging from $500 billion to $7.5 trillion, with each milestone reflecting $500 billion in additional valuation, and (ii) the Issuer's establishment of a permanent human colony on Mars with at least one million inhabitants, in each case, subject to the Reporting Person's continued employment ("SpaceX CEO Award"). 302,072,285 shares of restricted Class B Common Stock vest upon (i) the Issuer's achievement of specified market capitalization milestones across 12 equal tranches ranging from $1.065 trillion to $6.565 trillion, with each milestone reflecting $500 billion in additional valuation, and (ii) the Issuer's completion of non-Earth-based data centers capable of delivering 100 terawatts of compute per year, in each case, subject to the Reporting Person's continued employment ("AI CEO Award")
The way people just casually use that word again now is so sad. And I don't even mean that in an "I'm offended" way, but more of "I'm embarrassed by the way you're trying to be offensive" way.
> I'm embarrassed by the way you're trying to be offensive
Oooor, try this one on for size:
What if they're not out to cause offense and the malice you impute is just an illusion under which you yourself are laboring alone? What if it was a well understood and not particularly offensive vernacular usage from before people decided they ought to spend their time being offended on behalf of hypothetical listeners?
Why use a word that has some offensive quality to it when other words would be just as effective in communicating whatever you're trying to communicate? You're actively making a decision that you know will cause some level of offense. So the only conclusion I can make is that some level of offense is intended.
In 2004 I used to volunteer as a tutor at an afterschool center in a low income housing project. One day a middle schooler was complaining about how much homework they had and I ribbed them a little, "oh, poor baby."
They were stung. "I'm not poor!" I felt so bad about it that it's stuck with me all these years. Does that mean because I've seen first hand how hurtful it can be that I should chide people whenever they use the P word?
"Chide" is not the word I would use because there is a very obvious difference between the offensiveness of "poor" and "retard", you obviously know that. But yes, if I heard another volunteer at a program for low income kids use "poor" in that offhanded context and I saw the pain it caused in those kids, I think it's reasonable to take that volunteer aside and say "be careful using terminology like 'poor' as it can be surprisingly hurtful to kids that are self-conscious about that". You can do that sort of thing in an informative and compassionate way without being "chiding".
And that analogy isn't even accurate because I'm not the one informing you that the word can be hurtful. You're using it already knowing that. So a better question is did you continue to say "oh, poor baby" to the kids who were hurt by your original comment?
Why would you think I'd continue to say it after realizing I'd inadvertently hurt the kid's feelings? You are making assumptions of ill will from me in the anecdote I shared just like you are making assumptions about the OP intending offense because you didn't like their word choice.
>Why would you think I'd continue to say it after realizing I'd inadvertently hurt the kid's feelings?
I don't think you would continue using it, that was the point I was making and it sounds like you now agree with me that we shouldn't be knowingly offensive.
And my point is that I went out of my way not to use it any more in that circumstance.
Yet in the years since, I still talk about poor decisionmaking, poor luck, poor performance, and poor word choice. Because it would be poor logic to go through life auditing everything I say just in case a middle schooler with a somewhat poor vocabulary might mistake my meaning.
Which brings me right back to "there is a very obvious difference between the offensiveness of 'poor' and 'retard', you obviously know that."
"Poor" has non-offensive uses so you can continue to use it in other ways. You don't need to advocate for the non-offensive uses of a slur. You know regardless of the context, some people will be offended by its use.
Mercy me, a slur? Retarded also has a non-offensive meaning. It just means slowed.
In fact, while I wasn't around at the time I'd wager that "mentally retarded" came into an official usage specifically because it was a clinical, sterile, bloodless, and utterly anodyne descriptive term. Moron, imbecile, and idiot all were once clinical terms. And then people throw them back and forth at each other to call each other stupid, they gained an offensive connotation and new terms were needed.
In 20 years will you find it absurd if people say that "differently abled" is a slur? Will you say "this is nuts, we literally came up with that term to avoid offense?" I will!
Yes, I think there is an intention to cause or risk offense. On the other end, I think there is an intention to be offended and failure to mitigate.
It is a fairly common conflict that arises as a flashpoint in many areas. Different social and legal theories come up with radically different standards.
