> it is possible to put 500 to 1000 TW/year of AI satellites into deep space, meaningfully ascend the Kardashev scale and harness a non-trivial percentage of the Sun’s power
We currently make around 1 TW of photovoltaic cells per year, globally. The proposal here is to launch that much to space every 9 hours, complete with attached computers, continuously, from the moon.
edit: Also, this would capture a very trivial percentage of the Sun's power. A few trillionths per year.
We also shouldn't overlook the fact that the proposal entirely glosses over the implication of the alternative benefits we might realize if humanity achieved the incredible engineering and technical capacity necessary to make this version of space AI happen.
Think about it. Elon conjures up a vision of the future where we've managed to increase our solar cell manufacturing capacity by two whole orders of magnitude and have the space launch capability for all of it along with tons and tons of other stuff and the best he comes up with is...GPUs in orbit?
This is essentially the superhero gadget technology problem, where comic books and movies gloss over the the civilization changing implications of some technology the hero invents to punch bad guys harder. Don't get me wrong, the idea of orbiting data centers is kind of cool if we can pull it off. But being able to pull if off implies an ability to do a lot more interesting things. The problem is that this is both wildly overambitious and somehow incredibly myopic at the same time.
Honestly, there's not a lot else I can think of if your goal is find some practical and profitable way to take advantage of relatively cheap access to near-Earth space. Communication is a big one, but Starlink is already doing that.
One of the things space has going for it is abundant cheap energy in the form of solar power. What can you do with megawatts of power in space though? What would you do with it? People have thought about beaming it back to Earth, but you'd take a big efficiency hit.
AI training needs lots of power, and it's not latency sensitive. That makes it a good candidate for space-based compute.
I'm willing to believe it's the best low-hanging fruit at the moment. You don't need any major technological advances to build a proof-of-concept. Whether it's possible for this to work well enough that it's actually cheaper than an equivalent terrestrial datacenter now or in the near future is something I can't answer.
You really can't grasp that GPUs scaled at this level is the most ambitious thing possible? That it will be the foundation of unfathomable technological innovation?
All right, so how is it that all you geniuses out here are totally right about this, but all the dullards at SpaceX and XAI, who have accomplished nothing compared to you lot, are somehow wrong about what they do every day?
I know being right without responsibility feels amazing but results are a brutal filter.
Do they need to be maintained? If one compute node breaks, you just turn it off and don't worry about it. You just assume you'll have some amount of unrecoverable errors and build that into the cost/benefit analysis. As long as failures are in line with projections, it's baked in as a cost of doing business.
The idea itself may be sound, though that's unrelated to the question of whether Elon Musk can be relied on to be honest with investors about what their real failure projections and cost estimates are and whether it actually makes financial sense to do this now or in the near future.
There are a class of people who may seem smart until they start talking about a subject you know about. Hank Green is a great example of this.
For many on HN, Elon buying Twitter was a wake up call because he suddenly started talking about software and servers and data centers and reliability and a ton of people with experience with those things were like "oh... this guy's an idiot".
Data centers in space are exactly like this. Your comment (correctly) alludes to this.
Companies like Google, Meta, Amazon and Microsoft all have so many servers that parts are failing constantly. They fail so often on large scales that it's expected things like a hard drive will fail while a single job might be running.
So all of these companies build systems to detect failures, disable running on that node until it's fixed, alerting someone to what the problem is and then bringing the node back online once the problem it's addressed. Everything will fail. Hard drives, RAM, CPUs, GPUs, SSDs, power supplies, fans, NICs, cables, etc.
So all data centers will have a number of technicians who are constantly fixing problems. IIRC Google's ratio tended to be about 10,000 servers per technician. Good technicians could handle higher ratios. When a node goes offline it's not clear why. Techs would take known good parts and basically replacce all of them and then figure out what the problem is later, dispose of any bad parts and put tested good parts into the pool of known good parts for a later incident.
Data centers in space lose all of this ability. So if you have a large number of orbital servers, they're going to be failing constantly with no ability to fix them. You can really only deorbit them and replace them and that gets real expensive.
Electronics and chips on satellites also aren't consumer grade. They're not even enterprise grade. They're orders of magnitude more reliable than that because they have to deal with error correction terrestial components don't due to cosmic rays and the solar wind. That's why they're a fraction of the power of something you can buy from Amazon but they cost 1000x as much. Because they need to last years and not fail, something no home computer or data center server has to deal with.
Put it this way, a hardened satellite or probe CPU is like paying $1 million for a Raspberry Pi.
And anybody who has dealt with data centers knows this.
> We currently make around 1 TW of photovoltaic cells per year, globally.
Doubling every three years; at that rate it would take about 30 years for 1TW to become 1000TW. Whether on not the trend continues largely depends on demand, but as of right now humanity seems to have an insatiable demand for power.
We’re not going to use 100% of our solar panel manufacturing capacity to power space data centers, specifically because everyone else on the ground is so power-hungry. If we’re being generous, it could maybe top out at 1%, which adds another ~20 years to your timeline for a total of 50. I think it’s safe to say this part is bunk (along with everything else about this plan which is also bunk).
The quoted "1 TW of photovoltaic cells per year, globally" is the peak output, not the average output. They're only about 20% higher peak output in space… well, if you can keep them cool at least.
Doesn't this risk some unforeseen effects on Earth or the rest of the solar system at that scale? Disruption of magnetic shield, some not yet known law of physics suddenly getting felt etc.?
This is all based on bad math. The people proposing these things don't even have proper scientific and mathematical training to determine what is achievable.
Context missing. This is in reference to a vision the (distant?) future where the satellites are manufactured in factories on the Moon and sent into space with mass drivers.
Full paragraph quote comes from:
> While launching AI satellites from Earth is the immediate focus, Starship’s capabilities will also enable operations on other worlds. Thanks to advancements like in-space propellant transfer, Starship will be capable of landing massive amounts of cargo on the Moon. Once there, it will be possible to establish a permanent presence for scientific and manufacturing pursuits. Factories on the Moon can take advantage of lunar resources to manufacture satellites and deploy them further into space. By using an electromagnetic mass driver and lunar manufacturing, it is possible to put 500 to 1000 TW/year of AI satellites into deep space, meaningfully ascend the Kardashev scale and harness a non-trivial percentage of the Sun’s power.
>
> This is in reference to a vision the (distant?) future where the satellites are manufactured in factories on the Moon and sent into space with mass drivers.
In the meantime, how about affordable insulin for everybody?
Why is it cheaper to ship all of the materials to space, then to the moon for assembly (which also includes shipping all of the people and supplies to keep them alive), then back into space vs just…
Why would satellites be manufactured on the moon? There's nothing on the moon. The raw materials would have to be ferried over first. What would be the point?
Nameplate capacity needs a derate for availability, so you can drop it down to about 200GW(e) equivalent continuous power assuming we're making and deploying enough batteries to support it. More, obviously, if those panels are going to an equatorial desert, less if they're going to sunny Svalbard in the winter time.
A former NASA engineer with a PhD in space electronics who later worked at Google for 10 years wrote an article about why datacenters in space are very technically challenging:
I don't have any specialized knowledge of the physics but I saw an article suggesting the real reason for the push to build them in space is to hedge against political pushback preventing construction on Earth.
I can't find the original article but here is one about datacenter pushback:
But even if political pushback on Earth is the real reason, it still seems datacenters in space are extremely technically challenging/impossible to build.
It's much easier to find a country or jurisdiction that doesn't care about a bunch of data centers vs launching them into space.
I don't get why we aren't building mixed use buildings, maybe the first floor can be retail and restaurants, the next two floors can be data centers, and then above that apartments.
We don’t even have a habitable structure in space when the ISS falls, there is no world in which space datacenters are a thing in the next 10, I’d argue even 30 years. People really need to ground themselves in reality.
Edit: okay Tiangong - but that is not a data center.
> We don’t even have a habitable structure in space
Silicon is way more forgiving than biology. This isn’t an argument for this proposal. But there is no technical connection between humans in space and data centers other than launch-cost synergies.
I don't know, 10 years seems reasonable for development. There's not that much new technology that needs to be developed. Cooling and communications would just require minor changes to existing designs. Other systems may be able to be lifted wholesale with minimal integration. I think if there were obstacles to building data centers on the ground then we might see them in orbit within the next ten years.
> A former NASA engineer with a PhD in space electronics who later worked at Google for 10 years wrote an article about why datacenters in space are very technically challenging
It's curious that we live in a world in which I think the majority of people somehow think this ISN'T complicated.
Like, have we long since reached the point where technology is suitably advanced to average people that it seems like magic, where people can almost literally propose companies that just "conjure magic" and the average person thinks that's reasonable?
> It(Solar) works, but it isn't somehow magically better than installing solar panels on the ground
Umm, if this is the point, I don't know whether to take rest of author's arguments seriously. Solar only works certain time of the day and certain period of year on land.
Also there is so limited calculations for the numbers in the article, while the article throws of numbers left and right.
Elon can't legally financially entangle Tesla to SpaceX due to Tesla being a public company, so his hands are tied.
Tesla is clearly benefiting from protectionism and its sales would collapse if BYD were allowed to openly sell in the US. Most people just want affordable, maintainable and reliable cars.
> Elon can't legally financially entangle Tesla to SpaceX due to Tesla being a public company, so his hands are tied.
He absolutely could do it, just like he did when Tesla bought SolarCity. It just isn’t as easy when one of the companies is public than when both are private.
He's broken pretty much all the other financial rules.... for example, the amount of blatant self-dealing he gets away with is staggering.
As long as the consequences of his actions continue to increase the paper value for investors, regulations don't really have teeth because there aren't damages. So the snowball gets bigger and the process repeats.
> Elon can't legally financially entangle Tesla to SpaceX
Bill Ackman has proposed taking SpaceX public by merging it with his Pershing Square SPARC Holdings, distributing 0.5 Special Purpose Acquisition Rights (SPARs) to Tesla shareholders for each share held. Each SPAR would be exercisable for two shares of SpaceX, aimed at enabling a 100% common stock capitalization without traditional underwriting fees or dilutive warrants.
With SpaceX IPO set to be one of the biggest of all time, this could have a pretty gnarly financial engineering impact on both companies -- especially if the short interest (direct or through derivatives) remains large.
It's "ironic?" considering Tesla launching in China is what created the necessary supply chain to turn BYD into the powerhouse it is today. Tesla's greed will become their own demise.
BYD are just affordable and maybe reliable, regarding maintenance their spares are hard to come by and are almost as hard to work with as Tesla and other brands.
Starship has a large number of critical milestones coming: Can it land and quickly reuse the upper stage? If not, it can't make refueling flights without building a dozen or two starships. Can it carry the full specified payload? If not, it can't even try to refuel in orbit. If it can't refuel in orbit, it can't go beyond earth orbit. Etc.
Everything has to go right or it will be irrelevant before it works.
> Everything has to go right or it will be irrelevant before it works.
Starship is not all of SpaceX. Saying, maybe because one hates Musk, that SpaceX is going to become irrelevant is wishful thinking.
In 2025 SpaceX launched more rockets into space than the entire world ever sent in a year up to 2022, something crazy like that.
Then out of, what, 14 000 active satellites in space more than half have been launched by SpaceX.
SpaceX is, so far, the biggest space success story of the history of the human race (and GP is right in saying that SpaceX is now a national security matter for the US).
Of course it isn't "too big to fail". Even banks aren't. Despite recent history large banks have failed often throughout history. There's no such thing. It may take down the supporting sovereign government (Dutch East Indies) but life goes on and new political orgs appear. People be people.
Too big to fail is a very recent modern myth. Go back 100+ years and lots of banks failed leading into the Great Depression.
Haven't you heard? Tesla is pivoting to building humanoid robots instead. They haven't sold a single one, but it toootally warrants retooling their car factories, pinky promise!
How vital is it really to national security? Starlink will have competition from Amazon Leo in the next few months. And while SpaceX is obviously in the lead in launch capability with Starship, there are multiple launch providers capable of providing roughly the same services the Falcon 9 and Heavy provide today.
SpaceX is slated to go public some time this year - June IIRC
The biggest selling point /was/ that Musk was being managed there, he wasn't tinkering with SpaceX like Twitter or Tesla, and his foolhardy direction was kept out of the company.
BUT, like Tesla, Musk cannot help himself and is making SpaceX look like a very bad investment - tying his other interests with SpaceX, allegedly using SpaceX money as a "war chest" in his battles.
There is also a danger that investors will see xAI as politically dangerous, which will really hurt SpaceX IPO
They want to go public, but have to sell the hell out of it in the meantime.
I'll bet SpaceX financials aren't as great as some people think. Remember, Elon was the guy who tried to take Tesla private, and talked a lot of smack about how silly it is to be a public company. All of a sudden he wants SpaceX to go public?
> The biggest selling point /was/ that Musk was being managed there, he wasn't tinkering with SpaceX like Twitter or Tesla, and his foolhardy direction was kept out of the company
The biggest selling point to who? Definitely not wall street
Let’s be honest - this is just a way to prop up Twitter/X. It makes SpaceX shareholders subsidize X, and also American taxpayers who are giving contracts to SpaceX for highly sensitive things. The government should ideally refuse to give SpaceX work unless it unwinds this.
What do you mean "let it fail?" SpaceX has the most profitable launch system in the world and now operates >50% of all satellites in orbit. They aren't exactly in need of a bailout.
NASA is still here. Unfortunately, NATA fell victim to enshitification by government contracting. NASA even if it wants to simply cannot today design and launch a rocket. :(
I am no fan of Musk the man. SpaceX is a strong company and Falcon is a solid vehicle. There is not a lot of competition, and NASA trying to in-source design and supply and construction of a new, reusable LEO rocket would be a complete nightmare.
I root for a competitive rocket market, but SpaceX is at the moment critical.
> > Boeing is too important to fail as well but it’s been terrible as a shareholder
Your opinion on Boeing being terrible as a shareholder vis-a-vis Tesla would be completely reversed if dividends and capital gains of the 2 companies were to be offered in the form of miles to be flown on Boeing planes and miles on Teslas Uber/Taxi/Autonomous taxis instead of dollars
The absolute overperformance on the stock market that Tesla has enjoyed vis-a-vis Boeing is not rooted in a concrete and tangible quality of life improvement for citizens. Not American citizens, nor global citizens for that matter.
It is my opinion that for all public companies in which it is possible to do so government should mandate payment in kind to all shareholders and board members to prevent the excessive promotional , cult and all around BS aspect of marketing to take over and allow people to profit just by riding off those, and Musk is the GOAT at that.
Twitter (X) owed $1.3B in debt every year in interest since Musk's takeover. This was before re-financing in a higher interest rate environment. The company was losing $200MM+ per year on ~$5B in revenue before the takeover, and there are reports that revenues have decreased by round 50%.
Best case scenario if we accept those numbers is that X makes $3B per year and about half of that goes immediately out the door in debt payments before paying a cent for the entire business to function.
However, if SpaceX acquires X, that ~$1.5B in interest is a fraction of the $8B In profits SpaceX is allegedly generating annually. Further, they can restructure the debt if it's SpaceX's debt, and not owned by X. Investors will be more likely to accept SpaceX shares as collateral than X.
> The company was losing $200MM+ per year on ~$5B in revenue before the takeover, and there are reports that revenues have decreased by round 50%.
X made a profit last year because they cut costs lower than the drop in ad revenue (which is also slowly recovering). The big question is if they will still be profitable in 2026 year without the US election driving big traffic numbers and ads.
He buys twitter at an inflated valuation. Runs it to the ground to a much lower valuation of $9B. [1] Then, his company Xai buys Twitter at a $33B, inflating the valuation up. Then SpaceX merges with Xai for no particular reason, but is expected to IPO at a $1T+ in the upcoming years. [3]
I’m not that smart, but if I were, I would be thinking this is an extended way to move the losses from the Twitter purchase on to the public markets.
It also makes it impossible for Twitter/X to die, as it deserves. It is by far the most toxic mainstream social network. It has an overwhelming amount of far right supremacist content. So bad that it literally resulted in Vivek Ramaswamy, a gubernatorial candidate in Ohio, to quit Twitter/X - nearly 100% of replies to his posts were from far right racists.
Obviously advertisers have not been fans. And it is a dying business. But rather than it dying, Elon has found a clever (and probably illegal) way to make it so that SpaceX, which has national security importance, is going to prop up Twitter/X. Now our taxpayer dollars are paying for this outrageous social network to exist.
ISS radiators run on water and ammonia. Think about how much a kg costs to lift to space and you'll see the economics of space data centers fall apart real fast. Plus, if the radiator springs a leak the satellite is scrap.
I don't pretend to understand the thermodynamics of all of this to do an actual calculation, but note that the ISS spends half its time in the shadow of the earth, which these satellites would not do.
Radiator size scales linearly with power but, crucially, coolant power, pumps, etc do not.
Imagine the capillary/friction losses, the force required, and the energy use(!) required to pump ammonia through a football-field sized radiator panel.
Moving electricity long distance is a lot easier than moving coolant long distances, which puts a soft limit on the reasonable size of the solar array of these satellites. But as long as you stay below that and pick a reasonable orbit it's indeed not too bad, you just have to properly plan for it
5,000 Starship launches to match the solar/heat budget of the 10GW "Stargate" OpenAI datacenter. The Falcon 9 family has achieved over 600 launches.
The ISS power/heat budget is like 240,000 BTU/hr. That’s equivalent to half of an Nvidia GB200 NVL72 rack. So two international space stations per rack. Or about 160,000 international space stations to cool the 10GW “Stargate” datacenter that OpenAI’s building in Abilene. There are 10,000 starlink satellites.
Starship could probably carry 250-300 of the new V2 Mini satellites which are supposed to have a power/heat budget of like 8kW. That's how I got 5,000 Starship launches to match OpenAI’s datacenter.
Weight seems less of an issue than size. 83,000 NVL72’s would weigh 270 million lbs or 20% of the lift capacity of 5000 starship launches. Leaving 80% for the rest of the satellite mass, which seems perhaps reasonable.
Elon's napkin math is definitely off though, by over an order of magnitude. "a million tons per year of satellites generating 100 kW of compute power per ton" The NVL72's use 74kW per ton. But that's just the compute, without including the rest of the fucking satellite (solar panels and radiators). So that estimate is complete garbage.
One note: If you could afford to send up one of your own personal satellites, it would be extremely difficult for the FBI to raid.
I'm not big on this subject, but I understand that heat transfer is difficult in space, because there's little to transfer to. If the solution is just making large radiators, then that means you're sending some big payloads full of radiators. Not to mention all the solar panels needed. I wanna live in sci-fi land too, but I don't see how it makes any sense compared to a terrestrial data center.
You can line the solar panels and radiators facing away from each other, and the radiators would take up less surface area. I think maybe the tricky part would be the weight of water + pipes to move heat from the compute to the radiators.
There's plenty of space in space, but there isn't plenty of space in rocket fairings, nor is there plenty of lift capacity for an unlimited amount of radiators.
It's such bullshit that we've decided this moron and others in his cohort can unilaterally reallocate such vast portions of humanity's labor at their whims.
This is an extremely stupid idea, but because of our shared delusion of capitalism and the idea that wealth accumulation at the top should be effectively limitless, this guy gets to screw around and divert actual human labor towards insane and useless projects like this rather than solving real world problems.
radiators can be made as long as desirable within the shade of the solar panels, hence the designer can pracitically set arbitrarily low temperatures above the background temperature of the universe.
Radiators can shadow each other, so that puts some kind of limit on the size of the individual satellite (which limits the size of training run it can be used for, but I guess the goal for these is mostly inference anyway). More seriously, heat conduction is an issue: If the radiator is too long, heat won't get from its base to its tip fast enough. Using fluid is possible, but adds another system that can fail. If nothing else, increasing the size of the radiator means more mass that needs to be launched into space.
Shading does work; JWST does this. However I don't see how you can make it work for satellite data centers. You would constantly be engaging attitude control as you realigned the panels to keep the radiators in shade. You'd run out of thruster fuel so fast you'd get like a month out of each satellite
Radiators can only be made as long as desirable because there's gravity for the fluid inside to go back down once it condenses. Even seen those copper heat pipes in your PC radiator?
these same comments pop up every time someone brings up satellite data-centers where people just assume the only way of dissipating heat is through convection with the environment.
