> In the long term, space-based AI is obviously the only way to scale.
> My estimate is that within 2 to 3 years, the lowest cost way to generate AI compute will be in space
Yes, Elon is very sane.
> In the long term, space-based AI is obviously the only way to scale.
In the long term - all mass and energy available is outside of Earth - what is here is not even a rounding error. If you wish to continue scaling compute it then becomes a question of time before you'd want to go off planet. Personally I'm quite keen to see near term space based compute explored, as it could end up becoming a much better trade-off than allocating ever more ground to power and operate terrestrial compute which directly conflict with the biosphere.
SpaceX started the Starlink design phase in 2015 - started launching Starlink satellites in 2019 - and they now have the most dominant satellite constellation ever deployed by a large factor. They have their own launch systems, launch sites, satellite bus, communication stack - both in-house designed and built.
What is really going to be that difficult with space-based compute? Radiation hardening and cooling? These are clear engineering challenges that can be simulated, tested with earth analogs, and then rapidly iterated across design generations. There's napkin math all over the internet on this, but it really seems like small challenges compared to the other engineering SpaceX have already sorted.
Beyond radiation / cooling / servicing - it seems like the biggest hurdle is to crack the scaling of designing / scaling the necessary amount of compute they will need to scale space based compute according to the laid out plans.
In case anyone is wondering how Tesla’s stock price remain wildly detached from its business reality, keep these four words in mind. If you can convince people that anything about you and your business has to be evaluated on a literally astronomical timescale, you can justify any valuation you desire, because your believers will give you infinite time to realize their investment returns. It has nothing to do with business. They are selling you a vision — which can also come in a pill form, labeled "salvia" and sold at gas stations.
I still see people say the cybertruck is built for mars environments, conveniently ignoring the vast technological and economical barriers stopping us from driving commercially produced vehicles on mars. This space data center thing is the same deal. It doesn't matter how long it will take to solve the technical issues with cooling, radiation, maintenance. It doesn't matter if it will make economical sense or not. It doesn't matter if spacex will be the one to actually do it. You just have to believe, and give them some time — a lot of time, so much time that a monkey can type out Hamlet and type it out again backwards.
See also the buffoonery coming out of Bay Area "effective altruist" and "longtermism" communities.
you also shrug off cooling. this is not a solved problem in any way. its not even approachable as of yet. the vast size of the radiators will be hilarious regardless.
you ignore power generation. solar is not an option. so we also need nuclear reactors for these orbital data centers. thats cool spacex can just branch out into nuclear too! love the idea of unmanned nuclear orbiting behemoths.
speaking of orbital.. what is their orbit? do they go out to Lagrange points? hilariously far? or do they stay close? hilariously fuel intensive to stay out of the atmosphere for such massive structures?
but hey, maybe we distribute spaceX-AI gpu's across starlinks. a couple solar panels and a tesla battery per gpu. all launched there by spacex
'all mass and energy available is outside of earth' Yeah, and out of range for compute data connections too.
I don't agree with the feasibility or ANY sort of practicality to this whatsoever. Im all for going for it, but I wish everyone could just admit that we're doing it because it's cool, not because it's useful. I get why Elon wont say that, but not us.
Every single satellite has sufficient cooling for its power production, otherwise they would be frying. Waste heat from a GPU is not materially different from waste heat from an amplifier. That's not cooling entire racks, but I don't think anybody talks about putting entire racks in space anymore.
I'm very much pro nuclear, but a solar cell in a sun synchronous orbit is pretty great too and eliminates most battery requirements
I very much doubt the economics of this makes sense, but I don't think a lot of your criticism is valid.
Manufacturing capabilities are quite lacking, though, in the short and medium terms, so this doesn't seem all that relevant.
Maybe a self-contained, modular solar panel / radiator / compute unit could be built, but it will be manufactured on Earth. (Where the fabs are.)
And it still seems easier to put solar panels and batteries near the data centers that SpaceX is already building on Earth.