If someone has a cold, should they not go shopping out of caution for others? If someone is immune compromised, is it their responsibility to take precautions in a store?
oh, no it's exactly as jimmy valmer puts it, there is nothing against mentally disabled people, it is just it's something so stupid that one can't even decide were to start to describe all the points in which is stupid, so stupid doesn't possibly cut it
Bringing back and pinning the word to not derail the discussion of "mental illness", "mental handicap", "slow learner", etc or its use as an offensive?
I think the main issue is that no matter which word/phrase is used, some people will use it as a slur, and changing it so often causes more issues than it solves.
For what it’s worth, I’m not trying to be offensive or edgy when I say that word with friends. “The grass is green and that thing (random topic) is retarded.”
You know that many people are offended by that word and yet you use it anyway when other words would get the exact same message across without the offense. The only reasons to use that specific word are either the desire to cause offense or to revel in the possibility of causing offense.
If someone refers to themselves by a particular slur, that does not grant you any social leniency to call them that too. Consider that exact situation with any other particular slur.
Musk using a slur about himself doesn't grant permission for others to use it. Would monegator have used the n-word if Musk had used it to describe himself? Hopefully not. And let's be honest with each other: Musk says things like this for shock value, because it's not an acceptable word.
Space is an abysmal environment for running compute. It offers no real advantages over doing the same thing on Earth, and it's more expensive, too! Energy is far cheaper and more abundant here than in space. And get ready to figure out things like:
- Heat dissipation
- Radiation shielding
- Either the most complex in-space construction ever undertaken, or the most complex distributed computing problem ever undertaken (no, Starlink satellites aren't good enough, we're orders of magnitude away from replicating the speed and reliability of connections within a single room)
- Zero flexibility, zero repairability, zero upgradability. Either it's working, or you make it burn up in the atmosphere with no in-between. Add on that the rationality of sending mountains of precision-manufactured tech containing many uncommon metals only for them to be completely lost. This makes the pricing even worse, in addition to
- Already high costs for designing, building and launching all that in addition to all the extra weight overhead you're taking in components that don't do computation, when the alternative is building a glorified warehouse in the middle of nowhere.
It just doesn't make any sense. It's a project tied up in hype and created solely so spaceflight can be hastily duct taped to the AI investment hysteria. Ask yourself why no one brought this up before or outside the context of AI, despite the lowering of space launch prices and data centers both existing before any of it.
> why no one brought this up before or outside the context of AI
AI compute is different (slightly higher latency is fine for inference, and there's no issue for training), and there has never been so much backlash against data centers or other infrastructure buildout. If an increasingly-non-minority of politicians get their way it will in fact be cheaper and faster to get some servers shipped to space than it will be to get the permits and build it out here on earth.
Also, most Datacenter maintenance is just dealing with the problems you see with space- power and cooling. Solar is 5x better in space and a lot more consistent. Here on land they're shutting down nuclear and imposing so many new regulations on gas and coal (and now on solar and wind) that there aren't many grids that can support growth at the scale that's being requested.
Nobody is claiming that all compute is going to space, which is what you seem to think you're arguing against. There's high-dollar demand for it right now, so if we want to be multi-planetary it's the perfect time to start tackling the "compute in space" problem that needs to be solved. Or you need to prove to future people that "All compute can be forever located on Earth" which seems a lot harder sell.
Assume reusable spaceflight eventually brings launch cost close to the cost of fuel. This is close to happening.
The overhead of building out grid and power infrastructure on land would then exceed the installation speed and cost relative to space based deployments.
Also assume the compute that does make it to space has a short shelf life anyways so lack of ability to repair is a non issue. As we scale manufacturing on land this will increasingly be the case.
China has already run experiments and served models from space, so we know the heat dissipation equation is solvable.
Finally you’d arrive at a similar model that’s already proven successful with Starlink but applied to serving inference.
The key question is speed to scale new deployments to meet demand. If the markets demand is near infinite, they will choose to fund space based deployments over slower land deployments.