I guess props to scamming Compaq into making a large investment that didn't pan out. He did personally make money so I guess win for him.
>In an effort to woo investors, Elon Musk built a large casing around a standard computer to give the impression that Zip2 was powered by a supercomputer.
>PayPal
Huh? He didn't found Paypal, his company was acquired by Paypal. You might as well give him credit for eBay while you're at it. Paypal released their first digital wallet in 1999. They acquired x.com (and Musk) in 2000. Paypal itself was then acquired by eBay in 2002.
>Tesla
Investor, not founder.
>SpaceX
Yup, props here.
>Grok/xAI
Hasn't made a penny, no signs it had any path to profitability, which is why it was shoved into Space-X to cover his personal losses.
hmm Tesla shipped millions of cars SpaceX launches 90% of space payloads, Starlink is working well. Thats hard to categorize as never delivered on any of his projects
The crucial thing is that Tesla's valuation has the hype projects baked in. The fact that it never delivered self driving or a robotaxi fleet and is now being saved solely by an import ban on Chinese EVs means that any success he had with Tesla is now an illusion.
The projects promised to be life altering for all mankind, they ended up being not even life altering for super rich Americans considering that Teslas are just EVs which without FSD are just regular cars with a different propellent that were made for political purposes and virtue signaling
The EV revolution has always been something almost dystopic : Trillions of dollars spent in order to not have the slightest amount of quality of life improvement, if anything a worse quality of life because you buy an EV that you cannot use 24/7/365 whereas you can an ICE car for much less .
As soon as something kinda elegant and hopeful as far as collective quality of life improvement is concerned (AI/ChatGPT) came around.....the whole green/EV revolution rightfully went out the window
If Musk was this genius you guys make him to be at 50 and with all the capital he burned he should have at least one company that if you disappeared the world would look drastically different, like if you disappeared Microsoft or Apple or Exxon or Aramco or Amazon or IBM....the world would come to a screeching halt.
Disappear one of Musk companies and everything would be the same as he's always involved in these sort of aspirational companies which have this great vision always 5 years into the future that never materialize into anything tangible or that improves the quality of life like the company I mentioned earlier
In a way, its kind of cool to see how robber barons work in real time in our generation. Its also insanely depressing as they will systematically enshittify and extract as much wealth from society as is possible.
I don't actually think the Robber Barons in the 1920s had people going out of their way to defend them and insist they had special knowledge.
The New Deal happened with massive popular support because people did not like the Barons, and wanted to stop them and actually have a life worth living.
I'm amazed at this kind of thinking. I get it, obviously, and it's not uncommon, but still.
Elon Musk has already revolutionized three industries:
1. EVs: Before Tesla, no one thought electric cars could be a mass-market product. And even today, the Model 3 and Model Y are at the top of almost all sales lists.
2. Orbital Launch: No one expected Space X to succeed. What does a software guy know about real engineering? But today, re-usable rockets are the way of the future, and Space X is at least 5 to 10 years ahead of any other company.
3. Satellite Communications: Every single major military power is trying to deploy their own version of Starlink. Before Starlink, 50 satellites was considered a big constellation. Starlink has 8,000 satellites and they are literally launching hundreds every month.
I know it's impossible to prove a counter-factual, but I'm convinced that none of these three would have happened without Elon. No other Western car company has (even now) produced a profitable EV. No other space company has prices as low as Space X. No one even has the capability to build a Starlink competitor (not yet at least). Without Elon pushing these projects, they simply would not have happened or would have happened decades later (after China or someone else beat us to it).
Even his not-yet-successful projects are far beyond most other companies:
Boring Company has actually built tunnels and passengers are actually riding it. No one else is even trying.
Neuralink has actually helped patients.
Tesla FSD actually does work (I use it all the time), and even if Waymo is ahead, Tesla is easily in second place.
I 100% get the hatred for Elon Musk. His political positions are absolutely worth criticizing and I cringe most of the time he tweets. But to deny his business and engineering ability is just motivated reasoning.
Such illusions are ultimately self-defeating. The more opposed one is to Elon Musk (in business or politics) the more important it is to see his capabilities clearly.
> Boring Company has actually built tunnels and passengers are actually riding it. No one else is even trying.
Boring Company bought an existing tunnel boring machine (TBM), and used it to dig a car tunnel. Their only “innovation” in terms of any cost savings is to dig smaller tunnels - which we already knew could be done (tunnel cost grows with diameter), and which we don’t do for good reasons (capacity, emergency egress).
The branding and marketing exercise was excellent though.
> "The only logical solution therefore is to transport these resource-intensive efforts to a location with vast power and space. I mean, space is called 'space' for a reason. [crying laughing emoji]"
This is all the reasoning provided. It is quite sad how a company I admired so much has become embroiled in financial doohickery.
> 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?
You 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.
How 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.
Unless I missed something the Microsoft underwater data center was basically a publicity stunt.
Anyone who thinks it makes sense to blast data centers into space has never seen how big and heavy they are, or thought about their immense power consumption, much less the challenge of radiating away that much waste heat into space.
An 8 GPU B200 cluster goes for about $500k right now. You'd need to put thousands of those into space to mimic a ground-based data center. And the launch costs are best case around 10x the cost of the cluster itself.
Letting them burn up in the atmosphere every time there's an issue does not sound sustainable.
You mean you operate them like Microsoft's failed submerged data center project [1]. When pointing at validating past examples you are generally supposed to point at successes.
But if we’re going down that line of thinking then it’s a poor comparison. I could open a data centre on the ground and use the same principle of zero maintenance, and it would be way cheaper and way more powerful.
This will totally work since we have an unlimited amount of rare earth elements we can just ship off into space never to see again. Infinite raw materials + infinite power equals infinite AI!!!
The 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"?
There'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.
Not to be an Elon defender, but can you back up your 0.15/ton? My own searching puts ROSA orders of magnitude higher. Each array is 600kg (0.6t) and puts out 20kw (https://space.skyrocket.de/doc_sdat/irosa-1.htm) which makes 20/0.6 = 33.333 kw/ton
Maybe I'm just out of the loop, but is solar substantially more efficient in space? I assume the satellites won't orbit in a way that follows the sun. And presumably the arrays of panels they can attach to a satellite don't exceed the size of the panels you could slap on and around a data center (at least without being insanely expensive).
Elon is a pathological liar and it’s crazy that he still gets sanewashed after all he’s done. It’s insanity that he hasn’t been kicked out of leading his companies, and it’s also insanity that he hasn’t been prosecuted by the SEC.
You’ve spent too much life force trying to even understand the liar’s fake logic.
Let’s start right here: there is no such thing as becoming power/grid constrained on earth. If you replaced just the cornfields that the United States uses just to grow corn for ethanol in gasoline just in the corn belt, you could power the entire country with solar+batteries+wind. Easily, and cheaply.
If you don’t even believe that solar+batteries are cheap (they are), fine, choose your choice of power plant. Nuclear works fine.
The truth is, xAI combining with SpaceX is almost certainly corrupt financial engineering. SpaceX as a government contractor and that means Elon’s pal Trump can now siphon money into xAI via the federal government.
This 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.
Currently, 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?
That launch cost is remarkably cheap to someone that's handled a $1.5million dollar 5U server filled with GPUs and RAM that weighs under 100kg.
Obviously the solar and cooling for the above would both weigh and cost a ton but... It's feels surprisingly close to being within an order of magnitude of current costs when you ballpark it?
Like i don't think it's actually viable, it's just a little shocking that the idea isn't as far out of line as i expected.
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.
Do 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.
Here 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?
Aside 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
What, creating a huge patchwork of self sufficient AIs, forming their own sky based net, seems bad to you, considering the whole torment nexus/Sky Net connotations? It's not like he's planning to attach it to his giant humanoid robot program. Oh. Ohhhhh. Oh no.
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.
We’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.
Any 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.
Never 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.
The 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.
But 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.
Sometimes, I wonder how people in the middle ages accepted the whole "Divine Right" of their ruling Kings, while simultaneously suffering under their rule.
> Larry Ellison, another hyper-informed genius business man
"King George, another royal blessed by the divine."
> He's got more information about space based compute deployments than any other person in the world...
He also had more information about self-driving progress than any other person in the world - yet he was wrong with his predictions every year for last 10 years.
> I will say there's a MASSIVE cost to getting power infrastructure, land, legal stuff done on terra firma; all that just sort of .. goes away when you're deploying to space, at least if you're deploying to space early and fast.
You need both power infrastructure and structures to build within for deploying in space too. And you have to build them and then put it all into space.
Cost per square foot of land is not that high basically anywhere you could build a datacentre to offset that.
Elon 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.
You 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.
The 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.
Elon Musk’s AI startup, xAI (x.ai = grok), officially acquired X Corp. (the parent company of the social media platform X or x.com, formerly Twitter/x.com) in an all-stock deal. Both now operate under a unified holding entity, frequently referred to in corporate filings as X.AI Holdings Corp. (or simply xAI). Now SpaceX has moved to acquire or merge with xAI. This effectively brings the social media data from x.com, the AI development of x.ai, and the satellite infrastructure of Starlink/SpaceX under one "super-conglomerate" roof.
Anyone remember the quote by Russ Hanneman on SV [0] - "No Revenue, means you're potential pure play"
We know datacenters in space - sound plausible enough - yet not practical - hence they're potential pure play - also you can have massive solar in space - unlimited space -- etc -- all true -- but how economical / practical is it ?
yet we know on earth - to power the whole earth with solar - only a fraction of the land is needed. Hell it's even in the Tesla Master Plan v3 docs [1] - current limitation being storage & distribution
so all you - are now witnessing to the greatest scam ever pulled on earth.
Jared Vennett (narration):
"In the years that followed, hundreds of bankers and rating agency's executives went to jail. The SEC was completely overhauled, and Congress had no choice but to break up the big banks and regulate the mortgage and derivatives industries."
"Just kidding. Banks took the money the American people gave them, and they used it to pay themselves huge bonuses, and lobby the Congress to kill big reform. And then they blamed immigrants and poor people, and this time even teachers."
Because this move is entirely financial engineering to hide losses just like the roll up of X in to xAI.
None of this has anything to do with business or innovation. Do you not immediately see that? Most of my friends reaction to this news was that this is so obvious it's almost funny (or actually it is funny, since most were laughing as they read the headline).
I'm curious how you could not understand the relevance of the quote unless you were aggressively trying to not understanding it.
Legitimately, did you not immediately conclude it was for financial shenanigans? What did you think? I'm not trying to be shitty, but what else could there be?
without having watched the Big Short or having read the article, my first impression from the quote is "Megacorporations are failing dramatically, and the billionaires at their helm are freely doing financial gymnastics to pull the covers over the eyes of shareholders, while gaming the system to fully circumvent taxes and regulation -- the people with the power to do anything about it (legislators and regulators) watch idly (maybe profiting), the oligarchs make off like bandits despite copious failures, and the end consumer/taxpayer is either robbed or clueless this is going on, but most likely both, when there was a world where accountability could have been had and the common man was treated better."
the article headline immediately screams "financial gymnastics" to me so the rest followed from the quote.
This merger smells like a bubble. Servers in space? They don't make enough to cover costs here on earth. Americans will be forced to bail this mess up because we need Falcon 9, Starship one, etc.
It is pretty clear. Socialize the losses of Musk investments by recovery via SpaceX contracts, supporting the US space program and the new Golden Dome program.
He sold FSD for 12 years, now is going to sell a Dyson Sphere for the next 30. This guy makes Ponzi look like a street hustler.
I think the argument is that it's messed up that a large debt swap from xAI kept Musk's margin on Twitter from being called by his investors, and now that debt is being absorbed by SpaceX.
Serious question .. are you long on cash or what? What investment class is not overvalued? Gold is but a store of value and doesn't really grow - look at how high it has gone.
I really don't know what you're trying to say. From your comment alone the conclusion I drew was that we should spend more on the industries that makes physical instead of financial products, such as SpaceX.
The sequence of events: Elon doing a leveraged buyout of X, then xAI funding, then debt transfers to X, then the xAI–X stock deal. Now the proposed SpaceX–xAI merger appears to have shifted X’s financial burden from Musk personally toward xAI investors and, potentially, future SpaceX shareholders.
This is speculative, of course, but yeah seems likely.
Who is SpaceX’s biggest customer? And which industry are we being told by any number of governments around the world is the most importantly thing ever and must be subsidized and forced on people?
Just a neat bit of financial engineering. You can tell because Elon picked SpaceX instead of Tesla – which would have actually made sense at some level (Optimus Robots + AI). But Tesla is public and so he'd need to follow laws and reporting requirements.
You can tell it's just financial engineering because in the entire press release xAI is only mentioned in the first two sentences. Everything after that is Elon talking about space data centers to distract from the actual topic. Which seems to be working
This is fairly naive, Elon isn't the only investor in SpaceX.
My guess is "that they did the math" and had an engineering study which convinced them that getting AI datacenters into space will make sense.
It's also not hard to imagine why, the process alone once perfected could be reused for asteroid mining for example, then mirogravity manufacturing, either of which alone would be enormous capital intensive projects. Even if AI dataenters in space are break-even it would be a massive win for SpaceX and leave their competition far behind.
Tesla is still very profitable, as is SpaceX I assume. Twitter/X has been a $44 billion dollar failure, and xAI is a vanity project so Musk can go around saying he is a player in the AI space. Investors in both X and xAI need to be bailed out, hence this announcement.
I suspect SpaceX will acquire Tesla at some point. It’s the most profitable of these companies. So basically SpaceX employees and shareholders are covering up for the failing Tesla business and the already-failed xAI business.
Let’s not forget, xAI is the parent of Twitter/X (the social network). So now, taxpayers are paying to keep Twitter/X alive. After all, it is taxpayer money going to the contracts the government gives SpaceX for launches. Nice way to subsidize what is effectively a one sided campaign machine for the GOP and far right.
I guess the difference is Tesla is a public company, so requires more paperwork. SpaceX isn't public yet, but will be soon, meaning it will have a cash infusion.
With that number of more satellites in orbit, launching a manned rocket into space is likely to be too risky due the amount of more debris and satellites.
And will it be net energy positive solution ?
This is why I come to this site. Obviously, Twitter's financials are struggling and theres more than a few people rich people who don't want to take the hit... but we can all drop that for a second to discuss the plausibility of data centers in space. Some links and comments I enjoyed:
* https://taranis.ie/datacenters-in-space-are-a-terrible-horrible-no-good-idea/
* https://en.wikipedia.org/wiki/Tiangong_space_station
* "Technically challenging", a nice way to say "impossible"
* "I’m not that smart, but if I were, I would be thinking this is an extended way to move the losses from the Twitter purchase on to the public markets."
* "ISS radiators run on water and ammonia. Think about how much a kg costs to lift to space and you'll see the economics of space data centers fall apart real fast. Plus, if the radiator springs a leak the satellite is scrap."
* "5,000 Starship launches to match the solar/heat budget of the 10GW "Stargate" OpenAI datacenter. The Falcon 9 family has achieved over 600 launches." [nerdsniper]
* "No, we just "assume" (i.e. know) that radiation in a vacuum is a really bad way of dissipating heat, to the point that we use vacuum as a very effective insulator on earth."
* "World's Best At Surfing A Temporary Hyperinflation Wave is not a life goal to really be proud of tbh"
I can't tell how many layers of sarcasm are here, but I just want to highlight that aktshually cooling in space is quite difficult because there is no convection, so the only cooling option is radiative. Which gets a bit hard when the satellite gets blasted by the sun.
The ISS doesn't have problems staying warm, it has problems cooling off.
There's an audience for this. People have been taking these vapid statements and his trouble with communicating as evidence of genius for well over a decade now.
Presumably the cooling problem gets hand waved away as a technical detail, and the real selling point is data centers that aren't subject to any regional governments laws.
Send up a spacecraft with back-to-back / equal area solar panels and radiators (have to reject heat backwards, can't reject it to your neighboring sphere elements!). Push your chip temp as much as possible (90C? 100C?). Find a favorable choice of vapor for a heat pump / Organic Rankine Cycle (possibly dual-loop) to boost the temp to 150C for the radiator. Cool the chip with vapor 20C below its running temp. 20-40% of the solar power goes to run the pumps, leaving 60-80% for the workload (a resistor with extra steps).
There are a lot of degrees of freedom to optimize something like this.
Remember MoviePass, and how they were losing gobs of money by letting people see unlimited movies for $20/month?
It was so obviously stupid that a bunch of people went, "well, this so clearly can't work that they must have a secret plan to make money, we'll invest on that promise", and then it turned out there was no secret plan, it was as stupid as it looked and it went bankrupt.
The "datacenters in space" thing is a similar play: it's so obviously dumb that a bunch of smart people have tricked themselves into thinking "wow, SpaceX must have actually figured a way it can work!"; SpaceX has not and it is in fact exactly as stupid as it looks.
The pre-IPO timing of this narrative, combined with Musk’s history of Tesla’s stock pumping, leaves no room for doubt. But, that has worked for Tesla, I’m pretty confident it will work for SpaceX, which will IPO for $1T+.
But it won't end the same as MoviePass until Elon dies; he will keep moving things around, propping up failures with VC, IPO, federal/state (taxpayer) and profit making business money.
My guess would be just a regular radiator and cooling system like a liquid pump. The only obstacle should be the vacuum. That said I don't have any hopes Elon has any understanding of any of it.
Cooling a datacenter in space isn't really any harder than cooling a starlink in space, the ratio of solar panels to radiating area will have to be about the same. There is nothing uniquely heat-producing about GPUs, ultimately almost all energy collected by a satellite's solar panels ends up as heat in the satellite.
IMO the big problem is the lack of maintainability.
> Cooling a datacenter in space isn't really any harder than cooling a starlink in space
A watt is a watt and cooling isn't any different just because some heat came from a GPU. But a GPU cluster will consume order of magnitudes more electricity, and will require a proportionally larger surface area to radiate heat compared to a starlink satellite.
Best estimate I can find is that a single starlink satellite uses ~5KW of power and has a radiator of a few square meters.
Power usage for 1000 B200's would be in the ballpark of 1000kW. That's around 1000 square meters of radiators.
Then the heat needs to be dispersed evenly across the radiators, which means a lot of heat pipes.
Cooling GPU's in space will be anything but easy and almost certainly won't be cost competitive with ground-based data centers.
Sure, but cooling a starlink in space is a lot more difficult than cooling a starlink on earth would be. And unlike starlink which absolutely must be in space in order to function, data centers work just fine on the ground.
According to Gemini, Earth datacenters cost $7m per MW at the low end (without compute) and solar panel power plants cost $0.5-1.5m per MW, giving $7.5-8.5m per MW overall.
Starlink V2 mini satellites are around 10kW and costs $1-1.5m to launch, for a cost of $100-150m per MW.
So if Gemini is right it seems a datacenter made of Starlinks costs 10-20x more and has a limited lifetime, i.e. it seems unprofitable right now.
In general it seems unlikely to be profitable until there is no more space for solar panels on Earth.
I think that it's not just about the ratio. To me the difference is that Starlink sattelites are fixed-scope, miniature satellites that perform a limited range of tasks. When you talk about GPUs, though, your goal is maximizing the amount of compute you send up. Which means you need to push as many of these GPUs up there as possible, to the extent where you'd need huge megastructures with solar panels and radiators that would probably start pushing the limits of what in-space construction can do. Sure, the ratio would be the same, but what about the scale?
And you also need it to make sense not just from a maintenance standpoint, but from a financial one. In what world would launching what's equivalent to huge facilities that work perfectly fine on the ground make sense? What's the point? If we had a space elevator and nearly free space deployment, then yeah maybe, but how does this plan square with our current reality?
Oh, and don't forget about getting some good shielding for all those precise, cutting-edge processors.
I am concerned, and haven’t seen anyone else point this out yet, that Musk will move Grok’s CSAM generation capabilities to space to be beyond the reach of terrestrial policing. Does this create some sort of legal loophole here so Musk can do this with impunity?
No. For one, things in space are under the jurisdiction of the country they were launched from. For another, it's people that do crimes, not satellites or LLMs, and the people involved in making CSAM are all on Earth.