Instead of wasting huge amounts of land to farming, restaurants and transportation of food it would be so much better if everyone just had a Star-Trek style food replicator in their house.
None of the tech exists but fuck it. Why bother with realities of life?
I am raising 200 Trillion Dollars for AI Space FoodX. Who is in?
Seems more like a grift to me, after the car grift and the Mars grift didn't pan out.
Unless someone figures out how to break the laws of thermodynamics there's never going to be a cost effective DC in space.
Edit: https://en.wikipedia.org/wiki/Planetary_equilibrium_temperat... A blackbody sphere near Earth's orbit balances out to almost exactly 0C. A sphere has about 4x as much radiating surface as capturing surface. A flat surface facing the sun that would have 2x, front and back.
They can. But in Elons case, its going to be his style of sending failure after failure up in the space, getting something working part time, lying about it and exaggerating how good it is, and then making fun of others for not using his inferior product.
Pretty much like everything else he has done.
yawn, people keep making this excuse on behalf of the South African investor with poor technical expertise.
tbf, a 'sane' person wouldn't have started a rocket company and an ev company, at the same time, in a recession.
He has never been sane. and that has made all the difference.
This "space datacenters is more important than colonizing the universe" thing is just to deflect from what would be an inevitable failure because if they do this pivot, they can push out the timeline for that further than the original 2026 on Mars goal that they are about to wildly overshoot.
In other words good old fashioned plausibly deniable securities fraud.
It isn't... the hostile local government can seize the ssh keys you use to control it and take it over just fine.
The hostile international non-local super power just gained a new ability to jam communications or destroy it with a bit of deniability too.
Local governments in the US practically never exercise control over data centers by doing any of the things you just discussed. There's a reason why you're saying "this has been threatened". It's a strange new thing resulting from bizarre current behavior - behavior and a resulting trend that started after Elon started talking about space based data centers, and thus cannot be the cause of it.
I'm not caught up entirely, but I would imagine that NSA's capabilities have advanced beyond what has been published from slides created nearly 20 years ago.
you mean other than China, russia, NK and Iran?
I think if fusion is real, it might not be so advantageous until space mining is a thing.
We're simply out things we can profitably send to space so SpaceX and others are trying to come up with ideas to induce demand.
My understanding is that Starlink mostly grew out of the same need to justify scaling up rocket production.
Other than the occasional GNSS, weather, scientific, broadcast and surveillance satellite, there's not all that much worth sending into space.
I can only assume "too easy to track" is part of the logic.
Ditto for kinetic strikes. That was super hyped up.
Yeah, that's why it'd be a good way for SpaceX to make money.
IDK I think plenty of people will want to go to space or even cut 24 hour flights across the world to 90 minutes.
As for experience - it's going to be pricy, but look how many multi-million dollar yachts are out there, parked, doing nothing. People do have money for such experiences.
Sonic boom can't be the limiting factor forever.
Hell, it would be cheaper to figure out how to build them on the ocean.
I recognize that that is distressing to people, hell, it’s been obvious to me since I was at OWS in my 20s. But we are in a new world now and the old rules don’t apply. A company that has the backing of the government to launch their spacecraft will simply do it. You think Texas is going to stop them? Or Florida? Or even California? Of course not.
A lot changes in a world where you can plan things out with AI. A lot changes in a world with abundance. If we play our cards right we could have the culture, but that means letting go of the conservative yearning to put things back to how they were. The old world is 10 light years away now, it wasn’t as great as we remember it and it ain’t coming back.
And if I had to choose, I’d much rather have datacenters in orbit than one burning hydrocarbons loudly 2 blocks from my kids’ school.
Yeah, but that choice is nonsense. Mandate that datacenters on the ground are on 100% green power and quiet, and they'll still be way way more cost effective than the orbital option.
Looking at things right now? I would say no. We will see, maybe in up my own ass on this, but I see a pretty big set of changes coming down the pike. Adapt or die (as unpalatable as that may seem).