That makes no sense to me. You've cited one overhead for building on land but just didn't bring up the billions of overheads for deploying satellites. The associated costs of launching this stuff into space blows the price of boring old power grid infrastructure out of the water, even with launch prices assumed to be the cheapest they can be. For instance, the extreme additional costs of equipping each payload with their individual systems of power generation, flight computers, maneuvering systems, structural components, communications infrastructure, heat dissipation hardware, radiation shielding and so much more. You have to pay the price in weight and money for all of these inefficient small-scale components for each unit launch, compared to plugging in an additional server rack and taking advantage of the economies of scale in a centralized data center that already solve all of these things. And that's before you get into the costs of designing, mass-producing and operating these things, compared to the costs of running a normal data center. And that thing about reusability that I mentioned also cuts into the margins. Data centers can keep working even if their hardware is a generation out of date, they can sell off old equipment, they can repair individual components instead of throwing out a whole rack. This is literally setting your GPUs on fire. It's just one more thing that makes it even more expensive.
Heat dissipation in space is possible, of course, but every kilogram you spend on heat is a kilogram you could've spent on something else. When you're talking about boxes that generate so much heat, you're going to need to spend a lot on that ancillary hardware in each unit that, again, makes it even less rational.
Then the concerns about megastructures or distributed computing go unanswered - to my knowledge, we simply don't have the technology for either of these right now. Starlink isn't close to solving it - the bandwidth of a Starlink satellite is nothing in comparison to the bandwidth of a single current-gen server GPU connection.
> It offers no real advantages over doing the same thing on Earth
Abundant solar energy, free real estate, less regulation, less backlash from NIMBYs, simpler (yes) cooling.
> Energy is far cheaper and more abundant here than in space.
Huh? The sun is obviously the most abundant source of energy in the solar system.
Satellites in a dusk dawn sun synchronous orbit can be fully illuminated 24/7, so they receive ~6x more solar energy than panels on earth. They also don’t need batteries to operate 24/7, and the panels don’t need glass to deflect hail.
> Heat dissipation
Yes, it will require large radiators. They’re mechanically simpler than terrestrial cooling though.
> Radiation shielding
It turns out generative AI is somewhat uniquely robust to occasional bit flips.
- Either the most complex in-space construction ever undertaken, or the most complex distributed computing problem ever undertaken (no, Starlink satellites aren't good enough, we're orders of magnitude away from replicating the speed and reliability of connections within a single room)
It won’t be a single structure, and will only be used for inference, so latency between satellites/racks doesn’t matter.
- Zero flexibility, zero repairability, zero upgradability. Either it's working, or you make it burn up in the atmosphere with no in-between. Add on that the rationality of sending mountains of precision-manufactured tech containing many uncommon metals only for them to be completely lost. This makes the pricing even worse, in addition to
- Already high costs for designing, building and launching all that in addition to all the extra weight overhead you're taking in components that don't do computation, when the alternative is building a glorified warehouse in the middle of nowhere.
I think people vastly underestimate how much a fully reusable Starship will change the economics of space operations.
Not only the initial launch costs, but things like refueling and repairing satellites becomes more economical. I wouldn’t be surprise if SpaceX sends Starships to refuel/maintain satellites to keep them in orbit longer. In fact, SpaceX’s own animation shows modular servers sliding in and out of the satellites.
Energy is not cheaper on earth. Solar in space gets 24/7 power at a 30%+ rate vs the surface. Radiative cooling is passive and cheaper because there are no HVACs or chillers and high temperature chips reduce the need. The ISS does this already. Radiation really isn't a big issue with ECC and redundancy and optical links are fine for batch training.
The only hard part making the math work is the launch cost. They need reusable and reliable starship economics. If they hit that goal, it will become cheaper for pre-training, which is 70%+ of the budget for the frontier models.
Space-grade photovoltaics are >10x more expensive than ground based panels. Add some (Tesla) utility scale batteries and it can run 24/7. No need for expensive radiators or rocket launches. And personnel can upgrade the hardware every time there's a new generation of GPU's.
Putting datacenters in deserts around the equator is a much better idea than in Space. If you're really optimizing for cost that is. If you're optimizing for SpaceX meme-stock valuation the former wins
Which hasn't yet been proven to be either technically or economically viable, even on paper. It's a pipe dream.