AFAIK he can do whatever he wants in space, but CSAM is still illegal to view or even download in most (all?) countries of the world. So unless the degenerates also move out into space (which I'm sure they're eager to do), it wouldn't really ease the legal situation here on Earth.
Maybe if Elon launched himself and the dev team into orbit and didn't use any ground stations and just Starlink terminals he could start getting into legal loopholes.
Elon didn't want to get outshined after Sam Altman suggested that "maybe we build a big Dyson sphere around the solar system". When will people realize that these "geniuses" are only good at making money, and any benefit to society is coincidental.
There are regularly comments on HN from people who say they pay for Musk's various products, and I am always downvoted into oblivion for suggesting that that is ethically problematic.
There's obviously quite a lot of autocratist illiberals in tech.
But the "fraud" here is being done mostly to VC investors with deep pockets and lawyers, at least until he tries to take this entity public. And I can't imagine them just taking this lying down, but then again maybe they realize that offloading this steaming pile on public market investors is the best way out. But even then... SpaceX seemed like it was quite viable on its own, the investors there are the real losers here.
1. What in the circular funding? This feels more like a financing scheme founding it under X/Twitter and then spinning it over to SpaceX. I suspect some debt is disappearing or taxes aren't getting assessed because of this move.
2. The only thing harder than harnessing "a millionth of the sun's power" on Earth would be launching enough material into space to do the same thing. And that's not even a reason for SpaceX to own an AI company, at least not at this point. The current AI isn't going to help with the engineering to do that. Right now hiring 20-somethings fresh out of college is way cheaper and SpaceX has been very successful with that.
quick edit: dang, I even got point 1 backwards. xAI owns X/Twitter, and that means that SpaceX now owns X/Twitter as well as an AI company. Super suspicious that SpaceX could actually think that buying the social media part (a significant portion of xAI's value) would be worth it.
Most everyone here is just angry Elon opened the door for freedom of speech to exist online and started the process of uncovering the left's extreme corruption. They feel betrayed by him so he's enemy #1. Literally could cure cancer and he'd still be branded as evil by the left. Ironic as he's arguably done more for leftist causes (renewables, ev's, etc) than anyone else on the planet.
These are private companies, with private investors. If Elon really screwed all investors over and over again he wouldnt be able to raise money repeatedly. But he has no problem raising money. Because the claims that he's doing shady things or screwing investors is entirely unfounded. Same goes with the claims that he's an idiot or not that smart. If thats really the case, show me what you've accomplished with your genius.
It takes intelligence to recognize intelligence. An idiot cant recognize a genuis.
Elon's purchase of Twitter (now X) opened the door for conservative view points to be heard online. Prior to that, all other major platforms were censoring conservative views (with wide spread company leadership claiming due to pressure from Biden administration).
I regularly use X. I get blasted with political posts from both sides constantly on it.
I thought xAi previously merged with twitter, so all of this is now rolled into SpaceX? Atleast the investors in xAi and the original financiers in twitter get a breather. SpaceX is the new bandaid for this hot mess. Let’s see if this ends up rotting SpaceX or if it gets healed.
Seems like a way to put a lot of junk in space. If thats in earth orbit it will lead to a lot of junk falling from the sky in 10 years. If it all burns up that will be a lot of nasty shit in the atmosphere - millions of tons!
The really skeptical take here is that eventually all of Musk's companies merge, or at least the biggest ones, for juicing that market value to get that $1T payout. Looking at Tesla.
Related: NASA chief suggests SpaceX may be booted from moon mission [1]. Blue Origin could snatch SpaceX's Starship lander contract. This looks increasingly a good idea.
Sean Duffy is no longer acting administrator of NASA. This proposal was apparently part of a bid to get the support of a coalition of old-space companies and new-space non-SpaceX companies. As part of that strategy he apparently leaked Isaacman's Project Athena document and was backgrounding that he was a SpaceX plant.
But, Isaacman is administrator now, and whatever you think about Isaacman and his relationship to SpaceX, I don't think there's much merit in thinking one of Duffy's half thought out plans is likely to be carried out.
Sadly this seems correct. When Trump was re-elected Elon Musk pushed for Jared Isaacman to be appointed as NASA administrator. When the pick went another way, it led to some real friction between Musk and Trump. Now, with Isaacman finally at the helm of NASA, it looks like Musk’s influence over the agency has come full circle.
Kind of a bad look - but I can't precisely say why. Maybe he thinks he can raise more capital this way than he could for each company separately? Especially raising more money for X might be quite hard - they seem to be quite a bit behind on the revenue side compared to OpenAI / Anthropic. With both companies merged he might just find enough retail investors willing to buy at sky high valuations.
Doesn't the idea of Orbital Datacenters imply that the constraining resource right now is physical space, and not compute, electricity, etc?
Did we suddenly solve the electricity problem, or the compute problem?
As far as im aware there are still plenty of datacenters being planned and built right now.
They must have linked the wrong press release /s. I would have expected a press release about SpaceX acquiring xAI to talk about why they did that. Or at least mention xAI beyond the first paragraph. This is just Elon talking about space data centers
I'm not sure if this is meant to be sarcasm but is there really a need in the car market for on-demand CSAM? What actual use does Grok or any LLM have in a car?
This makes me genuinely sad. SpaceX was the one thing of his that Elon has largely avoided screwing up. Imho, this is in large part due to Gwynne Shotwell. She seems to have the personality (not to mention, personal wealth) to kick Elon in the head when he tries to mess things up.
What’s happening now is nothing more than a transparent effort to couple the AI hype-wagon to SpaceX in order to drive the valuation higher in the minds of investors who still think that LLMs will completely transform society.
I’ll be thrilled if the rocket folks can avoid being distracted by this nonsense, but I’m not optimistic.
I’ve been following SpaceX since something like the 2nd Falcon 1 launch and this is the worst thing I’ve seen happen. Sad times.
One thing to keep in mind. xAI and SpaceX both have contracts with the DoD. So it makes sense he moved it there rather than Tesla. Not sure I buy the needing AI for doing more in space or if this is to save sinking ship, but if one of his two big companies needed to buy it to keep it afloat it makes sense it was SpaceX and not Tesla.
I'm wondering if SpaceX's going public will be delayed. If not we'll see the first test of the public's appetite for what the AI companies' balance sheets look like
"Launching a constellation of a million satellites that operate as orbital data centers is a first step towards becoming a Kardashev II-level civilization"
So, basically give ourselves Kessler syndrome. Or is Elon trying to monopolize orbit entirely?
> Starship will deliver millions of tons to orbit and beyond per year
Excuse my naive physics, but is there a point at which if you take enough mass off of earth and launch it into space, it would have a measurable effect on earth's orbit? (Or if the mass is still tethered to earth via gravity, is there no net effect?)
Earth weighs about 5972200000000000000000000 kg. They are claiming to plan to launch ~5000000000 kg / year. That's 8x10^-14 %. You'd need some pretty accurate instruments to tell the difference.
Whenever computer chips go into space, they have to be hardened against radiation, because there is no atmosphere to protect them. Otherwise you get random bit flips.
This process takes a while, which is partly why all the computers in space seem out of date. Because they are.
No one is going to want to use chips that are a many years out of date or subject to random bit flips.
(Although now it got me thinking, do random bit flips matter when training a trillion parameter model?)
That's not exactly how LLM temperature works. :). Also that's on inference, not training. Presumably these would be used for training, the latency would be too high for inference.
This is terrible for Space-X. They're doing a great job. Musk has left running it to Gwynne Shotwell, who really is a rocket scientist. Now Space-X has a AI business unit they don't need, a new money drain, and more attention from Musk.
Should have merged xAI into Twitter. A failure there would not be a major setback.
Didn’t Elon say that orbital solar collection was a stupid idea due to energy loss in transmission? Using AI as an almost proof-of-work shows that it may potentially be more complex problem than previously thought. If we threw Bitcoin miners up to those satellites you could literally beam money down.
Let's call it for what it is, a payday for Elon. Paper billionaires have figured out they cannot cash out with out tanking their paper, so now you have these circular deals to extract as much as possible. If we had a functioning government they would step in and put an immediate stop to this on national security grounds.
Musk is moving value out of public hands and into his own. He overpaid $44B for Twitter, then rebranded it as an AI asset by folding X into xAI. He pushed Tesla to invest $2B of shareholder money into xAI despite shareholders voting no. Five days later, SpaceX acquired xAI, effectively turning Tesla’s cash into equity in a private company Musk owns far more of. Musk controlled every step, there was no real arm’s-length process, and he almost certainly knew the outcome in advance. Musk and his private investors get control, inflated valuations, and IPO upside. Tesla shareholders supply the cash, take the risk, and lose leverage.
I don't see the demand for space being there, OSS is driving costs down and there are still plenty of hardware and algorithmic optimizations we haven't deployed yet.
well, Musk has been overpromising and under delivering for a decade (or more?), so it seems pretty clear this too is shithousery, albeit possibly ambitious.
I suppose one of the ADR’s read something like “…who cares about bitflips, man. Isn’t AI all about probability?”
Knowing the insane level of hardening that goes into putting microcontrollers into space, how to the expect to use some 3nm process chip to stand a chance?
A trend at the moment is to just hope for the best in cubesats and other small satellites in LEO. If you’re below the radiation belt it’s apparently tenable. I worked somewhere designing satellite hardware for LEO and we simply opted to use consumer ARM hardware with a special OS with core level redundancy / consensus to manage bit flips. Obviously some problems will present for AI there… but there are arguably bigger problems with AI data centres like the fact that they offer almost no benefit with respects to the costs of putting and maintaining stuff in space!
Just checking (genuine question) there wouldn't be a sneaky way to weaponize a million satellites in orbit around the Earth, would there? I can't imagine it wouldn't have ever been looked into.
I have so many conflicting thoughts that I cannot properly articulate yet. I can say though, this is not going to end well for most, it is clumsily premeditated and starting to feel like dude is just trying to be a Neal Stephenson character.
I will not be left holding this bag. This is such financial engineering nonsense, and if we had any sort of regulatory controls this would never be allowed to happen - especially BECAUSE of national security reasons.
SpaceX launches about 60% of National Security Space Launch payloads. The only other active launcher cleared for these missions is ULA's Vulcan Centaur, which has launched 3 times total, ~once every 8 months.
Do this math include the cost and weight of the radiators? Because it obviously can't work without big radiators, and I don't see them mentioned in the math?
I think there is one more possible framing for this.
Recently xAI has been in the news for Groq's revenge-porn-like "undress them" feature, which seems pretty legally questionable.
Musk has also been in the news for his own Epstein-related activities.
If he can move Groq and X into space, well, there's not very many age-of-consent or revenge-porn laws in space as far as I know, so maybe he'll be able to do some sort of legal leverage where the space data-center can produce otherwise legally questionable AI responses with impunity.
This makes a lot of sense. The commercial launch business is not large enough to support all possible Falcon launches, so Starlink was created to take advantage of the low launch cost and vertical integration and is now a major profit center for SpaceX.
Starship launches are only going to make sense every 779.94 days (the approx 2 year Mars-Earth proximity). The rest of the time, the launches could similarly be used to deploy orbiting data centers for XAi/Grok etc. Brilliant move.
I really would like to see a cost and cooling breakdown. I just can't see how you can do radiative cooling on the scales required, not to mention hardening.
I thought this was a troll by Elon, now I'm leaning towards not. I don't see how whatever you build being dramatically faster and cheaper to do on land, even 100% grid independent with solar and battery. Even if the launch cost was just fuel, everything else that goes into putting data centers in space dwarfs the cost of 4x solar plus battery.
I cannot really tell satire apart from genuine opinions anymore.
(But I do hope it was satire, if not, cooling satelites was/is a big issue and they only have very modest heat creation. A data center would be in a quite different ballpark)
Pre-IPO price padding. xAI is going nowhere but at least for now it has some value. Move it under SpaceX, bump up SpaceX’s valuation and therefore it’s opening IPO price. Then kill xAI and write it off.
I know that per HN's guidelines we're supposed to be "kind and curious", and "reply to the argument instead of calling names". But with some texts, engaging with individual arguments loses sight of the more important bigger picture. So while unkind, the most "thoughtful and substantive" thing I think can be said about this text is:
If Musk and SpaceX are serious about putting 1 million datacenter satellites into space, then they are not serious about Mars.
You cannot simultaneously build and launch 10’s of thousands of Starships to deliver 1 million tons of equipment and supplies bound to Mars while also committing to launching 10’s of thousands of Starships to orbit full of satellites.
They would need to quadruple their launch rate, and half of those launches would be Starships bound for Mars, the vast majority of which would never return.
How many Falcon9’s have ever been built? It is incredible to say you can build that many rockets and use up that much fuel on any reasonable time scale. You might as well say the Tesla Roadster version 3 will be a Single Stage to Orbit rocket car.
Friendly reminder for anyone that forgot - xAI acquired Twitter, so now Space-X is the proud owner of a dying social media platform that they overpaid for.
Any claims that this is about putting compute in space is just a non-sense distraction. This was absolutely about bailing Elon out of his impulsive, drug-fueled Twitter purchase.
The only question now is: when they try to go public, will they be punished for wasting so much money or not? My guess is: not.
Musk’s schtick lately has always been to find a challenging problem, point everyone to it and say “ I will solve it by September “, have the stock shoot up and make money. His first dis that with self driving, then twitter, then xAi, and now robots and data center in space. These last two will last him a decade as these are both challenging problems to solve over night.
So they use a valid and valuable company to hide a giant dumpster fire company. To add to that, their best argument is "AI in space", which has some real "solar roadways" energy to it. I honestly don't know how any SpaceX shareholder could approve this.
We just had an X8.1 CME event, I just want to point out that at any moment we could have an x40 (we had a carrington event already in 1859) or higher event and all those sats at low earth orbit would be fried and start hitting each other, if SpaceX keeps launching more it becomes incredible probable that we might hit the Kessler Syndrome, and we would legit lose access to Space for a WHILE, including all of what satellites entail.
Are we ready for that as a modern society or are we going to start enacting regulation against it? I'm sorry but people wanting internet everywhere does not justify we going back to the dark ages for a decade or more.
US is dialing back even on global warming. There is no chance current government would risk space superiority for some kessler-shmessler that nobody has ever seen.
Now he just needs to work in crypto satellites down to users via all the new phones supporting satellite link to SpaceX. I kinda expected that one first. Distributed payment network outside of government control/oversight seems like something he would be in to.
Pretty terrible for SpaceX. Of course they paid a crazy inflated price for xAI in an attempt to cash in on the IPO. This just devalues SpaceX and exposes the investors to all the AI bubble risk.
Tesla has probably the most valuable shareholders on Earth. Over years of empty promises and meme status, the stock has pretty much purged all the level heads. So it's mostly deluded Elon sycophants giving placing their tithe on the alter of his sci-fi fantasy smoke and mirrors game.
In reality he will be dumping the debt of twitter and xAI (and maybe spacex?) on Tesla shareholders, and buoying that with the added layer of hyper that spaceX brings.
> SpaceX has acquired xAI to form the most ambitious, vertically-integrated innovation engine on (and off) Earth, with AI, rockets, space-based internet, direct-to-mobile device communications and the world’s foremost real-time information and free speech platform. This marks not just the next chapter, but the next book in SpaceX and xAI's mission: scaling to make a sentient sun to understand the Universe and extend the light of consciousness to the stars!
Man the real story in a few decades is going to be whether SpaceX was onto such a kille business that it survived being used asma fiscal dumping ground for the losses being incurred by mismanagement everywhere else.
Reminder that space only allows for radiative cooling (since there is no air to absorb heat) so data centers in space are going to have massive cooling panels.
Reminder that SpaceX has received an estimated $38 billion in government funding over the years, and all of its returns are going to a small set of private investors.
Socialized losses, privatized profits. As is the American way.
Hyperloop was never a company project, neuralink was a separate company, tesla is rolling out driverless robotaxis and fsd is amazing, robots inevitably are going to do the majority of work - there’s no real doubt about it is there?
> robots inevitably are going to do the majority of work - there’s no real doubt about it is there?
There's a lot of doubt that the AI and compute to enable that would happen on commercially relevant timescales.
Consider: "do the majority of work" is a strict superset of "get into car and drive it". The power envelope available for an android is much smaller than a car, and the recently observed rate of improvements for compute hardware efficiency says this will take 16-18 years to bridge that gap; that plus algorithmic efficiency improvements still requires a decade between "car that can drive itself" and "android that can drive a car". (For any given standard of driving).
And that's a decade gap even if it only had to do drive a car and no other labour.
You can't get around this (for an economy-wide significant number of androids) by moving the compute to a box plugged into the mains, for the same reason everyone's current getting upset about the effect of data centres on their electricity bills.
And note that I'm talking about a gap between them, not a time from today. Tesla's car-driving AI still has safety drivers/the owner in the driving seat, it is not a level 4 system. For all that there are anecdotes about certain routes and locations where it works well, there's a lot of others where it fails.
That said: Remote control units without much AI are still economically useful, e.g. a factory in Texas is staffed entirely by robots operated over a Starlink connection by a much cheaper team in Nairobi.
Major grift vibes, inventing half-baked reasoning to justify massive valuations.
If money wasnt so deeply entwined with politics at this point, this is the sort of news that would launch fraud investigations.
The way I read it, X AI is not really profitable and Elon's creditors/coinvestors asked for something tangible for their money, a.k.a shares in SpaceX, his only business that still has some solid foundation. The rest is emois.
When Elon Musk talks about benefiting humanity, remember that the one time he had unregulated power (DOGE) he used it mostly to cut benefits for poor people, and to push ideological agendas. His only agenda is self-aggrandizement, and this announcement is cover for passing the hot potato of Twitter debt around.
Whenever Elon, Anthropic or whoever mentions humanity, they don’t mean every 8 billion of us on Earth. They mean themselves, their mates and whoever they deem worthy of the posterity they believe only they have the merit to design.
xAI to cover X investors, SpaceX to cover xAI, us American public investors to cover SpaceX since it cannot go under for strategic reasons. Ultimate grift.
I thought this wasn't viable due to cooling requirements - how do you cool massive amounts of compute when the only option is to radiate it into space - nothing to convect it with?
Also, the incredible amount of grift here with the left hand paying the right is scarcely believable. Same story as Tesla buying Solarcity. Board of directors should be ashamed IMO.
Yes. It is very cold up there but there is also no matter, or very little matter. So head conduction and convection don't work, it's all radiation. When we are learning to solve heat transfer problems in engineering school we are generally taught to neglect radiation, because it's effect on cooling the system is typically second or third order when compared to the to "big C's"
Elon Musk is a genius, but he’s a financing genius. Look at the long history he has of false promises supporting financing deals between his companies and you’ll see this for what it is, a cash injection and a lie to justify it. He did the same thing with a fake solar roof demo when Tesla bought the almost bankrupt Solar City. He also shifted resources from Tesla and SpaceX to support X in the early days. Even founding xAI outside of Tesla, when so much of its valuation was built on its AI capabilities, was questionable.
Making a "sentient sun" is the most bald-faced asinine, drug-induced nonsense that should be the complete destruction of all credibility to anyone who said it or typed it or works for anything here.
Financial engineering. Twitter under Elon became a dumpster fire of porn and hate and big banks were holding 13B in bonds that wouldn’t be worth the paper they were printed on for the company alone so he just links it with his only company that actually is doing something worthwhile…
Not sure how X which “merged” wit X (formerly Twitter) and SpaceX really matter or synergize but here we are. It’s all about the money being protected. And this Ketamine using wierdo is gonna be the worlds first trillionaire. Yay all of us.
> By directly harnessing near-constant solar power with little operating or maintenance costs, these satellites will transform our ability to scale compute. It’s always sunny in space! Launching a constellation of a million satellites that operate as orbital data centers is a first step towards becoming a Kardashev II-level civilization, one that can harness the Sun’s full power, while supporting AI-driven applications for billions of people today and ensuring humanity’s multi-planetary future.
Apparently optimus robots don't work and he needs to start his final grift, space datacenters, while his datacenters on earth are powered by gas turbines.
Most likely he's just trying to bury his epstein involvement where was exposed lying by his own daughter.
Can someone convince me that this is not a) pure horseshit b) a plan for Elon to sneak enough mass into orbit to hold the Earth hostage?
If you can bring millions of tons of anything into orbit around Earth you can destroy civilization, or just France.
When does the market realize this is all just a shell game and the emperor really has no clothes?
We saw this on a much smalelr scale a decade ago when one of Elon's companies (Tesla) acquired a second one of Elon's companies (SolarCity) because it was broke and owed a ton of money to a third one of Elon's companies (SpaceX).
Elon was forced to go through with his impulsive Twitter acquisition by a Delaware court, an acquisition that was not only secured by a bunch of Tesla stock but also a bunch of Qatari and Saudi royal money. He then mismanaged Twitter so badly Fidelity wrote down its value by at least 80% [1].
So what did Elon do? Raised even more questionable foreign money into xAI, diverted GPUs intended for another of his companies (Tesla) into Twitter and then "merged" Twitter into xAI, effectively using other people's money to bail him out from an inevitable margin call on his Tesla stock.
Interestingly, Twitter was reportedly valued at $33 billion in this deal [2], significantly more than the less than $10 billion Fidelity valued Twitter at. Weird, huh? With a competent government, this would be securities fraud that would have you spend the rest of your life in jail. And even with all that, $11 billion was lost on the deal.
So here we are and it's time for the shell game to be played again. Now it's SpaceX's turn to bail out the xAI investors.
And what is the argument for all this? AI data centers in space. Words cannot describe how little sense this makes. Launch costs (even if the Starship launch costs get to their rosy projections), cooling in space, cosmic rays (and the resulting errors) and maintenance. Servers constantly need parts replaced. You can just deorbit the satellite instead but that seems like an expensive way of dealing with a bad SSD or RAM chip.
> This marks not just the next chapter, but the next book in SpaceX and xAI's mission: scaling to make a sentient sun to understand the Universe and extend the light of consciousness to the stars!
Slightly less high Musk aide: "But what is the synergy, where's the moat and how could that be done in practice and most importantly is there any limiting factor on Earth before we have to bring AI into.."
Musk : "SPACE!!!!!!"
It is incredible to think that the extremes of the stock market are actually pretty similar, pink sheets/cryptos and these mega companies are actually the same. News fueled pumps and dumps to win the cycle of hype of the week
This means that Grok, Elon's politically labotomized involuntary pornography generator, and X (formerly Twitter), Elon's Nazi-adjacent propaganda machine, are now completely intertwined with SpaceX, a too-big-to-fail government contractor that currently serves as America's only reliable option for manned orbital spaceflight.
Anyone who doesn't see how broken this situation is isn't paying attention. This is how people like Elon, who want to seize as much power from the government as they can, ensure that the means for seizing that power are untouchable.
Anyone who has ever used Grok or X lately knows that both of these products are heavily manipulated to align with the political, social, and economic views of Elon Musk, who is increasingly boosting "white power" language and full-throatedly backing America's most nationalistic and authoritarian president to date.
This is just another consolidation of power, and it's deeply worrisome. Any integrity one may have hoped remained at SpaceX just vanished when they aligned their mission with that of these deeply problematic digital services.
And this is not even scratching the surface of what looks like a deliberate attempt to create Kessler syndrome by launching millions of cheap short-term satellites into orbit, or the rationality of putting datacenters into orbit in the first place...
There's an epic conflict of interest here with Musk owning most of both companies. And they're in entirely separate fields, there is no plausible synergy here to be gained.
Musk has a history of having one of his more successful companies buying one of his less successful companies. xAI bought X, and Tesla bought SolarCity
Musk is notorious for shuffling assets across his companies to make some financials look better. For example, shuffling Twitter servers (and then all of Twitter) under xAI.
What's funny? Do you think the investors are against this? The investor's aren't idiots. I imagine the typical investor in Elon Musk's companies would approve of this sort of thing. So what's the problem? Besides, its a private company with Musk as majority shareholder in both. That's the beauty of private companies, you can just do things.
I wish more companies were private and ambitious. I'm tired of companies like Apple making marginal spec bumps to their phones and milking the same products for decades
The investors want to cash out, Musk needs lots of money to plow into his latest toy that so far only excels in ridiculing him and sexual harassment/CSAM, so they make a deal to take in xAI and go public. Win win.
It's widely reported that Musk is a majority shareholder of xAI and the controlling shareholder of SpaceX (close to 80% of voting shares). Not surprising that he would be looking to consolidate ownership under one entity especially if he perceives significant synergies (i.e., data centres in space).
I have never understood how Data centers in space ever make economic sense, the payload, latency and many other issues make it difficult at least for the immediate needs
Do you have any parties planned? I’ve been working to the edge of sanity this year and so, once my kids head home after Christmas, I really want to hit the party scene in St Barts or elsewhere and let loose.
wow this is one of the most incredibly fraudulent things that ever happened in American capitalism and I'm not ignoring Enron or the mortgage meltdown. I'm speechless the US has given up on any semblance of law and order in matter of financial markets and this stuff can happen without people going to jail.
People pooh-poohing space datacenters will obviously think this is a bad move. But Elon clearly believes space datacenters will work. Given that, and the fact that SpaceX will IPO this year, this acquisition was inevitable.
SpaceX and xAI would not be able to freely collaborate on space datacenters after the IPO because it would be self-dealing. SpaceX likes to be vertically integrated, so they wouldn't want to just be a contractor for OpenAI's or Anthropic's infrastructure. Merging before the IPO is the only way that SpaceX could remain vertically integrated as they build space datacenters.
> it is possible to put 500 to 1000 TW/year of AI satellites into deep space, meaningfully ascend the Kardashev scale and harness a non-trivial percentage of the Sun’s power
We currently make around 1 TW of photovoltaic cells per year, globally. The proposal here is to launch that much to space every 9 hours, complete with attached computers, continuously, from the moon.
edit: Also, this would capture a very trivial percentage of the Sun's power. A few trillionths per year.
We also shouldn't overlook the fact that the proposal entirely glosses over the implication of the alternative benefits we might realize if humanity achieved the incredible engineering and technical capacity necessary to make this version of space AI happen.
Think about it. Elon conjures up a vision of the future where we've managed to increase our solar cell manufacturing capacity by two whole orders of magnitude and have the space launch capability for all of it along with tons and tons of other stuff and the best he comes up with is...GPUs in orbit?
This is essentially the superhero gadget technology problem, where comic books and movies gloss over the the civilization changing implications of some technology the hero invents to punch bad guys harder. Don't get me wrong, the idea of orbiting data centers is kind of cool if we can pull it off. But being able to pull if off implies an ability to do a lot more interesting things. The problem is that this is both wildly overambitious and somehow incredibly myopic at the same time.
Yeah it does not make a whole lot of sense as the useful lifespan of the gpus in 4-6 years. Sooo what happens when you need to upgrade or repair?
1 reply →
So what are the other things? You said he glossed over them and didn't mention a single one.
2 replies →
Honestly, there's not a lot else I can think of if your goal is find some practical and profitable way to take advantage of relatively cheap access to near-Earth space. Communication is a big one, but Starlink is already doing that.
One of the things space has going for it is abundant cheap energy in the form of solar power. What can you do with megawatts of power in space though? What would you do with it? People have thought about beaming it back to Earth, but you'd take a big efficiency hit.
AI training needs lots of power, and it's not latency sensitive. That makes it a good candidate for space-based compute.
I'm willing to believe it's the best low-hanging fruit at the moment. You don't need any major technological advances to build a proof-of-concept. Whether it's possible for this to work well enough that it's actually cheaper than an equivalent terrestrial datacenter now or in the near future is something I can't answer.
3 replies →
This is such a hypebeast paragraph.
Datacenters in space are a TERRIBLE idea.
Figure out how to get rid of the waste heat and get back to me.
11 replies →
[dead]
You really can't grasp that GPUs scaled at this level is the most ambitious thing possible? That it will be the foundation of unfathomable technological innovation?
3 replies →
All right, so how is it that all you geniuses out here are totally right about this, but all the dullards at SpaceX and XAI, who have accomplished nothing compared to you lot, are somehow wrong about what they do every day?
I know being right without responsibility feels amazing but results are a brutal filter.
6 replies →
Only people who never interacted with data center reliability think it's doable to maintain servers with no human intervention.
Whoa there, space-faring sysadmin. You really want that off-world contract tho?
1 reply →
Do they need to be maintained? If one compute node breaks, you just turn it off and don't worry about it. You just assume you'll have some amount of unrecoverable errors and build that into the cost/benefit analysis. As long as failures are in line with projections, it's baked in as a cost of doing business.
The idea itself may be sound, though that's unrelated to the question of whether Elon Musk can be relied on to be honest with investors about what their real failure projections and cost estimates are and whether it actually makes financial sense to do this now or in the near future.
1 reply →
There are a class of people who may seem smart until they start talking about a subject you know about. Hank Green is a great example of this.
For many on HN, Elon buying Twitter was a wake up call because he suddenly started talking about software and servers and data centers and reliability and a ton of people with experience with those things were like "oh... this guy's an idiot".
Data centers in space are exactly like this. Your comment (correctly) alludes to this.
Companies like Google, Meta, Amazon and Microsoft all have so many servers that parts are failing constantly. They fail so often on large scales that it's expected things like a hard drive will fail while a single job might be running.
So all of these companies build systems to detect failures, disable running on that node until it's fixed, alerting someone to what the problem is and then bringing the node back online once the problem it's addressed. Everything will fail. Hard drives, RAM, CPUs, GPUs, SSDs, power supplies, fans, NICs, cables, etc.
So all data centers will have a number of technicians who are constantly fixing problems. IIRC Google's ratio tended to be about 10,000 servers per technician. Good technicians could handle higher ratios. When a node goes offline it's not clear why. Techs would take known good parts and basically replacce all of them and then figure out what the problem is later, dispose of any bad parts and put tested good parts into the pool of known good parts for a later incident.
Data centers in space lose all of this ability. So if you have a large number of orbital servers, they're going to be failing constantly with no ability to fix them. You can really only deorbit them and replace them and that gets real expensive.
Electronics and chips on satellites also aren't consumer grade. They're not even enterprise grade. They're orders of magnitude more reliable than that because they have to deal with error correction terrestial components don't due to cosmic rays and the solar wind. That's why they're a fraction of the power of something you can buy from Amazon but they cost 1000x as much. Because they need to last years and not fail, something no home computer or data center server has to deal with.
Put it this way, a hardened satellite or probe CPU is like paying $1 million for a Raspberry Pi.
And anybody who has dealt with data centers knows this.
6 replies →
But … but what if we had solar-powered AI SREs to fix the solar-powered AI satellites… /in space/?
5 replies →
I couldn't believe that was an actual quote from the article. It is.
These people are legit insane.
Not insane at all. They are perfectly sane and know words can be twisted to justify just about anything, when stating the actual goals is unsavory.
No it's just Musk's Big Idea for spacex to hype it before IPO. It's their version of FSD, robots etc.
You've got to hand it to him, he is a bullshitter par excellence.
> We currently make around 1 TW of photovoltaic cells per year, globally.
Doubling every three years; at that rate it would take about 30 years for 1TW to become 1000TW. Whether on not the trend continues largely depends on demand, but as of right now humanity seems to have an insatiable demand for power.
I think it largely depends on what bottlenecks exist that we haven’t hit yet.
We’re not going to use 100% of our solar panel manufacturing capacity to power space data centers, specifically because everyone else on the ground is so power-hungry. If we’re being generous, it could maybe top out at 1%, which adds another ~20 years to your timeline for a total of 50. I think it’s safe to say this part is bunk (along with everything else about this plan which is also bunk).
In fairness, solar cells can be about 5x more efficient in space (irradiance, uptime).
The quoted "1 TW of photovoltaic cells per year, globally" is the peak output, not the average output. They're only about 20% higher peak output in space… well, if you can keep them cool at least.
7 replies →
It is more than 5x less expensive to get surface area on earth’s surface.
16 replies →
Does that include all the required radiators to vent heat?
2 replies →
Fortunately there are no downsides to launching solar cells into space that would offset those gains.
Now do waste heat.
Solar cells have exactly the same power rating on earth as in space surely? What would change is their capacity factor and so energy generation.
18 replies →
And how much of that power would be spent on high speed communications with Earth that aren't, you know, a megabit or two per second
8 replies →
Doesn't this risk some unforeseen effects on Earth or the rest of the solar system at that scale? Disruption of magnetic shield, some not yet known law of physics suddenly getting felt etc.?
It's not really going to happen so we don't have to worry about that.
> We currently make around 1 TW of photovoltaic cells per year, globally.
China made 1.8 TW of solar cells in 2025.
The raw materials required to make these are incredibly abundant, we make as much as we need.
you realize the factor of 2 you introduce doesn't meaningfully change the order of magnitude that the previous poster is implying right?
5 replies →
This is all based on bad math. The people proposing these things don't even have proper scientific and mathematical training to determine what is achievable.
Context missing. This is in reference to a vision the (distant?) future where the satellites are manufactured in factories on the Moon and sent into space with mass drivers.
Full paragraph quote comes from:
> While launching AI satellites from Earth is the immediate focus, Starship’s capabilities will also enable operations on other worlds. Thanks to advancements like in-space propellant transfer, Starship will be capable of landing massive amounts of cargo on the Moon. Once there, it will be possible to establish a permanent presence for scientific and manufacturing pursuits. Factories on the Moon can take advantage of lunar resources to manufacture satellites and deploy them further into space. By using an electromagnetic mass driver and lunar manufacturing, it is possible to put 500 to 1000 TW/year of AI satellites into deep space, meaningfully ascend the Kardashev scale and harness a non-trivial percentage of the Sun’s power. >
> This is in reference to a vision the (distant?) future where the satellites are manufactured in factories on the Moon and sent into space with mass drivers.
In the meantime, how about affordable insulin for everybody?
Why is it cheaper to ship all of the materials to space, then to the moon for assembly (which also includes shipping all of the people and supplies to keep them alive), then back into space vs just…
building them on earth and then shipping them up?
We’re not exactly at a loss for land over here.
3 replies →
Why would satellites be manufactured on the moon? There's nothing on the moon. The raw materials would have to be ferried over first. What would be the point?
5 replies →
Is there a credible way to cool a space-based data center on that scale?
We will have cyber taxis and FSD 100% next year.
See Dyson Sphere: https://en.wikipedia.org/wiki/Dyson_sphere
Dyson's paper was literally written in jest.
4 replies →
Yeah, that's the point ... it's stupid to believe humanity is capable of deploying that much infrastructure. We cannot do even 0.01% of it.
22 replies →
Pfft that would just require setting up an entire lunar mineral extraction and refining system larger than we have on earth, just minor details.
Help me understand something. We make 1 TW of cells per year but we're struggling with bringing 1 GW consuming data centers online?
Nameplate capacity needs a derate for availability, so you can drop it down to about 200GW(e) equivalent continuous power assuming we're making and deploying enough batteries to support it. More, obviously, if those panels are going to an equatorial desert, less if they're going to sunny Svalbard in the winter time.
[flagged]
> And not delivering products
2024 revenue of >$100b is quite impressive for not delivering any products
4 replies →
At best, he should be a persona non grata across just about every aspect of society.
Even if you discount all the Nazi crap, he's directly responsible for deaths of 600,000+ people, mostly kids, for his illegal destruction of USAID.
What a tremendous failure it is that this guy is still allowed such a prominent place in society.
:eye roll:
This schtick is so, so tiresome.
1 reply →
A former NASA engineer with a PhD in space electronics who later worked at Google for 10 years wrote an article about why datacenters in space are very technically challenging:
https://taranis.ie/datacenters-in-space-are-a-terrible-horri...
I don't have any specialized knowledge of the physics but I saw an article suggesting the real reason for the push to build them in space is to hedge against political pushback preventing construction on Earth.
I can't find the original article but here is one about datacenter pushback:
https://www.bloomberg.com/opinion/articles/2025-08-20/ai-and...
But even if political pushback on Earth is the real reason, it still seems datacenters in space are extremely technically challenging/impossible to build.
It's much easier to find a country or jurisdiction that doesn't care about a bunch of data centers vs launching them into space.
I don't get why we aren't building mixed use buildings, maybe the first floor can be retail and restaurants, the next two floors can be data centers, and then above that apartments.
We don’t even have a habitable structure in space when the ISS falls, there is no world in which space datacenters are a thing in the next 10, I’d argue even 30 years. People really need to ground themselves in reality.
Edit: okay Tiangong - but that is not a data center.
Who is “we”? https://en.wikipedia.org/wiki/Tiangong_space_station
1 reply →
I don't think any of the companies that say they are working on space data centers intend them to be habitable.
> We don’t even have a habitable structure in space
Silicon is way more forgiving than biology. This isn’t an argument for this proposal. But there is no technical connection between humans in space and data centers other than launch-cost synergies.
2 replies →
I don't know, 10 years seems reasonable for development. There's not that much new technology that needs to be developed. Cooling and communications would just require minor changes to existing designs. Other systems may be able to be lifted wholesale with minimal integration. I think if there were obstacles to building data centers on the ground then we might see them in orbit within the next ten years.
I don't see those obstacles appearing though.
2 replies →
Ok then short SpaceX stock when it IPOs.
6 replies →
> A former NASA engineer with a PhD in space electronics who later worked at Google for 10 years wrote an article about why datacenters in space are very technically challenging
It's curious that we live in a world in which I think the majority of people somehow think this ISN'T complicated.
Like, have we long since reached the point where technology is suitably advanced to average people that it seems like magic, where people can almost literally propose companies that just "conjure magic" and the average person thinks that's reasonable?
you can’t tell me the microwave isn’t magic. it’s magic.
It's not just "very challenging", it's "very challenging and also solves no actual problem we face".
It's not like launching stuff into space doesn't have pushback, either. See: starlink satellites.
Yeah but he’s an expert his opinion can be dismissed bro this is 2026
"Technically challenging", a nice way to say "impossible"
Just like rockets landing themselves
15 replies →
> It(Solar) works, but it isn't somehow magically better than installing solar panels on the ground
Umm, if this is the point, I don't know whether to take rest of author's arguments seriously. Solar only works certain time of the day and certain period of year on land.
Also there is so limited calculations for the numbers in the article, while the article throws of numbers left and right.
No one is interested in excuses on why it can't be done. Were in interested in the plan on how they plan to do it.
The guy is saying satellite communication is restricted to 1Gbps ffs. SpaceX is way past that.
SpaceX is too big to fail. It's important for national security.
I wonder if Elon wants to tangle all his businesses into SpaceX so they are all kept afloat by SpaceX's importance.
Elon can't legally financially entangle Tesla to SpaceX due to Tesla being a public company, so his hands are tied.
Tesla is clearly benefiting from protectionism and its sales would collapse if BYD were allowed to openly sell in the US. Most people just want affordable, maintainable and reliable cars.
> Elon can't legally financially entangle Tesla to SpaceX due to Tesla being a public company, so his hands are tied.
He absolutely could do it, just like he did when Tesla bought SolarCity. It just isn’t as easy when one of the companies is public than when both are private.
1 reply →
I’m old enough to remember when this was said about Solar City
1 reply →
He's broken pretty much all the other financial rules.... for example, the amount of blatant self-dealing he gets away with is staggering.
As long as the consequences of his actions continue to increase the paper value for investors, regulations don't really have teeth because there aren't damages. So the snowball gets bigger and the process repeats.
> Elon can't legally financially entangle Tesla to SpaceX
Bill Ackman has proposed taking SpaceX public by merging it with his Pershing Square SPARC Holdings, distributing 0.5 Special Purpose Acquisition Rights (SPARs) to Tesla shareholders for each share held. Each SPAR would be exercisable for two shares of SpaceX, aimed at enabling a 100% common stock capitalization without traditional underwriting fees or dilutive warrants.
With SpaceX IPO set to be one of the biggest of all time, this could have a pretty gnarly financial engineering impact on both companies -- especially if the short interest (direct or through derivatives) remains large.
19 replies →
It's "ironic?" considering Tesla launching in China is what created the necessary supply chain to turn BYD into the powerhouse it is today. Tesla's greed will become their own demise.
1 reply →
BYD are just affordable and maybe reliable, regarding maintenance their spares are hard to come by and are almost as hard to work with as Tesla and other brands.