The cynical viewpoint is that this is Elon capitalizing on current datacenter hype to inflate SpaceX's valuation based on theoretically overcoming tremendous amounts of hard physics problems, over the next 5-10 years. As he did with FSD, Boring Company / Hyperloop, Twitter, etc.
We've been through it over and over. "Tesla is not a car company it's a X company" where X is the current trending theme.
So far, Tesla has been a blockchain, energy, robotics, and now a compute company.
Neither was reuse of rockets and I remember the ex-boss of Arianespace laughing at those bozos of SpaceX who try to pretend that they are a serious space business.
Musk made some bad bets, but also some good ones (Falcon rockets, Starlink) and some at least promising ones (Starship, Neuralink). And Twitter bought him enormous political influence - I wouldn't consider this a failure either, from the realistically-cynical point of view. That cannot be measured by revenue alone.
Twitter was a "success" for musk sure, it was also a catastrophic failure for the rest of western civilisation.
Musk has been coasting on his successes from 10+ years ago. He has nothing good to offer anyone in 2026.
>90% of the launch market and a great WORLDWIDE isp.
> He has nothing good to offer anyone in 2026.
Falcon launches cost dramatically _less_ than comparables like Ariane. In Fact, Ariane had to beg europe of subsidies to keep the program competitive.
Meanwhile, Starship is well on it's way.
You don't know what you're talking about.
Wait, you think Elon solved these problems?
Like it or not he founded (sometimes in-part) but drove these companies to success by demanding deliverables others thought were crazy.
For most of his sucsessful ideas he had sophisticated investors, VC's, to judge the idea and take the bet(at an early stage).
Would any VC(one without a conflict of interest) invest now, for the long term, based on those visions?
I still have no idea how that would work. Imagine launching an entire data center building into space, and then imagine also launching a solar array to power it, and then also launching a gigantic radiator to cool it... and the radiator is full of some kind of liquid that can never leak even though it's in a vacuum.
Like, sure, but also, that seems like a lot of work, a lot of extra cost, and a lot of risk, all just to avoid building it in Kansas.
It was also hard for many people to imagine a reusable booster, a belly flopping Starship, catching the largest booster ever built with “chopsticks”, a 10,000 satellite constellation, etc.
Orbital compute is technically very feasible. We’re not talking about a datacenter-sized structure, but a lot of rack-sized satellites connected by laser links. SpaceX has gotten pretty good at building, launching, and managing large constellations.
Economically it obviously it has challenges, but there are some advantages (6x solar production, free real estate, less regulation, arguably simpler cooling) to balance the extra costs (launch, radiators, lack of access for maintenance, limited lifespan, etc).
>It was also hard for many people to imagine a reusable booster, a belly flopping Starship, catching the largest booster ever built with “chopsticks”, a 10,000 satellite constellation, etc.
I don't actually think this would be hard to imagine. I've been a huge fan of Space X since it's launch exactly because these types of things do seem feasible because they save so much of money if they are achievable.
A moon base with a secondary launch site, yes. Mining asteroids for precious metals, definitely. I'm not some Luddite.
My only point here is that you can build a data center on the ground trivially easily. Any data center that can exist in space could much more easily exist on the ground... where you can update it and fix things that go wrong. The only issue is politics. I'm entirely happy to be wrong here. If someone can explain the thesis, I'd be happy to get on board.
Solar power is already one of the cheaper (and cleaner!) forms of power generation. In dawn dusk sun synchronous orbits where satellites are always fully illuminated the panels will produce around 6 times as much electricity as those on the ground. And you don't need batteries to operate 24/7 (and as a bonus, the satellites will follow the work day demand curve through the day, reducing latency, at least for some people).
A fully reusable Starship will drastically change the economics of both initial launch as well as maintenance. I expect it will become more feasible to send vehicles to refuel/repair/replace components and keep satellites flying longer. Especially for orbital compute where there will be relatively few dense orbital planes.
> catching the largest booster ever built with “chopsticks”
I've heard this often. It's not what happens. More correct to say that engineers (not Musk) got a booster to land in a pre-specified spot. The "chopsticks" aren't waving around to catch anything flying by. The booster comes to them.
LMAO, ok, how about “precisely landing the largest flying structure ever built such that it could be grabbed by two mechanical arms attached to the launch pad”?
That will make it clear it’s not actually a giant Mechazilla robot reaching out to grab the booster using literal chopsticks where ever it happens to come down.
There's no laws in space, which is the key.
That's the cool part, scoofy. You don't need to understand how it will work, you don't need to take any physics classes, round up enough investors, seek out and explain the basic ideas with the people who will make it happen, or invent anything new. Nor do you need to understand political sciences, taxation, jurisdictions, supply chains, or anything else needed to understand the question behind the Data Centers in Space solution.
You aren't even being roped into it with taxes, nor do you have to buy a single share. Other than willingly reading about it on whichever news sources you choose, your observed life will not change a single bit.
You can choose to seek out that info, or you can remain blissfully ignorant. But please don't join the online cacophony of people polluting the threads thinking everyone wants to understand just how ignorant they are.
I get it, I really do. It's a hard task and you don't understand it. But WHY do you feel the need to share that you don't understand? Do you think it makes you look smarter? Do you feel like you fit in more now? If you seek to understand, why aren't you asking questions??
>You aren't even being roped into it with taxes, nor do you have to buy a single share.
Because of the eventual index inclusions, and insane market cap, this affects nearly everyone with a retirement account. Unless you aren't tracking big indexes for some reason.
And some of our compute will run in space, the point is that it's all happening in the background. Most people have no idea how their retirement accounts work beyond knowing it's a bunch of companies pooled together. You don't need to understand how stocks get traded through Alternative Trading Systems, how the companies can decide to take profits vs paying out dividends, etc. A lot of the information is freely available, but you don't need to understand it or take any action.
You should actually stop and try to imagine it, or failing that, read some of the proposals from companies who want to do it.
None involve launching buildings.
You can look at their models for comparisons with Kansas. There is literature and papers you can read about this. Some go back decades if you include space power transmission, which are related.
It's much more likely SpaceX will continue building more ground data centers and using their sat relays to make global connection faster than ground connections can allow.
Do they, really? Because putting data centers in space would mean multiplying the infrastructure cost by a few orders of magnitude, while being far, far away from cheap energy - photovoltaics would work, certainly, but it will take a lot of it, and it's not like you can just slap panels on the roof - easy cooling, and people.
It's a ridiculous idea, and I don't believe it's what they are really pursuing.
xAI could tie in just fine without Cursor in the picture.
A great business for the rocket logistics company since radiators are a thing.
A less good business for the data center company.
yes musk said that, but that's retarded, a statement made to fill as many bingo spaces as possible
Why is this the case? I want him to be correct, but I am also skeptical.
> I want him to be correct
Why?
If you're curious, I explained why space data centers are such an irredeemably stupid idea in my eyes a few comments up.
Yes but the statement is in fact a milestone to meet in order to vest Class B stock options, specifically SPCX needs to put 100 terawatts of compute [1] in outer space and beam it back to somewhere, my guess is Earth.
There's even more rewards for putting a million people on Mars and reaching a market cap of 7.5T by a certain date. Oh yeah he has to stay employed too.
From the SEC Form 3 filed June 12th: 1) This Form 3 does not include 1,302,072,285 shares of restricted Class B Common Stock issued to and held of record by the Reporting Person, which may be voted by the Reporting Person, and the vesting of which is subject to the satisfaction of certain performance and other conditions. 1,000,000,000 shares of restricted Class B Common Stock vest upon (i) the Issuer's achievement of specified market capitalization milestones across 15 equal tranches ranging from $500 billion to $7.5 trillion, with each milestone reflecting $500 billion in additional valuation, and (ii) the Issuer's establishment of a permanent human colony on Mars with at least one million inhabitants, in each case, subject to the Reporting Person's continued employment ("SpaceX CEO Award"). 302,072,285 shares of restricted Class B Common Stock vest upon (i) the Issuer's achievement of specified market capitalization milestones across 12 equal tranches ranging from $1.065 trillion to $6.565 trillion, with each milestone reflecting $500 billion in additional valuation, and (ii) the Issuer's completion of non-Earth-based data centers capable of delivering 100 terawatts of compute per year, in each case, subject to the Reporting Person's continued employment ("AI CEO Award")
The way people just casually use that word again now is so sad. And I don't even mean that in an "I'm offended" way, but more of "I'm embarrassed by the way you're trying to be offensive" way.