4 replies →
[dead]
[Nearly] all is possible when you have a board of simps/cultists
Starship has a large number of critical milestones coming: Can it land and quickly reuse the upper stage? If not, it can't make refueling flights without building a dozen or two starships. Can it carry the full specified payload? If not, it can't even try to refuel in orbit. If it can't refuel in orbit, it can't go beyond earth orbit. Etc.
Everything has to go right or it will be irrelevant before it works.
> Everything has to go right or it will be irrelevant before it works.
Starship is not all of SpaceX. Saying, maybe because one hates Musk, that SpaceX is going to become irrelevant is wishful thinking.
In 2025 SpaceX launched more rockets into space than the entire world ever sent in a year up to 2022, something crazy like that.
Then out of, what, 14 000 active satellites in space more than half have been launched by SpaceX.
SpaceX is, so far, the biggest space success story of the history of the human race (and GP is right in saying that SpaceX is now a national security matter for the US).
14 replies →
Of course it isn't "too big to fail". Even banks aren't. Despite recent history large banks have failed often throughout history. There's no such thing. It may take down the supporting sovereign government (Dutch East Indies) but life goes on and new political orgs appear. People be people.
Too big to fail is a very recent modern myth. Go back 100+ years and lots of banks failed leading into the Great Depression.
Every system has a break point.
I think he will spin Tesla off since electrification and autonomy are no longer cool (he can’t build good quality cars or reliable FSD)
Haven't you heard? Tesla is pivoting to building humanoid robots instead. They haven't sold a single one, but it toootally warrants retooling their car factories, pinky promise!
FSD is incredibly reliable. Build quality of US built cars is middle of the pack, Europe/China built Teslas are top of the pack.
17 replies →
How vital is it really to national security? Starlink will have competition from Amazon Leo in the next few months. And while SpaceX is obviously in the lead in launch capability with Starship, there are multiple launch providers capable of providing roughly the same services the Falcon 9 and Heavy provide today.
> SpaceX is too big to fail. It's important for national security
So was GM. Didn’t stop it from going bankrupt.
SpaceX is slated to go public some time this year - June IIRC
The biggest selling point /was/ that Musk was being managed there, he wasn't tinkering with SpaceX like Twitter or Tesla, and his foolhardy direction was kept out of the company.
BUT, like Tesla, Musk cannot help himself and is making SpaceX look like a very bad investment - tying his other interests with SpaceX, allegedly using SpaceX money as a "war chest" in his battles.
There is also a danger that investors will see xAI as politically dangerous, which will really hurt SpaceX IPO
They want to go public, but have to sell the hell out of it in the meantime.
I'll bet SpaceX financials aren't as great as some people think. Remember, Elon was the guy who tried to take Tesla private, and talked a lot of smack about how silly it is to be a public company. All of a sudden he wants SpaceX to go public?
1 reply →
> The biggest selling point /was/ that Musk was being managed there, he wasn't tinkering with SpaceX like Twitter or Tesla, and his foolhardy direction was kept out of the company
The biggest selling point to who? Definitely not wall street
And our tax dollars.
Let’s be honest - this is just a way to prop up Twitter/X. It makes SpaceX shareholders subsidize X, and also American taxpayers who are giving contracts to SpaceX for highly sensitive things. The government should ideally refuse to give SpaceX work unless it unwinds this.
Why? The government is paying less for SpaceX than alternatives. It th cheapest and best service.
10 replies →
Why? Let it fail. Bring back NASA.
What do you mean "let it fail?" SpaceX has the most profitable launch system in the world and now operates >50% of all satellites in orbit. They aren't exactly in need of a bailout.
3 replies →
> Bring back NASA.
NASA is still here. Unfortunately, NATA fell victim to enshitification by government contracting. NASA even if it wants to simply cannot today design and launch a rocket. :(
I am no fan of Musk the man. SpaceX is a strong company and Falcon is a solid vehicle. There is not a lot of competition, and NASA trying to in-source design and supply and construction of a new, reusable LEO rocket would be a complete nightmare.
I root for a competitive rocket market, but SpaceX is at the moment critical.
You want to bring back the biggest loser? NASA kept missing deadlines for 30 years
1 reply →
NASA just splurges money. The private sector is far better when it comes to money.
11 replies →
When you’re connected to Epstein, you’ll always be too big to fail
national security is pretty felixaeble
Being too big to fail is not really a desirable outcome, it’s just better than failure.
Boeing is too important to fail as well but it’s been terrible as a shareholder
> > Boeing is too important to fail as well but it’s been terrible as a shareholder
Your opinion on Boeing being terrible as a shareholder vis-a-vis Tesla would be completely reversed if dividends and capital gains of the 2 companies were to be offered in the form of miles to be flown on Boeing planes and miles on Teslas Uber/Taxi/Autonomous taxis instead of dollars
The absolute overperformance on the stock market that Tesla has enjoyed vis-a-vis Boeing is not rooted in a concrete and tangible quality of life improvement for citizens. Not American citizens, nor global citizens for that matter.
It is my opinion that for all public companies in which it is possible to do so government should mandate payment in kind to all shareholders and board members to prevent the excessive promotional , cult and all around BS aspect of marketing to take over and allow people to profit just by riding off those, and Musk is the GOAT at that.
1 reply →
Merging SpaceX with a public company like Tesla would create a lot of issues for the classified projects SpaceX does.
What sort of issues are you thinking?
Plenty of defense contractors with classified projects are already publicly listed, so this is not uncharted territory.
Lockhead Martin for example: https://investors.lockheedmartin.com/news-releases/news-rele...
Gives this level of detail:
> Aeronautics classified program losses $(950)
> MFC classified program losses -
It seems very safe from a national security perspective.
No? Almost every big defense contractor is publicly traded.
I imagine those are surmountable challenges. Boeing somehow manages.
But more likely that merger would consist of SpaceX acquiring Tesla and taking it private
1 reply →
Raytheon is public.
As a SpaceX fan, I am saddened by this news.
The only reason for xAI to join SpaceX is to offload Elon's Twitter debt in the upcoming IPO.
This is just what I was thinking.
Twitter (X) owed $1.3B in debt every year in interest since Musk's takeover. This was before re-financing in a higher interest rate environment. The company was losing $200MM+ per year on ~$5B in revenue before the takeover, and there are reports that revenues have decreased by round 50%.
Best case scenario if we accept those numbers is that X makes $3B per year and about half of that goes immediately out the door in debt payments before paying a cent for the entire business to function.
However, if SpaceX acquires X, that ~$1.5B in interest is a fraction of the $8B In profits SpaceX is allegedly generating annually. Further, they can restructure the debt if it's SpaceX's debt, and not owned by X. Investors will be more likely to accept SpaceX shares as collateral than X.
> The company was losing $200MM+ per year on ~$5B in revenue before the takeover, and there are reports that revenues have decreased by round 50%.
X made a profit last year because they cut costs lower than the drop in ad revenue (which is also slowly recovering). The big question is if they will still be profitable in 2026 year without the US election driving big traffic numbers and ads.
4 replies →
As an old twitter fan, same. I was hoping elon would lose twitter to the banks in a bankruptcy.
Either this is a straight up con, or Musk found a glitch in physics. It's extremely difficult to keep things cold in space.
He buys twitter at an inflated valuation. Runs it to the ground to a much lower valuation of $9B. [1] Then, his company Xai buys Twitter at a $33B, inflating the valuation up. Then SpaceX merges with Xai for no particular reason, but is expected to IPO at a $1T+ in the upcoming years. [3]
I’m not that smart, but if I were, I would be thinking this is an extended way to move the losses from the Twitter purchase on to the public markets.
[1] https://www.axios.com/2023/12/31/elon-musks-x-fidelity-valua...
[2] https://www.reuters.com/markets/deals/musks-xai-buys-social-...
[3] https://www.cnbc.com/amp/2026/02/02/elon-musk-spacex-xai-ipo...
It also makes it impossible for Twitter/X to die, as it deserves. It is by far the most toxic mainstream social network. It has an overwhelming amount of far right supremacist content. So bad that it literally resulted in Vivek Ramaswamy, a gubernatorial candidate in Ohio, to quit Twitter/X - nearly 100% of replies to his posts were from far right racists.
Obviously advertisers have not been fans. And it is a dying business. But rather than it dying, Elon has found a clever (and probably illegal) way to make it so that SpaceX, which has national security importance, is going to prop up Twitter/X. Now our taxpayer dollars are paying for this outrageous social network to exist.
6 replies →
This isn't really true, though? The ISS does it with radiators that are ~1/2 the area of its solar panels, and both should scale linearly with power?
ISS radiators run on water and ammonia. Think about how much a kg costs to lift to space and you'll see the economics of space data centers fall apart real fast. Plus, if the radiator springs a leak the satellite is scrap.
1 reply →
The ISS creates radically less heat than a datacenter
I don't pretend to understand the thermodynamics of all of this to do an actual calculation, but note that the ISS spends half its time in the shadow of the earth, which these satellites would not do.
3 replies →
Radiator size scales linearly with power but, crucially, coolant power, pumps, etc do not.
Imagine the capillary/friction losses, the force required, and the energy use(!) required to pump ammonia through a football-field sized radiator panel.
Moving electricity long distance is a lot easier than moving coolant long distances, which puts a soft limit on the reasonable size of the solar array of these satellites. But as long as you stay below that and pick a reasonable orbit it's indeed not too bad, you just have to properly plan for it
The ISS isn't consuming and generating megawatts+ of power.
5 replies →
Setting aside the possibility it's window dressing for a financial bailout, there would be two ways compute in space makes sense:
1) new technology improves vacuum heat radiation efficiency
2) new technology reduces waste heat generation from compute
All the takes I've seen have been focused on #1, but I'm starting to wonder about #2... Specifically spintronics and photonic chips.
If you solve 2, heat dissipation goes away on earth too, so what’s the advantage of space
9 replies →
1. It's cheaper to make a vacuum on earth around a computer than it is to send a computer into space.
2. That would also presumably work on earth, unless it somehow relied on low-gravity, and would also be cheaper to benefit from on earth.
> new technology improves vacuum heat radiation efficiency
Isn't this fixed by blackbody radiation equations?
1 reply →
Maybe Karpathy has been hired to design a Full Self Cooling system.
It's a con, his AI business is failing, so he's rolling it up into the profitable business. Did a similar thing with Twitter.
This is so obvious, but it's so stupid and at this scale that people find it hard to believe.
Existing satellites manage to keep their equipment that already can consume several kW cool just fine.
You might need space for radiators, but there is plenty space in space.
5,000 Starship launches to match the solar/heat budget of the 10GW "Stargate" OpenAI datacenter. The Falcon 9 family has achieved over 600 launches.
The ISS power/heat budget is like 240,000 BTU/hr. That’s equivalent to half of an Nvidia GB200 NVL72 rack. So two international space stations per rack. Or about 160,000 international space stations to cool the 10GW “Stargate” datacenter that OpenAI’s building in Abilene. There are 10,000 starlink satellites.
Starship could probably carry 250-300 of the new V2 Mini satellites which are supposed to have a power/heat budget of like 8kW. That's how I got 5,000 Starship launches to match OpenAI’s datacenter.
Weight seems less of an issue than size. 83,000 NVL72’s would weigh 270 million lbs or 20% of the lift capacity of 5000 starship launches. Leaving 80% for the rest of the satellite mass, which seems perhaps reasonable.
Elon's napkin math is definitely off though, by over an order of magnitude. "a million tons per year of satellites generating 100 kW of compute power per ton" The NVL72's use 74kW per ton. But that's just the compute, without including the rest of the fucking satellite (solar panels and radiators). So that estimate is complete garbage.
One note: If you could afford to send up one of your own personal satellites, it would be extremely difficult for the FBI to raid.
Several kW is nothing for a bank of GPUs.
Radiators in space are extremely inefficient because there's no conduction.
Also you have huge heat inputs from the sun. So you need substantial cooling before you get around to actually cooling the GPUs.
24 replies →
I'm not big on this subject, but I understand that heat transfer is difficult in space, because there's little to transfer to. If the solution is just making large radiators, then that means you're sending some big payloads full of radiators. Not to mention all the solar panels needed. I wanna live in sci-fi land too, but I don't see how it makes any sense compared to a terrestrial data center.
2 replies →
> keep their equipment that already can consume several kW cool just fine
That's equivalent to a couple datacenter GPUs.
> You might need space for radiators, but there is plenty space in space.
Finding space in space is the least difficult problem. Getting it up there is not easy.
You can line the solar panels and radiators facing away from each other, and the radiators would take up less surface area. I think maybe the tricky part would be the weight of water + pipes to move heat from the compute to the radiators.
2 replies →
There's plenty of space in space, but there isn't plenty of space in rocket fairings, nor is there plenty of lift capacity for an unlimited amount of radiators.
> It's extremely difficult to keep things cold in space.
This is one of those things that's not obvious till you think about it.
It's such bullshit that we've decided this moron and others in his cohort can unilaterally reallocate such vast portions of humanity's labor at their whims.
This is an extremely stupid idea, but because of our shared delusion of capitalism and the idea that wealth accumulation at the top should be effectively limitless, this guy gets to screw around and divert actual human labor towards insane and useless projects like this rather than solving real world problems.
Just put a fan in a window.
what makes you believe this?
radiators can be made as long as desirable within the shade of the solar panels, hence the designer can pracitically set arbitrarily low temperatures above the background temperature of the universe.
Radiators can shadow each other, so that puts some kind of limit on the size of the individual satellite (which limits the size of training run it can be used for, but I guess the goal for these is mostly inference anyway). More seriously, heat conduction is an issue: If the radiator is too long, heat won't get from its base to its tip fast enough. Using fluid is possible, but adds another system that can fail. If nothing else, increasing the size of the radiator means more mass that needs to be launched into space.
1 reply →
Shading does work; JWST does this. However I don't see how you can make it work for satellite data centers. You would constantly be engaging attitude control as you realigned the panels to keep the radiators in shade. You'd run out of thruster fuel so fast you'd get like a month out of each satellite
2 replies →
Radiators can only be made as long as desirable because there's gravity for the fluid inside to go back down once it condenses. Even seen those copper heat pipes in your PC radiator?
1 reply →
these same comments pop up every time someone brings up satellite data-centers where people just assume the only way of dissipating heat is through convection with the environment.
18 replies →
what? the heat is coming from inside the house
1 reply →
The Goalpost shift continues, If elon were working for me, I would have fired him for having never delivered on any of his projects.
Hyperloop > Neuralink > Self-Driving Cars > Robotaxi fleets > Personal Robots > Orbital Datacenters > [Insert next]
> If elon were working for me, I would have fired him for having never delivered on any of his projects.
Never? For the sheer amount of moonshot bets he's doing, his track record would make any VC jealous. Zip2, PayPal, Tesla, SpaceX, Grok/xAI.
World's Best At Surfing A Temporary Hyperinflation Wave is not a life goal to really be proud of tbh.
1 reply →
>Zip2
I guess props to scamming Compaq into making a large investment that didn't pan out. He did personally make money so I guess win for him.
>In an effort to woo investors, Elon Musk built a large casing around a standard computer to give the impression that Zip2 was powered by a supercomputer.
>PayPal
Huh? He didn't found Paypal, his company was acquired by Paypal. You might as well give him credit for eBay while you're at it. Paypal released their first digital wallet in 1999. They acquired x.com (and Musk) in 2000. Paypal itself was then acquired by eBay in 2002.
>Tesla
Investor, not founder.
>SpaceX
Yup, props here.
>Grok/xAI
Hasn't made a penny, no signs it had any path to profitability, which is why it was shoved into Space-X to cover his personal losses.
1 reply →
Hey now. Huperliop was designed to scuttle California’s light rail project. Which it did. Mission accomplished.
I think you mean "California High-Speed Rail", not light rail.
Light rail, generally refers to urban rail, "trams".
California's light rail still exists, as does California High-Speed Rail, which is not light rail.
This is the funniest conspiracy theory.
5 replies →
What are you talking about?
hmm Tesla shipped millions of cars SpaceX launches 90% of space payloads, Starlink is working well. Thats hard to categorize as never delivered on any of his projects
The crucial thing is that Tesla's valuation has the hype projects baked in. The fact that it never delivered self driving or a robotaxi fleet and is now being saved solely by an import ban on Chinese EVs means that any success he had with Tesla is now an illusion.
5 replies →
The Starlink achievement alone is one of the most insane projects ever attempted and works really well.
The projects promised to be life altering for all mankind, they ended up being not even life altering for super rich Americans considering that Teslas are just EVs which without FSD are just regular cars with a different propellent that were made for political purposes and virtue signaling
The EV revolution has always been something almost dystopic : Trillions of dollars spent in order to not have the slightest amount of quality of life improvement, if anything a worse quality of life because you buy an EV that you cannot use 24/7/365 whereas you can an ICE car for much less .
As soon as something kinda elegant and hopeful as far as collective quality of life improvement is concerned (AI/ChatGPT) came around.....the whole green/EV revolution rightfully went out the window
If Musk was this genius you guys make him to be at 50 and with all the capital he burned he should have at least one company that if you disappeared the world would look drastically different, like if you disappeared Microsoft or Apple or Exxon or Aramco or Amazon or IBM....the world would come to a screeching halt.
Disappear one of Musk companies and everything would be the same as he's always involved in these sort of aspirational companies which have this great vision always 5 years into the future that never materialize into anything tangible or that improves the quality of life like the company I mentioned earlier
3 replies →
Yeah, but there are enough people to buy the hype.
In a way, its kind of cool to see how robber barons work in real time in our generation. Its also insanely depressing as they will systematically enshittify and extract as much wealth from society as is possible.
I don't actually think the Robber Barons in the 1920s had people going out of their way to defend them and insist they had special knowledge.
The New Deal happened with massive popular support because people did not like the Barons, and wanted to stop them and actually have a life worth living.
It only took like 30 years of suffering.
1 reply →
I'm amazed at this kind of thinking. I get it, obviously, and it's not uncommon, but still.
Elon Musk has already revolutionized three industries:
1. EVs: Before Tesla, no one thought electric cars could be a mass-market product. And even today, the Model 3 and Model Y are at the top of almost all sales lists.
2. Orbital Launch: No one expected Space X to succeed. What does a software guy know about real engineering? But today, re-usable rockets are the way of the future, and Space X is at least 5 to 10 years ahead of any other company.
3. Satellite Communications: Every single major military power is trying to deploy their own version of Starlink. Before Starlink, 50 satellites was considered a big constellation. Starlink has 8,000 satellites and they are literally launching hundreds every month.
I know it's impossible to prove a counter-factual, but I'm convinced that none of these three would have happened without Elon. No other Western car company has (even now) produced a profitable EV. No other space company has prices as low as Space X. No one even has the capability to build a Starlink competitor (not yet at least). Without Elon pushing these projects, they simply would not have happened or would have happened decades later (after China or someone else beat us to it).
Even his not-yet-successful projects are far beyond most other companies:
Boring Company has actually built tunnels and passengers are actually riding it. No one else is even trying.
Neuralink has actually helped patients.
Tesla FSD actually does work (I use it all the time), and even if Waymo is ahead, Tesla is easily in second place.
I 100% get the hatred for Elon Musk. His political positions are absolutely worth criticizing and I cringe most of the time he tweets. But to deny his business and engineering ability is just motivated reasoning.
Such illusions are ultimately self-defeating. The more opposed one is to Elon Musk (in business or politics) the more important it is to see his capabilities clearly.
> Boring Company has actually built tunnels and passengers are actually riding it. No one else is even trying.
Boring Company bought an existing tunnel boring machine (TBM), and used it to dig a car tunnel. Their only “innovation” in terms of any cost savings is to dig smaller tunnels - which we already knew could be done (tunnel cost grows with diameter), and which we don’t do for good reasons (capacity, emergency egress).
The branding and marketing exercise was excellent though.
The guy also was super excited to go rap3 some k1ds in the island. Gladly he also failed at this.
> "The only logical solution therefore is to transport these resource-intensive efforts to a location with vast power and space. I mean, space is called 'space' for a reason. [crying laughing emoji]"
This is all the reasoning provided. It is quite sad how a company I admired so much has become embroiled in financial doohickery.
It’s not SpaceX’s fault. It’s still a company to admire, it’s just that nobody appears to be able to stop Musk.