> I'm embarrassed by the way you're trying to be offensive
Oooor, try this one on for size:
What if they're not out to cause offense and the malice you impute is just an illusion under which you yourself are laboring alone? What if it was a well understood and not particularly offensive vernacular usage from before people decided they ought to spend their time being offended on behalf of hypothetical listeners?
Why use a word that has some offensive quality to it when other words would be just as effective in communicating whatever you're trying to communicate? You're actively making a decision that you know will cause some level of offense. So the only conclusion I can make is that some level of offense is intended.
In 2004 I used to volunteer as a tutor at an afterschool center in a low income housing project. One day a middle schooler was complaining about how much homework they had and I ribbed them a little, "oh, poor baby."
They were stung. "I'm not poor!" I felt so bad about it that it's stuck with me all these years. Does that mean because I've seen first hand how hurtful it can be that I should chide people whenever they use the P word?
"Chide" is not the word I would use because there is a very obvious difference between the offensiveness of "poor" and "retard", you obviously know that. But yes, if I heard another volunteer at a program for low income kids use "poor" in that offhanded context and I saw the pain it caused in those kids, I think it's reasonable to take that volunteer aside and say "be careful using terminology like 'poor' as it can be surprisingly hurtful to kids that are self-conscious about that". You can do that sort of thing in an informative and compassionate way without being "chiding".
And that analogy isn't even accurate because I'm not the one informing you that the word can be hurtful. You're using it already knowing that. So a better question is did you continue to say "oh, poor baby" to the kids who were hurt by your original comment?
Why would you think I'd continue to say it after realizing I'd inadvertently hurt the kid's feelings? You are making assumptions of ill will from me in the anecdote I shared just like you are making assumptions about the OP intending offense because you didn't like their word choice.
>Why would you think I'd continue to say it after realizing I'd inadvertently hurt the kid's feelings?
I don't think you would continue using it, that was the point I was making and it sounds like you now agree with me that we shouldn't be knowingly offensive.
And my point is that I went out of my way not to use it any more in that circumstance.
Yet in the years since, I still talk about poor decisionmaking, poor luck, poor performance, and poor word choice. Because it would be poor logic to go through life auditing everything I say just in case a middle schooler with a somewhat poor vocabulary might mistake my meaning.
Which brings me right back to "there is a very obvious difference between the offensiveness of 'poor' and 'retard', you obviously know that."
"Poor" has non-offensive uses so you can continue to use it in other ways. You don't need to advocate for the non-offensive uses of a slur. You know regardless of the context, some people will be offended by its use.
Mercy me, a slur? Retarded also has a non-offensive meaning. It just means slowed.
In fact, while I wasn't around at the time I'd wager that "mentally retarded" came into an official usage specifically because it was a clinical, sterile, bloodless, and utterly anodyne descriptive term. Moron, imbecile, and idiot all were once clinical terms. And then people throw them back and forth at each other to call each other stupid, they gained an offensive connotation and new terms were needed.
In 20 years will you find it absurd if people say that "differently abled" is a slur? Will you say "this is nuts, we literally came up with that term to avoid offense?" I will!
Yes, I think there is an intention to cause or risk offense. On the other end, I think there is an intention to be offended and failure to mitigate.
It is a fairly common conflict that arises as a flashpoint in many areas. Different social and legal theories come up with radically different standards.
If someone has a cold, should they not go shopping out of caution for others? If someone is immune compromised, is it their responsibility to take precautions in a store?
Do you just.. never swear?
It takes a big ego to look past fundamental attribution errors.
https://en.wikipedia.org/wiki/Fundamental_attribution_error
oh, no it's exactly as jimmy valmer puts it, there is nothing against mentally disabled people, it is just it's something so stupid that one can't even decide were to start to describe all the points in which is stupid, so stupid doesn't possibly cut it
If you can't think of any other intensified synonyms for "stupid", you just might be...