I wonder why SpaceX investors aren’t revolting.
> 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?
You 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.
How 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.
15 replies →
Unless I missed something the Microsoft underwater data center was basically a publicity stunt.
Anyone who thinks it makes sense to blast data centers into space has never seen how big and heavy they are, or thought about their immense power consumption, much less the challenge of radiating away that much waste heat into space.
11 replies →
An 8 GPU B200 cluster goes for about $500k right now. You'd need to put thousands of those into space to mimic a ground-based data center. And the launch costs are best case around 10x the cost of the cluster itself.
Letting them burn up in the atmosphere every time there's an issue does not sound sustainable.
8 replies →
You mean you operate them like Microsoft's failed submerged data center project [1]. When pointing at validating past examples you are generally supposed to point at successes.
[1] https://en.wikipedia.org/wiki/Project_Natick
3 replies →
But if we’re going down that line of thinking then it’s a poor comparison. I could open a data centre on the ground and use the same principle of zero maintenance, and it would be way cheaper and way more powerful.
But you could just run a “zero maintenance” data center on Earth and not pay to blast it into orbit.
How many submerged datacenters is Microsoft operating?
2 replies →
You also have to get rid of waste heat :P
This will totally work since we have an unlimited amount of rare earth elements we can just ship off into space never to see again. Infinite raw materials + infinite power equals infinite AI!!!
Accountants love this
The financial system is already freaking out about depreciation on land based data centers. I don't think it could survive what you're talking about.
Being under the ocean in a metal box you don't get too many micro-meteors or cosmic rays though.
3 replies →
As soon as a statement contains a timeframe estimate by Musk you know to disregard it entirely.
The 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"?
5 replies →
There'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
edit: iROSA = 33 kW per ton, thanks friends
Not to be an Elon defender, but can you back up your 0.15/ton? My own searching puts ROSA orders of magnitude higher. Each array is 600kg (0.6t) and puts out 20kw (https://space.skyrocket.de/doc_sdat/irosa-1.htm) which makes 20/0.6 = 33.333 kw/ton
2 replies →
The company lists their ISS solar panels as 28 kW for 331 kg, which comes pretty near to 100 W/kg.
Company website:
https://rdw.com/wp-content/uploads/2023/06/redwire-roll-out-...
And their Opal configuration beats the metric: 5.3 kW for 42.7 kg.
Maybe I'm just out of the loop, but is solar substantially more efficient in space? I assume the satellites won't orbit in a way that follows the sun. And presumably the arrays of panels they can attach to a satellite don't exceed the size of the panels you could slap on and around a data center (at least without being insanely expensive).
3 replies →
> But assuming that thermal rejection is good in space
Don't assume this. Why would you assume this?
Just put a slightly larger solar array on the same equipment on earth?
10 replies →
"There's clearly rhetorical hyperbole happening there" in a business paper is called lying
1 reply →
Elon is a pathological liar and it’s crazy that he still gets sanewashed after all he’s done. It’s insanity that he hasn’t been kicked out of leading his companies, and it’s also insanity that he hasn’t been prosecuted by the SEC.
You’ve spent too much life force trying to even understand the liar’s fake logic.
Let’s start right here: there is no such thing as becoming power/grid constrained on earth. If you replaced just the cornfields that the United States uses just to grow corn for ethanol in gasoline just in the corn belt, you could power the entire country with solar+batteries+wind. Easily, and cheaply.
If you don’t even believe that solar+batteries are cheap (they are), fine, choose your choice of power plant. Nuclear works fine.
The truth is, xAI combining with SpaceX is almost certainly corrupt financial engineering. SpaceX as a government contractor and that means Elon’s pal Trump can now siphon money into xAI via the federal government.
1 reply →
This 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.
It's always 2-3 years with this guy
Conveniently, about the amount of time it takes for the average person to forget and/or rematerialize in a new parallel dimension
1 reply →
You’re just not going at the speed of light as this guy’s brain is, time dilation is a thing
Currently, 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?
That's the price before Starship which would be the prerequisite for this whole project.
9 replies →
That launch cost is remarkably cheap to someone that's handled a $1.5million dollar 5U server filled with GPUs and RAM that weighs under 100kg.
Obviously the solar and cooling for the above would both weigh and cost a ton but... It's feels surprisingly close to being within an order of magnitude of current costs when you ballpark it?
Like i don't think it's actually viable, it's just a little shocking that the idea isn't as far out of line as i expected.
AI and space based economy
> 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 :)
Satellites have large operational costs. Satellite fleets even more so.
4 replies →
The 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_Inter...
The famous Musk timeline. "By next year, 2 year tops".
By his original timeline we should have landed on Mars… 4 years ago.
https://www.planetary.org/articles/20170929-spacex-updated-c...
1 reply →
> 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.
Do 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.
1 reply →
Here 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.
He just says stuff to convince people of things that benefit him. Internal consistency was never the plan.
1 reply →
Aside 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
11 replies →
What, creating a huge patchwork of self sufficient AIs, forming their own sky based net, seems bad to you, considering the whole torment nexus/Sky Net connotations? It's not like he's planning to attach it to his giant humanoid robot program. Oh. Ohhhhh. Oh no.
Launching alone consumes about 75-150kWh per tonne of energy for fuels only (as per ChatGPT).
Planned lifespan of Starlink satellites is 5years.
A million tons a year would be over 18 Starship launches per day.
>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.
We’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.
Any 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.
Never 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.
The 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.
2 replies →
But 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.
Do you honestly believe that nobody involved has ever considered that?
1 reply →
You lost me on million tons.
[flagged]
Sometimes, I wonder how people in the middle ages accepted the whole "Divine Right" of their ruling Kings, while simultaneously suffering under their rule.
> Larry Ellison, another hyper-informed genius business man
"King George, another royal blessed by the divine."
11 replies →
> 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.
2 replies →
Starlink runs special rad-hard computers from AMD. None of that transfers to top of the line GPUs. This is crazy.
4 replies →
> He's got more information about space based compute deployments than any other person in the world...
He also had more information about self-driving progress than any other person in the world - yet he was wrong with his predictions every year for last 10 years.
> I will say there's a MASSIVE cost to getting power infrastructure, land, legal stuff done on terra firma; all that just sort of .. goes away when you're deploying to space, at least if you're deploying to space early and fast.
You need both power infrastructure and structures to build within for deploying in space too. And you have to build them and then put it all into space.
Cost per square foot of land is not that high basically anywhere you could build a datacentre to offset that.
3 replies →
We should take a hyperloop trip together, connect our nuralinks and figure this out together. Or perhaps our optimus bots can help us understand?
> like Larry Ellison, another hyper-informed genius business man
Don’t anthropomorphize Larry Ellison.
Elon 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.
1 reply →
This just reads like Elon trying to leverage the AI bubble to prop up SpaceX stock to me.
3 replies →
They'll just be decommissioned and burn up in space. nVidia will make space-grade GPUs on a 2-3 year cycle.
They don't need to be space grade, consumer hardware will do just fine.
For AI a random bit flip doesn't matter much.
6 replies →
You 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.
1 reply →
> 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.
> 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.
The 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.
I fly satellites. None of them have a zero operational cost. None. Even the most automated cost money to keep running.
The problem is solved by making satellites extremely expensive.
If you had told me 4 years ago that Twitter would be merged into SpaceX I would have called you crazy. Yet here we are..
Is xAI Twitter? I thought they were separate companies, but honestly I don’t know anymore.
Yes X was merged into xAI last year I believe.
3 replies →
Elon Musk’s AI startup, xAI (x.ai = grok), officially acquired X Corp. (the parent company of the social media platform X or x.com, formerly Twitter/x.com) in an all-stock deal. Both now operate under a unified holding entity, frequently referred to in corporate filings as X.AI Holdings Corp. (or simply xAI). Now SpaceX has moved to acquire or merge with xAI. This effectively brings the social media data from x.com, the AI development of x.ai, and the satellite infrastructure of Starlink/SpaceX under one "super-conglomerate" roof.
I'm just waiting for the year 2350 when Walmart buys Weyland-Yutani.
Anyone remember the quote by Russ Hanneman on SV [0] - "No Revenue, means you're potential pure play"
We know datacenters in space - sound plausible enough - yet not practical - hence they're potential pure play - also you can have massive solar in space - unlimited space -- etc -- all true -- but how economical / practical is it ?
yet we know on earth - to power the whole earth with solar - only a fraction of the land is needed. Hell it's even in the Tesla Master Plan v3 docs [1] - current limitation being storage & distribution
so all you - are now witnessing to the greatest scam ever pulled on earth.
[0]: https://www.youtube.com/watch?v=BzAdXyPYKQo [1]: https://www.tesla.com/ns_videos/Tesla-Master-Plan-Part-3.pdf
Seems like a great way to play games with moving money around. Come up with a "valuation" and then "acquire".
It worked for his Twitter co-investors. I guess they overlap enough with Xai investors for them to think it's more clever than it is a rip off.
I'm pretty amazed one can play a shell game for so long and so openly with the public.
<elon venture> rescues failing <elon venture> here have some <unattainable goals> the shareholders love that shiz.
From the Big Short (movie)
Jared Vennett (narration): "In the years that followed, hundreds of bankers and rating agency's executives went to jail. The SEC was completely overhauled, and Congress had no choice but to break up the big banks and regulate the mortgage and derivatives industries."
"Just kidding. Banks took the money the American people gave them, and they used it to pay themselves huge bonuses, and lobby the Congress to kill big reform. And then they blamed immigrants and poor people, and this time even teachers."
I feel like without adding some commentary with these quotes this comment lacks enough info to see how it relates to the linked article.
Because this move is entirely financial engineering to hide losses just like the roll up of X in to xAI.
None of this has anything to do with business or innovation. Do you not immediately see that? Most of my friends reaction to this news was that this is so obvious it's almost funny (or actually it is funny, since most were laughing as they read the headline).
I'm curious how you could not understand the relevance of the quote unless you were aggressively trying to not understanding it.
65 replies →
Legitimately, did you not immediately conclude it was for financial shenanigans? What did you think? I'm not trying to be shitty, but what else could there be?
25 replies →
without having watched the Big Short or having read the article, my first impression from the quote is "Megacorporations are failing dramatically, and the billionaires at their helm are freely doing financial gymnastics to pull the covers over the eyes of shareholders, while gaming the system to fully circumvent taxes and regulation -- the people with the power to do anything about it (legislators and regulators) watch idly (maybe profiting), the oligarchs make off like bandits despite copious failures, and the end consumer/taxpayer is either robbed or clueless this is going on, but most likely both, when there was a world where accountability could have been had and the common man was treated better."
the article headline immediately screams "financial gymnastics" to me so the rest followed from the quote.
1 reply →
You don't see how the common thread is Elon Musk buying out his own businesses, largely on the basis of overinflated stocks and corporate welfare?
It's baffling that the market has stayed so irrational because of Musk. It will collapse because of him.
1 reply →
This merger smells like a bubble. Servers in space? They don't make enough to cover costs here on earth. Americans will be forced to bail this mess up because we need Falcon 9, Starship one, etc.
4 replies →
It is pretty clear. Socialize the losses of Musk investments by recovery via SpaceX contracts, supporting the US space program and the new Golden Dome program.
He sold FSD for 12 years, now is going to sell a Dyson Sphere for the next 30. This guy makes Ponzi look like a street hustler.
[flagged]
[flagged]
It's messed up that Grok underwrote all those subprime mortgages in 2008
I think the argument is that it's messed up that a large debt swap from xAI kept Musk's margin on Twitter from being called by his investors, and now that debt is being absorbed by SpaceX.
18 replies →
Serious question .. are you long on cash or what? What investment class is not overvalued? Gold is but a store of value and doesn't really grow - look at how high it has gone.
I really don't know what you're trying to say. From your comment alone the conclusion I drew was that we should spend more on the industries that makes physical instead of financial products, such as SpaceX.
The sequence of events: Elon doing a leveraged buyout of X, then xAI funding, then debt transfers to X, then the xAI–X stock deal. Now the proposed SpaceX–xAI merger appears to have shifted X’s financial burden from Musk personally toward xAI investors and, potentially, future SpaceX shareholders.
This is speculative, of course, but yeah seems likely.
5 replies →
SpaceX is a private company
Check back in 2 years to see if your statement is still true.
SpaceX is planning the largest IPO in history aiming for over a trillion dollars in market cap
1 reply →
so are the banks?
1 reply →
Step 1: Merge xAI and SpaceX
Step 2: IPO SpaceX
Step 3: Merge Tesla and SpaceX x xAI (which would have been tricky if they were still private).
Where's the government bailout in this transaction that would make this a relevant comparison?
https://arstechnica.com/tech-policy/2026/01/starlink-demands...
Who is SpaceX’s biggest customer? And which industry are we being told by any number of governments around the world is the most importantly thing ever and must be subsidized and forced on people?
7 replies →
There has been rumblings of speculation that when the AI bubble pops that the government will bail the companies out.
Just a neat bit of financial engineering. You can tell because Elon picked SpaceX instead of Tesla – which would have actually made sense at some level (Optimus Robots + AI). But Tesla is public and so he'd need to follow laws and reporting requirements.
You can tell it's just financial engineering because in the entire press release xAI is only mentioned in the first two sentences. Everything after that is Elon talking about space data centers to distract from the actual topic. Which seems to be working
This is fairly naive, Elon isn't the only investor in SpaceX.
My guess is "that they did the math" and had an engineering study which convinced them that getting AI datacenters into space will make sense.
It's also not hard to imagine why, the process alone once perfected could be reused for asteroid mining for example, then mirogravity manufacturing, either of which alone would be enormous capital intensive projects. Even if AI dataenters in space are break-even it would be a massive win for SpaceX and leave their competition far behind.
Are you a bot or are you just stupid?
3 replies →
does he need spacex/xai to prop up tesla or the other way around?
Tesla is still very profitable, as is SpaceX I assume. Twitter/X has been a $44 billion dollar failure, and xAI is a vanity project so Musk can go around saying he is a player in the AI space. Investors in both X and xAI need to be bailed out, hence this announcement.
21 replies →
I suspect SpaceX will acquire Tesla at some point. It’s the most profitable of these companies. So basically SpaceX employees and shareholders are covering up for the failing Tesla business and the already-failed xAI business.
Let’s not forget, xAI is the parent of Twitter/X (the social network). So now, taxpayers are paying to keep Twitter/X alive. After all, it is taxpayer money going to the contracts the government gives SpaceX for launches. Nice way to subsidize what is effectively a one sided campaign machine for the GOP and far right.
4 replies →
I guess the difference is Tesla is a public company, so requires more paperwork. SpaceX isn't public yet, but will be soon, meaning it will have a cash infusion.
With that number of more satellites in orbit, launching a manned rocket into space is likely to be too risky due the amount of more debris and satellites. And will it be net energy positive solution ?
This is why I come to this site. Obviously, Twitter's financials are struggling and theres more than a few people rich people who don't want to take the hit... but we can all drop that for a second to discuss the plausibility of data centers in space. Some links and comments I enjoyed:
> In the long term, space-based AI is obviously the only way to scale.
I never questioned it.
Space is also extremely cold, and if it's as dense as Musk cooling won't be an issue.
I can't tell how many layers of sarcasm are here, but I just want to highlight that aktshually cooling in space is quite difficult because there is no convection, so the only cooling option is radiative. Which gets a bit hard when the satellite gets blasted by the sun.
The ISS doesn't have problems staying warm, it has problems cooling off.
> the only cooling option is radiative.
It does say he's planning an AI sun, I'm guessing that's the temperature you need to run at for radiation to work.
6 replies →
Hence the “dense as Musk” comment
> scaling to make a sentient sun to understand the Universe and extend the light of consciousness to the stars!
I didn't realize SpaceX's media press is even cringer than Elon's average tweet...
There's an audience for this. People have been taking these vapid statements and his trouble with communicating as evidence of genius for well over a decade now.
Microdosing might not be the word for it.
I mean this is clearly directly written by Elon. I suspect the SpaceX comms people are equally eye-rolling.
I still don't understand the "data center in space" narrative. How are they going to solve the cooling issue?
Maintenance cost must be pretty fucking insane
This "Space Datacenter" sounds like biggest bullshit in last decade, which is pretty damn fucking high bar.
Hes committed to building thousands of Optimus robots for a market that does not exist while cutting back on building evs (a market that does).
I think its pretty clear that Musk has lost his goddamn mind. And the American corporate system and Government seem powerless to do anything.
2 replies →
There is no maintenance, you have many cheap satellites - if one fails you just deorbit it.
2 replies →
I think it's fair to say past 30 years. Dotcom boom only had modest cons by contrast
Same way the dude solved Roadster, full self driving etc.
Presumably the cooling problem gets hand waved away as a technical detail, and the real selling point is data centers that aren't subject to any regional governments laws.
> and the real selling point is data centers that aren't subject to any regional governments laws.
No? ISS isn't exempt from legal systems.
https://www.esa.int/Science_Exploration/Human_and_Robotic_Ex...
1 reply →
So the csam generator can live on? Why not just pirate radio style it and cool it with ocean water....
Send up a spacecraft with back-to-back / equal area solar panels and radiators (have to reject heat backwards, can't reject it to your neighboring sphere elements!). Push your chip temp as much as possible (90C? 100C?). Find a favorable choice of vapor for a heat pump / Organic Rankine Cycle (possibly dual-loop) to boost the temp to 150C for the radiator. Cool the chip with vapor 20C below its running temp. 20-40% of the solar power goes to run the pumps, leaving 60-80% for the workload (a resistor with extra steps).
There are a lot of degrees of freedom to optimize something like this.
Spacecraft radiator system using a heat pump - https://patents.google.com/patent/US6883588B1/en
Periodically send a crew to a comet to bring back a large slab of ice to put in the data center.
beltalowda!
Remember MoviePass, and how they were losing gobs of money by letting people see unlimited movies for $20/month?
It was so obviously stupid that a bunch of people went, "well, this so clearly can't work that they must have a secret plan to make money, we'll invest on that promise", and then it turned out there was no secret plan, it was as stupid as it looked and it went bankrupt.
The "datacenters in space" thing is a similar play: it's so obviously dumb that a bunch of smart people have tricked themselves into thinking "wow, SpaceX must have actually figured a way it can work!"; SpaceX has not and it is in fact exactly as stupid as it looks.
The pre-IPO timing of this narrative, combined with Musk’s history of Tesla’s stock pumping, leaves no room for doubt. But, that has worked for Tesla, I’m pretty confident it will work for SpaceX, which will IPO for $1T+.
But it won't end the same as MoviePass until Elon dies; he will keep moving things around, propping up failures with VC, IPO, federal/state (taxpayer) and profit making business money.
that's the best part! They don't!
My guess would be just a regular radiator and cooling system like a liquid pump. The only obstacle should be the vacuum. That said I don't have any hopes Elon has any understanding of any of it.
Cooling a datacenter in space isn't really any harder than cooling a starlink in space, the ratio of solar panels to radiating area will have to be about the same. There is nothing uniquely heat-producing about GPUs, ultimately almost all energy collected by a satellite's solar panels ends up as heat in the satellite.
IMO the big problem is the lack of maintainability.
> Cooling a datacenter in space isn't really any harder than cooling a starlink in space
A watt is a watt and cooling isn't any different just because some heat came from a GPU. But a GPU cluster will consume order of magnitudes more electricity, and will require a proportionally larger surface area to radiate heat compared to a starlink satellite.
Best estimate I can find is that a single starlink satellite uses ~5KW of power and has a radiator of a few square meters.
Power usage for 1000 B200's would be in the ballpark of 1000kW. That's around 1000 square meters of radiators.
Then the heat needs to be dispersed evenly across the radiators, which means a lot of heat pipes.
Cooling GPU's in space will be anything but easy and almost certainly won't be cost competitive with ground-based data centers.
Sure, but cooling a starlink in space is a lot more difficult than cooling a starlink on earth would be. And unlike starlink which absolutely must be in space in order to function, data centers work just fine on the ground.
5 replies →
According to Gemini, Earth datacenters cost $7m per MW at the low end (without compute) and solar panel power plants cost $0.5-1.5m per MW, giving $7.5-8.5m per MW overall.
Starlink V2 mini satellites are around 10kW and costs $1-1.5m to launch, for a cost of $100-150m per MW.
So if Gemini is right it seems a datacenter made of Starlinks costs 10-20x more and has a limited lifetime, i.e. it seems unprofitable right now.
In general it seems unlikely to be profitable until there is no more space for solar panels on Earth.
3 replies →
I think that it's not just about the ratio. To me the difference is that Starlink sattelites are fixed-scope, miniature satellites that perform a limited range of tasks. When you talk about GPUs, though, your goal is maximizing the amount of compute you send up. Which means you need to push as many of these GPUs up there as possible, to the extent where you'd need huge megastructures with solar panels and radiators that would probably start pushing the limits of what in-space construction can do. Sure, the ratio would be the same, but what about the scale?
And you also need it to make sense not just from a maintenance standpoint, but from a financial one. In what world would launching what's equivalent to huge facilities that work perfectly fine on the ground make sense? What's the point? If we had a space elevator and nearly free space deployment, then yeah maybe, but how does this plan square with our current reality?
Oh, and don't forget about getting some good shielding for all those precise, cutting-edge processors.
7 replies →
I am concerned, and haven’t seen anyone else point this out yet, that Musk will move Grok’s CSAM generation capabilities to space to be beyond the reach of terrestrial policing. Does this create some sort of legal loophole here so Musk can do this with impunity?
No. For one, things in space are under the jurisdiction of the country they were launched from. For another, it's people that do crimes, not satellites or LLMs, and the people involved in making CSAM are all on Earth.
AFAIK he can do whatever he wants in space, but CSAM is still illegal to view or even download in most (all?) countries of the world. So unless the degenerates also move out into space (which I'm sure they're eager to do), it wouldn't really ease the legal situation here on Earth.
Ground stations would be the major problem.
Maybe if Elon launched himself and the dev team into orbit and didn't use any ground stations and just Starlink terminals he could start getting into legal loopholes.
So my floating data center in international waters idea has potential investors?
Remember that he wants to make X a bank. An orbital tax haven.
But seriously, I think legally satellites are under the jurisdiction of the country they were launched from.
Well, offshore launches are already a thing.
Or he could just buy a small island in the Carribean. There's one in particular that is available.
1 reply →
Elon didn't want to get outshined after Sam Altman suggested that "maybe we build a big Dyson sphere around the solar system". When will people realize that these "geniuses" are only good at making money, and any benefit to society is coincidental.
Too me this smells of projected cash desperation. Do people actually pay for Grok?
You get Grok with paid x.com ; so there is some sort of cash from that, I would guess.
Is Grok driving any x.com subscriptions?
There are regularly comments on HN from people who say they pay for Musk's various products, and I am always downvoted into oblivion for suggesting that that is ethically problematic.
There's obviously quite a lot of autocratist illiberals in tech.
as you can see from the Epstein files, people with minor interests have a lot of money
Grrreat! Grok in Space... now AI-generated non-consensual sexual materials can be made completely outside the jurisdiction of any earthly body!
Rah rah. Line goes up!
https://tinyurl.com/xai-joins-spacex
isn't this just fraud in broad daylight? I don't get it. Why not at least try to hide it?
It's crazy but legally there's nothing fraudulent here. I'm sure the deal was approved by the boards of both companies.
Its only fraud if poor people do it. Welcome to American politics.
But the "fraud" here is being done mostly to VC investors with deep pockets and lawyers, at least until he tries to take this entity public. And I can't imagine them just taking this lying down, but then again maybe they realize that offloading this steaming pile on public market investors is the best way out. But even then... SpaceX seemed like it was quite viable on its own, the investors there are the real losers here.
It is all very puzzling to me.
no I get it, but I mean fraud is usually kept out of the public. this is fraud in broad daylight?
1 reply →
It's only fraud if rich people lose money
Poor people are using their public car company to buy their private space company?
How do you keep these kind of things under the radar anyway?
5 days after Tesla gave xAi 2 billion.
Funnel money from a public company to a private entity and then make it disappear. Poof! Magic.
Thoughts:
1. What in the circular funding? This feels more like a financing scheme founding it under X/Twitter and then spinning it over to SpaceX. I suspect some debt is disappearing or taxes aren't getting assessed because of this move.
2. The only thing harder than harnessing "a millionth of the sun's power" on Earth would be launching enough material into space to do the same thing. And that's not even a reason for SpaceX to own an AI company, at least not at this point. The current AI isn't going to help with the engineering to do that. Right now hiring 20-somethings fresh out of college is way cheaper and SpaceX has been very successful with that.
quick edit: dang, I even got point 1 backwards. xAI owns X/Twitter, and that means that SpaceX now owns X/Twitter as well as an AI company. Super suspicious that SpaceX could actually think that buying the social media part (a significant portion of xAI's value) would be worth it.
There is literally an emoji in the middle of the announcement post. Very on brand for Elon.
Datacenters in space are a terrible, horrible, no good idea: https://news.ycombinator.com/item?id=46087616
Most everyone here is just angry Elon opened the door for freedom of speech to exist online and started the process of uncovering the left's extreme corruption. They feel betrayed by him so he's enemy #1. Literally could cure cancer and he'd still be branded as evil by the left. Ironic as he's arguably done more for leftist causes (renewables, ev's, etc) than anyone else on the planet.
These are private companies, with private investors. If Elon really screwed all investors over and over again he wouldnt be able to raise money repeatedly. But he has no problem raising money. Because the claims that he's doing shady things or screwing investors is entirely unfounded. Same goes with the claims that he's an idiot or not that smart. If thats really the case, show me what you've accomplished with your genius.
It takes intelligence to recognize intelligence. An idiot cant recognize a genuis.
Unfortunately you led with the claim that Musk cares about freedom of speech, a claim so laughable it renders anything else you say unbelievable.
Elon's purchase of Twitter (now X) opened the door for conservative view points to be heard online. Prior to that, all other major platforms were censoring conservative views (with wide spread company leadership claiming due to pressure from Biden administration).
I regularly use X. I get blasted with political posts from both sides constantly on it.
I thought xAi previously merged with twitter, so all of this is now rolled into SpaceX? Atleast the investors in xAi and the original financiers in twitter get a breather. SpaceX is the new bandaid for this hot mess. Let’s see if this ends up rotting SpaceX or if it gets healed.
Seems like a way to put a lot of junk in space. If thats in earth orbit it will lead to a lot of junk falling from the sky in 10 years. If it all burns up that will be a lot of nasty shit in the atmosphere - millions of tons!
The really skeptical take here is that eventually all of Musk's companies merge, or at least the biggest ones, for juicing that market value to get that $1T payout. Looking at Tesla.
Then he will spin off and then remerge again. He has several more small companies before he plays that game. Maybe next quarter!
Related: NASA chief suggests SpaceX may be booted from moon mission [1]. Blue Origin could snatch SpaceX's Starship lander contract. This looks increasingly a good idea.
[1] https://www.cnn.com/2025/10/20/science/nasa-spacex-moon-land...
Sean Duffy is no longer acting administrator of NASA. This proposal was apparently part of a bid to get the support of a coalition of old-space companies and new-space non-SpaceX companies. As part of that strategy he apparently leaked Isaacman's Project Athena document and was backgrounding that he was a SpaceX plant.
But, Isaacman is administrator now, and whatever you think about Isaacman and his relationship to SpaceX, I don't think there's much merit in thinking one of Duffy's half thought out plans is likely to be carried out.
Sadly this seems correct. When Trump was re-elected Elon Musk pushed for Jared Isaacman to be appointed as NASA administrator. When the pick went another way, it led to some real friction between Musk and Trump. Now, with Isaacman finally at the helm of NASA, it looks like Musk’s influence over the agency has come full circle.
> This looks increasingly a good idea.
Why?
The reason for the entire moon mission is national prestige.
Is financial fraud consistent with our national prestige?
There are better companies.
4 replies →
SpaceX hits delays according to the article
2 replies →
https://www.spacex.com/updates#xai-joins-spacex additionally the longer article on SpaceX site
IF we develop beyond earth...then AI, robots and connectivity will likely be huge parts of it.
+ spacex already is the best way for many payloads to get to space.
+ starlink is already the best low orbit based connectivity solution.
+ x is already a great way to train virtual world AI.
+ tesla (and its robots) is a great way to train physical world AI.
+ space takes big $ and talent - this combo would have both.
the IF at the top is just that. but feels like an interesting convo for this crowd.
Kind of a bad look - but I can't precisely say why. Maybe he thinks he can raise more capital this way than he could for each company separately? Especially raising more money for X might be quite hard - they seem to be quite a bit behind on the revenue side compared to OpenAI / Anthropic. With both companies merged he might just find enough retail investors willing to buy at sky high valuations.
Accountants will be studying the deals and cyclical valuations of AI companies in the same way we study bank runs and FDIC insurance today.
Doesn't the idea of Orbital Datacenters imply that the constraining resource right now is physical space, and not compute, electricity, etc?
Did we suddenly solve the electricity problem, or the compute problem? As far as im aware there are still plenty of datacenters being planned and built right now.
[flagged]
They must have linked the wrong press release /s. I would have expected a press release about SpaceX acquiring xAI to talk about why they did that. Or at least mention xAI beyond the first paragraph. This is just Elon talking about space data centers
This is definitely better than merging with Tesla.
They can sell xAI/Grok to all automobile companies along with Tesla and other businesses(X.com included) just like the SpaceX services.
It would good to see how it was valued.
I'm not sure if this is meant to be sarcasm but is there really a need in the car market for on-demand CSAM? What actual use does Grok or any LLM have in a car?
This makes me genuinely sad. SpaceX was the one thing of his that Elon has largely avoided screwing up. Imho, this is in large part due to Gwynne Shotwell. She seems to have the personality (not to mention, personal wealth) to kick Elon in the head when he tries to mess things up.
What’s happening now is nothing more than a transparent effort to couple the AI hype-wagon to SpaceX in order to drive the valuation higher in the minds of investors who still think that LLMs will completely transform society.
I’ll be thrilled if the rocket folks can avoid being distracted by this nonsense, but I’m not optimistic.
I’ve been following SpaceX since something like the 2nd Falcon 1 launch and this is the worst thing I’ve seen happen. Sad times.
I think it’s just financial, I don’t see this as being detrimental or disruptive to SpaceX much at all.
If investors are falling for it, I guess all we can hope for is that the government doesn't bail it out!
One thing to keep in mind. xAI and SpaceX both have contracts with the DoD. So it makes sense he moved it there rather than Tesla. Not sure I buy the needing AI for doing more in space or if this is to save sinking ship, but if one of his two big companies needed to buy it to keep it afloat it makes sense it was SpaceX and not Tesla.
I'm wondering if SpaceX's going public will be delayed. If not we'll see the first test of the public's appetite for what the AI companies' balance sheets look like
Trails of those low orbit satellites wasn't bad enough.
Can't wait to see pictures of night sky ruined by... A data-center in the frame.
"Launching a constellation of a million satellites that operate as orbital data centers is a first step towards becoming a Kardashev II-level civilization"
So, basically give ourselves Kessler syndrome. Or is Elon trying to monopolize orbit entirely?
> buy a dying social network for 44bil
> merge it with a company created out of thin air for 20bil.
> have a third company buy it.
put it back on the market for 1.5 trillion.
> Starship will deliver millions of tons to orbit and beyond per year
Excuse my naive physics, but is there a point at which if you take enough mass off of earth and launch it into space, it would have a measurable effect on earth's orbit? (Or if the mass is still tethered to earth via gravity, is there no net effect?)
Earth weighs about 5972200000000000000000000 kg. They are claiming to plan to launch ~5000000000 kg / year. That's 8x10^-14 %. You'd need some pretty accurate instruments to tell the difference.
Whenever computer chips go into space, they have to be hardened against radiation, because there is no atmosphere to protect them. Otherwise you get random bit flips.
This process takes a while, which is partly why all the computers in space seem out of date. Because they are.
No one is going to want to use chips that are a many years out of date or subject to random bit flips.
(Although now it got me thinking, do random bit flips matter when training a trillion parameter model?)
LLMs specifically need random bits flipped for the results to be 'creative'.
That's not exactly how LLM temperature works. :). Also that's on inference, not training. Presumably these would be used for training, the latency would be too high for inference.
1 reply →
not a problem for "AI". it's just a bit more spice (temperature) which Grok possibly need! jk!
> My estimate is that within 2 to 3 years, the lowest cost way to generate AI compute will be in space.
I have never been so tempted to join Kalshi
Genuine question: is it even theoretically possible to find some way to dump the heat that would be generated by a "data center" in space?
Yes, satellites and the ISS successfully radiate heat today.
ISS solar panels only make 200kw at max.
https://taranis.ie/datacenters-in-space-are-a-terrible-horri...
It's not comparable to any data center.
This is terrible for Space-X. They're doing a great job. Musk has left running it to Gwynne Shotwell, who really is a rocket scientist. Now Space-X has a AI business unit they don't need, a new money drain, and more attention from Musk.
Should have merged xAI into Twitter. A failure there would not be a major setback.
So... Elon wants to literally build Skynet?
Is there anything substantially different about Google's announcement https://news.ycombinator.com/item?id=45813267 that makes it any more sane than the Space-X announcement?
Does this mean the foreign software engineers in xAI are now subject to ITAR?
Didn’t Elon say that orbital solar collection was a stupid idea due to energy loss in transmission? Using AI as an almost proof-of-work shows that it may potentially be more complex problem than previously thought. If we threw Bitcoin miners up to those satellites you could literally beam money down.
Let's call it for what it is, a payday for Elon. Paper billionaires have figured out they cannot cash out with out tanking their paper, so now you have these circular deals to extract as much as possible. If we had a functioning government they would step in and put an immediate stop to this on national security grounds.
Elon investors should try buying a lottery ticket, it also lets you dream of the future while not providing returns.
Musk is moving value out of public hands and into his own. He overpaid $44B for Twitter, then rebranded it as an AI asset by folding X into xAI. He pushed Tesla to invest $2B of shareholder money into xAI despite shareholders voting no. Five days later, SpaceX acquired xAI, effectively turning Tesla’s cash into equity in a private company Musk owns far more of. Musk controlled every step, there was no real arm’s-length process, and he almost certainly knew the outcome in advance. Musk and his private investors get control, inflated valuations, and IPO upside. Tesla shareholders supply the cash, take the risk, and lose leverage.
I've yet to attain full-stack mastery in my job, but Musk has already attained capital stack mastery.
I don't see the demand for space being there, OSS is driving costs down and there are still plenty of hardware and algorithmic optimizations we haven't deployed yet.
This is either insanely ambitious genius or pure shithousery. I guess we'll find out which one it is in 10 years
Given some of Musk’s previous statements, I think I know which one it is already.
well, Musk has been overpromising and under delivering for a decade (or more?), so it seems pretty clear this too is shithousery, albeit possibly ambitious.
Banksters struggled to sell off Twitter notes. Did they get out intact finally?
I suppose one of the ADR’s read something like “…who cares about bitflips, man. Isn’t AI all about probability?”
Knowing the insane level of hardening that goes into putting microcontrollers into space, how to the expect to use some 3nm process chip to stand a chance?
A trend at the moment is to just hope for the best in cubesats and other small satellites in LEO. If you’re below the radiation belt it’s apparently tenable. I worked somewhere designing satellite hardware for LEO and we simply opted to use consumer ARM hardware with a special OS with core level redundancy / consensus to manage bit flips. Obviously some problems will present for AI there… but there are arguably bigger problems with AI data centres like the fact that they offer almost no benefit with respects to the costs of putting and maintaining stuff in space!
Just checking (genuine question) there wouldn't be a sneaky way to weaponize a million satellites in orbit around the Earth, would there? I can't imagine it wouldn't have ever been looked into.
I have so many conflicting thoughts that I cannot properly articulate yet. I can say though, this is not going to end well for most, it is clumsily premeditated and starting to feel like dude is just trying to be a Neal Stephenson character.
Any details regarding valuations etc?
I will not be left holding this bag. This is such financial engineering nonsense, and if we had any sort of regulatory controls this would never be allowed to happen - especially BECAUSE of national security reasons.
The Howard Hughes of our time. Soon enough he will start pissing in milk jugs.
What's to stop president AOC from pulling the clearances of everyone working for SpaceX?
SpaceX launches about 60% of National Security Space Launch payloads. The only other active launcher cleared for these missions is ULA's Vulcan Centaur, which has launched 3 times total, ~once every 8 months.
And? Donald Trump's presidency has made it clear that "this is bad for our country" isn't a sufficient argument.
1 reply →
They both have X in their names, just imagine the synergies!
Do this math include the cost and weight of the radiators? Because it obviously can't work without big radiators, and I don't see them mentioned in the math?
I think there is one more possible framing for this.
Recently xAI has been in the news for Groq's revenge-porn-like "undress them" feature, which seems pretty legally questionable.
Musk has also been in the news for his own Epstein-related activities.
If he can move Groq and X into space, well, there's not very many age-of-consent or revenge-porn laws in space as far as I know, so maybe he'll be able to do some sort of legal leverage where the space data-center can produce otherwise legally questionable AI responses with impunity.
[dead]
I swear Prof G mentioned this exact same thing happening today
It was also foreshadowed in the Benedict Evans newsletter.
This makes a lot of sense. The commercial launch business is not large enough to support all possible Falcon launches, so Starlink was created to take advantage of the low launch cost and vertical integration and is now a major profit center for SpaceX.
Starship launches are only going to make sense every 779.94 days (the approx 2 year Mars-Earth proximity). The rest of the time, the launches could similarly be used to deploy orbiting data centers for XAi/Grok etc. Brilliant move.
I hope all the Tesla shareholders understand that they’re about to get hosed.
Musks making Tesla seem like a good fit into the portfolio.
At least there will be AI and Agentic stuff in mars.
Has SpaceX figured something out related to photonic chips that dramatically reduces waste heat generation of compute?
I suspended my disbelief and gave it a chance but I couldn’t hold it anymore after the emoji.
> Current advances in AI are dependent on large terrestrial data centers, which require immense amounts of power and cooling.
You know what's even harder to cool?
> Orbital Data Centers
I really would like to see a cost and cooling breakdown. I just can't see how you can do radiative cooling on the scales required, not to mention hardening.
I thought this was a troll by Elon, now I'm leaning towards not. I don't see how whatever you build being dramatically faster and cheaper to do on land, even 100% grid independent with solar and battery. Even if the launch cost was just fuel, everything else that goes into putting data centers in space dwarfs the cost of 4x solar plus battery.
Nobody knows cooling satellites better than SpaceX
I cannot really tell satire apart from genuine opinions anymore.
(But I do hope it was satire, if not, cooling satelites was/is a big issue and they only have very modest heat creation. A data center would be in a quite different ballpark)
Maybe so, but the actual SpaceX engineers are powerless to stop Elon running his mouth.
This is basic physics lol
Perhaps SpaceX incentive is to lie?
The first M&A announcement I've seen in my entire life that includes a laughing emoji; maybe that's what it is!
Why?
Because it's another tool to move money on books and make it seem that spaceX and or xAi look good to investors when needed. That would be my guess.
Pre-IPO price padding. xAI is going nowhere but at least for now it has some value. Move it under SpaceX, bump up SpaceX’s valuation and therefore it’s opening IPO price. Then kill xAI and write it off.
These are no longer tech companies, to they are financial, and power, instruments of the billionaire class
> These are no longer tech companies, to they are financial, and power, instruments of the billionaire class
SpaceX has made numerous breakthroughs in reusable launch vehicles, human spaceflight, satellite constellation, and rocket propulsion.
SpaceX is the world's dominant space launch provider with its launch cadence eclipsing all others, including private and national programs.
2 replies →
I know that per HN's guidelines we're supposed to be "kind and curious", and "reply to the argument instead of calling names". But with some texts, engaging with individual arguments loses sight of the more important bigger picture. So while unkind, the most "thoughtful and substantive" thing I think can be said about this text is:
The man's a moron.
If Musk and SpaceX are serious about putting 1 million datacenter satellites into space, then they are not serious about Mars.
You cannot simultaneously build and launch 10’s of thousands of Starships to deliver 1 million tons of equipment and supplies bound to Mars while also committing to launching 10’s of thousands of Starships to orbit full of satellites.