Ironically, we can thank Elon Musk for that too.
edit: Gross that you're being downvoted. HN crowd needs a serious look in the mirror.
Bringing back and pinning the word to not derail the discussion of "mental illness", "mental handicap", "slow learner", etc or its use as an offensive?
I think the main issue is that no matter which word/phrase is used, some people will use it as a slur, and changing it so often causes more issues than it solves.
For what it’s worth, I’m not trying to be offensive or edgy when I say that word with friends. “The grass is green and that thing (random topic) is retarded.”
You know that many people are offended by that word and yet you use it anyway when other words would get the exact same message across without the offense. The only reasons to use that specific word are either the desire to cause offense or to revel in the possibility of causing offense.
exactly.
There's no need to use slurs.
Musk himself has identified as such: “For the record, I am a fat retard.”
So, that is in fact, his word.
If someone refers to themselves by a particular slur, that does not grant you any social leniency to call them that too. Consider that exact situation with any other particular slur.
Musk using a slur about himself doesn't grant permission for others to use it. Would monegator have used the n-word if Musk had used it to describe himself? Hopefully not. And let's be honest with each other: Musk says things like this for shock value, because it's not an acceptable word.
Energy will be the biggest bottleneck to data centers on land. Is not an issue in space. Space is the perfect env for running compute.
Space is an abysmal environment for running compute. It offers no real advantages over doing the same thing on Earth, and it's more expensive, too! Energy is far cheaper and more abundant here than in space. And get ready to figure out things like:
- Heat dissipation
- Radiation shielding
- Either the most complex in-space construction ever undertaken, or the most complex distributed computing problem ever undertaken (no, Starlink satellites aren't good enough, we're orders of magnitude away from replicating the speed and reliability of connections within a single room)
- Zero flexibility, zero repairability, zero upgradability. Either it's working, or you make it burn up in the atmosphere with no in-between. Add on that the rationality of sending mountains of precision-manufactured tech containing many uncommon metals only for them to be completely lost. This makes the pricing even worse, in addition to
- Already high costs for designing, building and launching all that in addition to all the extra weight overhead you're taking in components that don't do computation, when the alternative is building a glorified warehouse in the middle of nowhere.
It just doesn't make any sense. It's a project tied up in hype and created solely so spaceflight can be hastily duct taped to the AI investment hysteria. Ask yourself why no one brought this up before or outside the context of AI, despite the lowering of space launch prices and data centers both existing before any of it.
> why no one brought this up before or outside the context of AI
AI compute is different (slightly higher latency is fine for inference, and there's no issue for training), and there has never been so much backlash against data centers or other infrastructure buildout. If an increasingly-non-minority of politicians get their way it will in fact be cheaper and faster to get some servers shipped to space than it will be to get the permits and build it out here on earth.
Also, most Datacenter maintenance is just dealing with the problems you see with space- power and cooling. Solar is 5x better in space and a lot more consistent. Here on land they're shutting down nuclear and imposing so many new regulations on gas and coal (and now on solar and wind) that there aren't many grids that can support growth at the scale that's being requested.
Nobody is claiming that all compute is going to space, which is what you seem to think you're arguing against. There's high-dollar demand for it right now, so if we want to be multi-planetary it's the perfect time to start tackling the "compute in space" problem that needs to be solved. Or you need to prove to future people that "All compute can be forever located on Earth" which seems a lot harder sell.
Assume reusable spaceflight eventually brings launch cost close to the cost of fuel. This is close to happening.
The overhead of building out grid and power infrastructure on land would then exceed the installation speed and cost relative to space based deployments.
Also assume the compute that does make it to space has a short shelf life anyways so lack of ability to repair is a non issue. As we scale manufacturing on land this will increasingly be the case.
China has already run experiments and served models from space, so we know the heat dissipation equation is solvable.
Finally you’d arrive at a similar model that’s already proven successful with Starlink but applied to serving inference.
The key question is speed to scale new deployments to meet demand. If the markets demand is near infinite, they will choose to fund space based deployments over slower land deployments.