They would need to quadruple their launch rate, and half of those launches would be Starships bound for Mars, the vast majority of which would never return.
How many Falcon9’s have ever been built? It is incredible to say you can build that many rockets and use up that much fuel on any reasonable time scale. You might as well say the Tesla Roadster version 3 will be a Single Stage to Orbit rocket car.
Why would you put data centres in space?
Seemingly to prop up the value of your companies
Maybe this is beyond parody…
https://x.com/elonmusk/status/1005577738332172289
Self-driving single stage to orbit rocket cars. It can drive you to work and then go on a Starlink launch to LEO and be back in time to pick you up.
Add your car to the SpaceX fleet and get paid to own a Tesla!
Right, just to meet the most minimal of the scenarios for datacenters, someone upthread has calculated "launches every 9 hours, 24/7" as a minimum.
Is he also talking about moving X's servers (since xAI owns X) into space?
Friendly reminder for anyone that forgot - xAI acquired Twitter, so now Space-X is the proud owner of a dying social media platform that they overpaid for.
Any claims that this is about putting compute in space is just a non-sense distraction. This was absolutely about bailing Elon out of his impulsive, drug-fueled Twitter purchase.
The only question now is: when they try to go public, will they be punished for wasting so much money or not? My guess is: not.
SpaceX has jumped the shark.
Wait so now SpaceX is the company letting you undress people? Why mix gold and poop? SpaceX plus xAI.
Makes about as much sense as Twitch buying Curse about.. a decade ago?
Musk’s schtick lately has always been to find a challenging problem, point everyone to it and say “ I will solve it by September “, have the stock shoot up and make money. His first dis that with self driving, then twitter, then xAi, and now robots and data center in space. These last two will last him a decade as these are both challenging problems to solve over night.
Purely financial shenanigans. Nothing to see here, please move along.
"Thou shalt not make a machine in the likeness of a human mind."
And so it began. The seed was sent into space. All going according to plan.
So Elon has more shares when SpaceX IPOs?
So they use a valid and valuable company to hide a giant dumpster fire company. To add to that, their best argument is "AI in space", which has some real "solar roadways" energy to it. I honestly don't know how any SpaceX shareholder could approve this.
Perfect timing to offload that debt onto the bag— I mean shareholders.
We just had an X8.1 CME event, I just want to point out that at any moment we could have an x40 (we had a carrington event already in 1859) or higher event and all those sats at low earth orbit would be fried and start hitting each other, if SpaceX keeps launching more it becomes incredible probable that we might hit the Kessler Syndrome, and we would legit lose access to Space for a WHILE, including all of what satellites entail.
Are we ready for that as a modern society or are we going to start enacting regulation against it? I'm sorry but people wanting internet everywhere does not justify we going back to the dark ages for a decade or more.
US is dialing back even on global warming. There is no chance current government would risk space superiority for some kessler-shmessler that nobody has ever seen.
Now he just needs to work in crypto satellites down to users via all the new phones supporting satellite link to SpaceX. I kinda expected that one first. Distributed payment network outside of government control/oversight seems like something he would be in to.
Pretty terrible for SpaceX. Of course they paid a crazy inflated price for xAI in an attempt to cash in on the IPO. This just devalues SpaceX and exposes the investors to all the AI bubble risk.
What kind of financial engineering is going on here? Is xAI about to go bankrupt or something?
Probably. They are building the largest data centers in the world with very little revenue.
I asked Gemini for a two word summary and it wrote "financial engineering"
Grok: Integrated Ambition
The next step will be merging SpaceX and Tesla.
Tesla has probably the most valuable shareholders on Earth. Over years of empty promises and meme status, the stock has pretty much purged all the level heads. So it's mostly deluded Elon sycophants giving placing their tithe on the alter of his sci-fi fantasy smoke and mirrors game.
In reality he will be dumping the debt of twitter and xAI (and maybe spacex?) on Tesla shareholders, and buoying that with the added layer of hyper that spaceX brings.
> SpaceX has acquired xAI to form the most ambitious, vertically-integrated innovation engine on (and off) Earth, with AI, rockets, space-based internet, direct-to-mobile device communications and the world’s foremost real-time information and free speech platform. This marks not just the next chapter, but the next book in SpaceX and xAI's mission: scaling to make a sentient sun to understand the Universe and extend the light of consciousness to the stars!
I think Elon's taken one too many puffs of hopium
You'd think he'd have a pretty huge tolerance at this point.
So he doesn't want to go to mars he wants to make a big space chatbot?
Plus I suspect the sun is already sentient no need to reinvent the sun
> I mean, space is called “space” for a reason. [Face with Tears of Joy]
Man the real story in a few decades is going to be whether SpaceX was onto such a kille business that it survived being used asma fiscal dumping ground for the losses being incurred by mismanagement everywhere else.
Can we at least agree this is an awesome thing to try?
Way more exciting than spending $70 billion on VR like Meta did when we all just wanted to play games.
Reminder that space only allows for radiative cooling (since there is no air to absorb heat) so data centers in space are going to have massive cooling panels.
In other news, Kessler Syndrome: https://www.youtube.com/watch?v=8ag6gSzsGbc
Don't forget to opt out of SpaceX's product Starlink using your data to train AI: https://www.pcmag.com/news/starlink-wants-your-data-for-ai-m...
xAI owns Twitter... So now space company owns Twitter? Wtf
From a technical point of view this doesn't make any sense.
From a finance and accounting point of view this makes everything more cloudy. Which certain types of people really like.
Reminder that SpaceX has received an estimated $38 billion in government funding over the years, and all of its returns are going to a small set of private investors.
Socialized losses, privatized profits. As is the American way.
What if enemy or some anti-ai activities or etc attack it? How to protect it? It's just a too easy target.
It's just a dumbest idea ever if Elon truly believes it. I'm pretty sure he doesn't.
What about security?
It's a CONSTANT stream of new ideas with no payoff at this point.
Hyperloop > Neuralink > Self-Driving Cars > Robotaxi fleets > Personal Robots > Orbital Datacenters > [insert next vibe shift]
At what point do people start to see the ever-shifting goalposts for what they are?
Hyperloop was never a company project, neuralink was a separate company, tesla is rolling out driverless robotaxis and fsd is amazing, robots inevitably are going to do the majority of work - there’s no real doubt about it is there?
Datacenters in orbit seem insane so idk we’ll see
> robots inevitably are going to do the majority of work - there’s no real doubt about it is there?
There's a lot of doubt that the AI and compute to enable that would happen on commercially relevant timescales.
Consider: "do the majority of work" is a strict superset of "get into car and drive it". The power envelope available for an android is much smaller than a car, and the recently observed rate of improvements for compute hardware efficiency says this will take 16-18 years to bridge that gap; that plus algorithmic efficiency improvements still requires a decade between "car that can drive itself" and "android that can drive a car". (For any given standard of driving).
And that's a decade gap even if it only had to do drive a car and no other labour.
You can't get around this (for an economy-wide significant number of androids) by moving the compute to a box plugged into the mains, for the same reason everyone's current getting upset about the effect of data centres on their electricity bills.
And note that I'm talking about a gap between them, not a time from today. Tesla's car-driving AI still has safety drivers/the owner in the driving seat, it is not a level 4 system. For all that there are anecdotes about certain routes and locations where it works well, there's a lot of others where it fails.
That said: Remote control units without much AI are still economically useful, e.g. a factory in Texas is staffed entirely by robots operated over a Starlink connection by a much cheaper team in Nairobi.
1 reply →
Don’t forget LEO satellite internet! Oh wait….
Musk earns a $1tn payout when Tesla hits $8.5tn dollars.
I expect the next step in this series of moves is to turn Tesla into a SPAC & have it acquire SpaceX, bringing its valuation nearer that 8.5t.
What this tells me - xAI is essentially a failure, though at what level I'm not sure.
Major grift vibes, inventing half-baked reasoning to justify massive valuations. If money wasnt so deeply entwined with politics at this point, this is the sort of news that would launch fraud investigations.
Terrible news for SpaceX.
Could this spell the beginning of the resurrection of the Nazi Space Program?
The way I read it, X AI is not really profitable and Elon's creditors/coinvestors asked for something tangible for their money, a.k.a shares in SpaceX, his only business that still has some solid foundation. The rest is emois.
Discussion on previous speculation: https://news.ycombinator.com/item?id=46814701
“SpaceX is doing great but I want some of that AI investment money.”
When Elon Musk talks about benefiting humanity, remember that the one time he had unregulated power (DOGE) he used it mostly to cut benefits for poor people, and to push ideological agendas. His only agenda is self-aggrandizement, and this announcement is cover for passing the hot potato of Twitter debt around.
Whenever Elon, Anthropic or whoever mentions humanity, they don’t mean every 8 billion of us on Earth. They mean themselves, their mates and whoever they deem worthy of the posterity they believe only they have the merit to design.
Musk always merges his companies when one is suffering:
Twitter/X in xAI
SolarCity into Tesla
xAI into SpaceC
I am just waiting now for Tesla to be acquired by SpaceC as it has run into issues.
What does Gwynne Shotwell think of this. She seemed level-headed, but is she also batshit insane now by osmosis?
xAI to cover X investors, SpaceX to cover xAI, us American public investors to cover SpaceX since it cannot go under for strategic reasons. Ultimate grift.
The whole universe was supposed to be turned into paperclips, now it is being turned into graphics cards to produce images of barely legal girls on X.
And Musk keeps grifting about Kardashev 2 civilizations while his rockets do not even reach the moon.
If SpaceX goes public, that will rescue his xAi shares. I wonder how he will rescue his Tesla shares.
What metric does reaching the moon fulfill?
Nothing, save for advertising that you can. And Musk obviously can't, or he would have by now.
> the world’s foremost real-time information and free speech platform
What a joke.
I thought this wasn't viable due to cooling requirements - how do you cool massive amounts of compute when the only option is to radiate it into space - nothing to convect it with?
Also, the incredible amount of grift here with the left hand paying the right is scarcely believable. Same story as Tesla buying Solarcity. Board of directors should be ashamed IMO.
Yes. It is very cold up there but there is also no matter, or very little matter. So head conduction and convection don't work, it's all radiation. When we are learning to solve heat transfer problems in engineering school we are generally taught to neglect radiation, because it's effect on cooling the system is typically second or third order when compared to the to "big C's"
It would take roughly 5000 square meter area to cool a typical small data center heat output (1 MW). Not great, not terrible.
4 replies →
Cooling and maintenance (part swaps, etc.) are one of many obvious reasons why this is bullshit.
Doesn't stop grifters, tough.
in actual datacenters you often don't even bother swapping parts and just let things die in place until you replace whole racks
1 reply →
hmmmm
Elon Musk is a genius, but he’s a financing genius. Look at the long history he has of false promises supporting financing deals between his companies and you’ll see this for what it is, a cash injection and a lie to justify it. He did the same thing with a fake solar roof demo when Tesla bought the almost bankrupt Solar City. He also shifted resources from Tesla and SpaceX to support X in the early days. Even founding xAI outside of Tesla, when so much of its valuation was built on its AI capabilities, was questionable.
Now those dumb articles about AI data centers in space from a couple months back make so much more sense.
Who got the money? Hahah
Making a "sentient sun" is the most bald-faced asinine, drug-induced nonsense that should be the complete destruction of all credibility to anyone who said it or typed it or works for anything here.
Financial engineering. Twitter under Elon became a dumpster fire of porn and hate and big banks were holding 13B in bonds that wouldn’t be worth the paper they were printed on for the company alone so he just links it with his only company that actually is doing something worthwhile…
Not sure how X which “merged” wit X (formerly Twitter) and SpaceX really matter or synergize but here we are. It’s all about the money being protected. And this Ketamine using wierdo is gonna be the worlds first trillionaire. Yay all of us.
fool me once (Solar city) shame on you, fool me twice...
> By directly harnessing near-constant solar power with little operating or maintenance costs, these satellites will transform our ability to scale compute. It’s always sunny in space! Launching a constellation of a million satellites that operate as orbital data centers is a first step towards becoming a Kardashev II-level civilization, one that can harness the Sun’s full power, while supporting AI-driven applications for billions of people today and ensuring humanity’s multi-planetary future.
Apparently optimus robots don't work and he needs to start his final grift, space datacenters, while his datacenters on earth are powered by gas turbines.
Most likely he's just trying to bury his epstein involvement where was exposed lying by his own daughter.
Can someone convince me that this is not a) pure horseshit b) a plan for Elon to sneak enough mass into orbit to hold the Earth hostage? If you can bring millions of tons of anything into orbit around Earth you can destroy civilization, or just France.
knowing elon, he will make this actually work, thus fully vindicating both the financial engineering and his arrogance!
Disgusting.
When does the market realize this is all just a shell game and the emperor really has no clothes?
We saw this on a much smalelr scale a decade ago when one of Elon's companies (Tesla) acquired a second one of Elon's companies (SolarCity) because it was broke and owed a ton of money to a third one of Elon's companies (SpaceX).
Elon was forced to go through with his impulsive Twitter acquisition by a Delaware court, an acquisition that was not only secured by a bunch of Tesla stock but also a bunch of Qatari and Saudi royal money. He then mismanaged Twitter so badly Fidelity wrote down its value by at least 80% [1].
So what did Elon do? Raised even more questionable foreign money into xAI, diverted GPUs intended for another of his companies (Tesla) into Twitter and then "merged" Twitter into xAI, effectively using other people's money to bail him out from an inevitable margin call on his Tesla stock.
Interestingly, Twitter was reportedly valued at $33 billion in this deal [2], significantly more than the less than $10 billion Fidelity valued Twitter at. Weird, huh? With a competent government, this would be securities fraud that would have you spend the rest of your life in jail. And even with all that, $11 billion was lost on the deal.
So here we are and it's time for the shell game to be played again. Now it's SpaceX's turn to bail out the xAI investors.
And what is the argument for all this? AI data centers in space. Words cannot describe how little sense this makes. Launch costs (even if the Starship launch costs get to their rosy projections), cooling in space, cosmic rays (and the resulting errors) and maintenance. Servers constantly need parts replaced. You can just deorbit the satellite instead but that seems like an expensive way of dealing with a bad SSD or RAM chip.
[1]: https://www.cnn.com/2024/10/02/business/elon-musk-twitter-x-...
[2]: https://www.cnbc.com/2025/03/28/elon-musk-says-xai-has-acqui...
> This marks not just the next chapter, but the next book in SpaceX and xAI's mission: scaling to make a sentient sun to understand the Universe and extend the light of consciousness to the stars!
One of the dumbest things I've ever read.
Musk: "What do we have that OpenAi doesn't have?"
Musk aide while high: "sPaCe"
Slightly less high Musk aide: "But what is the synergy, where's the moat and how could that be done in practice and most importantly is there any limiting factor on Earth before we have to bring AI into.."
Musk : "SPACE!!!!!!"
It is incredible to think that the extremes of the stock market are actually pretty similar, pink sheets/cryptos and these mega companies are actually the same. News fueled pumps and dumps to win the cycle of hype of the week
They were all inside his head or bots replying to him on Xitter
Ahahaha, who got the money?
Going to be marked at some delusional valuation and at IPO retail bag holders are going to get absolutely massacred.
Like Musk's other great public company failures?
They are failures as far as products
1 reply →
Tesla is marked at purely delusional prices as well. Total hopium on Optimus taking off while their core business craters.
1 reply →
This means that Grok, Elon's politically labotomized involuntary pornography generator, and X (formerly Twitter), Elon's Nazi-adjacent propaganda machine, are now completely intertwined with SpaceX, a too-big-to-fail government contractor that currently serves as America's only reliable option for manned orbital spaceflight.
Anyone who doesn't see how broken this situation is isn't paying attention. This is how people like Elon, who want to seize as much power from the government as they can, ensure that the means for seizing that power are untouchable.
Anyone who has ever used Grok or X lately knows that both of these products are heavily manipulated to align with the political, social, and economic views of Elon Musk, who is increasingly boosting "white power" language and full-throatedly backing America's most nationalistic and authoritarian president to date.
This is just another consolidation of power, and it's deeply worrisome. Any integrity one may have hoped remained at SpaceX just vanished when they aligned their mission with that of these deeply problematic digital services.
And this is not even scratching the surface of what looks like a deliberate attempt to create Kessler syndrome by launching millions of cheap short-term satellites into orbit, or the rationality of putting datacenters into orbit in the first place...
Seems more risky to me; it takes only one Mercurial temper swing for SpaceX to be nationalized, and now Twitter and xAI are bundled along with it.
[dead]
[dead]
https://x.ai/news/xai-joins-spacex wrong link?
I'm sure these comments will be rational and non-biased by political emotions /s
Well surely this acquisition is above board. Nothing funny going on here, just good old business as usual.
> Well surely this acquisition is above board.
What makes you think it isn’t?
There's an epic conflict of interest here with Musk owning most of both companies. And they're in entirely separate fields, there is no plausible synergy here to be gained.
10 replies →
Musk has a history of having one of his more successful companies buying one of his less successful companies. xAI bought X, and Tesla bought SolarCity
Musk is notorious for shuffling assets across his companies to make some financials look better. For example, shuffling Twitter servers (and then all of Twitter) under xAI.
3 replies →
lol
Musk's involvement for one.
It's just "Elon zomg lulz" trolling for updoots. This place pretends to know better.
Well, at least this time both companies are fully private.
For the next four months, at least.
What's funny? Do you think the investors are against this? The investor's aren't idiots. I imagine the typical investor in Elon Musk's companies would approve of this sort of thing. So what's the problem? Besides, its a private company with Musk as majority shareholder in both. That's the beauty of private companies, you can just do things.
I wish more companies were private and ambitious. I'm tired of companies like Apple making marginal spec bumps to their phones and milking the same products for decades
> What's funny? Do you think the investors are against this? The investor's aren't idiots.
Any proof of that?
4 replies →
Apple just launched their own silicon chips just a few years ago. They're very ambitious but still calculated.
> > I'm tired of companies like Apple making marginal spec bumps to their phones and milking the same products for decades
At least what Apple does is real not make believe like everything Musk claims , disappear boring Apple or even boring Microsoft, Oracle, IBM etc.
And the world would come to a screeching halt, disappear all of Musk companies and people would barely notice.
You seem to be eager to be sold dreams , that's exactly what vaporware salesmen like Musk hope to find on their path
The investors want to cash out, Musk needs lots of money to plow into his latest toy that so far only excels in ridiculing him and sexual harassment/CSAM, so they make a deal to take in xAI and go public. Win win.
> The investor's aren't idiots
citation needed.
It's widely reported that Musk is a majority shareholder of xAI and the controlling shareholder of SpaceX (close to 80% of voting shares). Not surprising that he would be looking to consolidate ownership under one entity especially if he perceives significant synergies (i.e., data centres in space).
Data centers in space are a hilariously bad idea. Where would the heat go? This idea is like the opposite of liquid cooling.
10 replies →
I have never understood how Data centers in space ever make economic sense, the payload, latency and many other issues make it difficult at least for the immediate needs
2 replies →
[dead]
[flagged]
[flagged]
[flagged]
Do you have any parties planned? I’ve been working to the edge of sanity this year and so, once my kids head home after Christmas, I really want to hit the party scene in St Barts or elsewhere and let loose.
wow this is one of the most incredibly fraudulent things that ever happened in American capitalism and I'm not ignoring Enron or the mortgage meltdown. I'm speechless the US has given up on any semblance of law and order in matter of financial markets and this stuff can happen without people going to jail.
Besides the obvious, why do engineers with real skills still work for this guy?
People pooh-poohing space datacenters will obviously think this is a bad move. But Elon clearly believes space datacenters will work. Given that, and the fact that SpaceX will IPO this year, this acquisition was inevitable.
SpaceX and xAI would not be able to freely collaborate on space datacenters after the IPO because it would be self-dealing. SpaceX likes to be vertically integrated, so they wouldn't want to just be a contractor for OpenAI's or Anthropic's infrastructure. Merging before the IPO is the only way that SpaceX could remain vertically integrated as they build space datacenters.
How can you clearly believe anything Musk says at this point. It's not 'clear' at all what he actually believes and what he just makes up.
Space datacenters just so grok can undress women? this is the dumbest company on the planet