That makes no sense to me. You've cited one overhead for building on land but just didn't bring up the billions of overheads for deploying satellites. The associated costs of launching this stuff into space blows the price of boring old power grid infrastructure out of the water, even with launch prices assumed to be the cheapest they can be. For instance, the extreme additional costs of equipping each payload with their individual systems of power generation, flight computers, maneuvering systems, structural components, communications infrastructure, heat dissipation hardware, radiation shielding and so much more. You have to pay the price in weight and money for all of these inefficient small-scale components for each unit launch, compared to plugging in an additional server rack and taking advantage of the economies of scale in a centralized data center that already solve all of these things. And that's before you get into the costs of designing, mass-producing and operating these things, compared to the costs of running a normal data center. And that thing about reusability that I mentioned also cuts into the margins. Data centers can keep working even if their hardware is a generation out of date, they can sell off old equipment, they can repair individual components instead of throwing out a whole rack. This is literally setting your GPUs on fire. It's just one more thing that makes it even more expensive.
Heat dissipation in space is possible, of course, but every kilogram you spend on heat is a kilogram you could've spent on something else. When you're talking about boxes that generate so much heat, you're going to need to spend a lot on that ancillary hardware in each unit that, again, makes it even less rational.
Then the concerns about megastructures or distributed computing go unanswered - to my knowledge, we simply don't have the technology for either of these right now. Starlink isn't close to solving it - the bandwidth of a Starlink satellite is nothing in comparison to the bandwidth of a single current-gen server GPU connection.
> It offers no real advantages over doing the same thing on Earth
Abundant solar energy, free real estate, less regulation, less backlash from NIMBYs, simpler (yes) cooling.
> Energy is far cheaper and more abundant here than in space.
Huh? The sun is obviously the most abundant source of energy in the solar system.
Satellites in a dusk dawn sun synchronous orbit can be fully illuminated 24/7, so they receive ~6x more solar energy than panels on earth. They also don’t need batteries to operate 24/7, and the panels don’t need glass to deflect hail.
> Heat dissipation
Yes, it will require large radiators. They’re mechanically simpler than terrestrial cooling though.
> Radiation shielding
It turns out generative AI is somewhat uniquely robust to occasional bit flips.
- Either the most complex in-space construction ever undertaken, or the most complex distributed computing problem ever undertaken (no, Starlink satellites aren't good enough, we're orders of magnitude away from replicating the speed and reliability of connections within a single room)
It won’t be a single structure, and will only be used for inference, so latency between satellites/racks doesn’t matter.
- Zero flexibility, zero repairability, zero upgradability. Either it's working, or you make it burn up in the atmosphere with no in-between. Add on that the rationality of sending mountains of precision-manufactured tech containing many uncommon metals only for them to be completely lost. This makes the pricing even worse, in addition to - Already high costs for designing, building and launching all that in addition to all the extra weight overhead you're taking in components that don't do computation, when the alternative is building a glorified warehouse in the middle of nowhere.
I think people vastly underestimate how much a fully reusable Starship will change the economics of space operations.
Not only the initial launch costs, but things like refueling and repairing satellites becomes more economical. I wouldn’t be surprise if SpaceX sends Starships to refuel/maintain satellites to keep them in orbit longer. In fact, SpaceX’s own animation shows modular servers sliding in and out of the satellites.
Energy is not cheaper on earth. Solar in space gets 24/7 power at a 30%+ rate vs the surface. Radiative cooling is passive and cheaper because there are no HVACs or chillers and high temperature chips reduce the need. The ISS does this already. Radiation really isn't a big issue with ECC and redundancy and optical links are fine for batch training.
The only hard part making the math work is the launch cost. They need reusable and reliable starship economics. If they hit that goal, it will become cheaper for pre-training, which is 70%+ of the budget for the frontier models.
Space-grade photovoltaics are >10x more expensive than ground based panels. Add some (Tesla) utility scale batteries and it can run 24/7. No need for expensive radiators or rocket launches. And personnel can upgrade the hardware every time there's a new generation of GPU's.
Putting datacenters in deserts around the equator is a much better idea than in Space. If you're really optimizing for cost that is. If you're optimizing for SpaceX meme-stock valuation the former wins
Perfect, minus that pesky law of thermodynamics.
What about the heat?
[dead]