So what kind of laws would lead to the orbital option being preferred over the ground-based clean option?
Pendulums swing. Anyone advocating for the development of more advanced technologies should be in favor of a system of fair laws enforced robustly. One need only look to countries that lack this foundation to understand why.
So I'll just say the same sentence again. The NIMBY factor isn't so powerful as to keep datacenters off entire continents.
https://www.chaotropy.com/why-jeff-bezos-is-probably-wrong-p...
Like, it's not "great" but if you're not flying around the sun every 72 minutes or whatever and you can keep your panels sun on and radiate into deep space, the numbers aren't bananas.
And you need to get the heat away from the central point to the extremities of the radiator as much as possible. So you can maximize how much energy can be radiated away.
Seems like the weight of the system would be an issue with whatever gas or liquid you used to fill those radiators, but maybe I'm wrong...
"Is It Really Impossible To Cool A Datacenter In Space?" - Scott Manley
tl;dr -> not impossible.
But anyway, conservatively, about 20 tons each, it seems like you could fit at least 5 of these per starship, assuming it's weight and not volume limited. Doesn't seem like fuel's a prohibitive portion of the cost here. But if they can't get it to their no-refurb-between-launches target, then that might be a significant part of the cost.
Its to the point where anything he says is guaranteed to be wrong just on the merit that its coming out of his mouth.
in the very broad shoulders of long term, he's probably right.. its why the concept of a dysonsphere is around. you can get uninterrupted 24/7 free energy.
but yeah, the tech is a long way away.
*Edit: lol My estimate is that within 2 to 3 years, the lowest cost way to generate AI compute will be in space.
i think 2-3 years is a very unlikely outcome.
Freeman Dyson invented the concept as a joke against SETI, especially designing it to sound quasi-plausible.
In reality, there is no way to create a stable structure of this size, it would be like trying to balance a building on the top of a pinhead - except the pinhead is a chaotic, unpredictable star. And the amount of energy required to displace multiple planets worth of mass, manufacture some amount of it into complex satellites, and then displace this amount again to a "stable" Solar orbit simply doesn't exist in the Solar system, on any plausible time scale (it would take many thousands if not millions of years worth of solar power to do so).
But 2 to 3 years?! Seems crazy
We used to eliminate Nazis, not invest in them.
US history is more complicated than that, and aside from those four years of hot war, more ambiguous.
Henry Ford was a big Nazi sympathizer, and the Apollo program was led by an actual card-carrying Nazi engineer with a history of overseeing slave labor in a concentration camp.
Which is not meant to defend Nazis, just correct the myth that the US was once somehow morally pure in this regard.
For all the lies, bad behavior, and broken promises, SpaceX's achievements and reliability record is still incredible, X/Twitter hasn't crashed and burned after all the layoffs and drama, and Tesla (until recently due to his meddling) had a lock on the leading the car industry's direction & doing a lot to drive practical electrification globally.
Every journalist keeps talking about "data centers" in space, like they're giant buildings with square miles of solar arrays.
"Hurr, durr, the tech bros are clearly crazy!" say the arts majors that didn't listen to, or understand the talks by the people proposing to put AI compute in space.
What Elon and co are talking about is simply a matter of putting a somewhat bigger solar panel on a Starlink satellite along with a handful of inference-optimised neural processor chips onto its motherboard, side-by-side with the existing electronics.
They're just talking about launching a slightly bigger and slighty more power hungry version of Starlink. That's it.
"We'll need thousands of them!
Yes, they know.
Starlink is already planned for a scale of tens of thousands of satellites. With Starship V3 scaling it to hundreds of thousands is (almost) no big deal.
The number of political actors that can stop you from building in Morocco (or confiscate/damage your invested capital once you deploy it) are numerous. The number that can do so in space? Maybe a half dozen. We’re already seeing states and municipalities in the US moving to ban data centers and the energy infrastructure needed to power them. Building in space faces no such procedural roadblocks.
The economics still seem like an open question, but if the demand for compute is high enough, space based data centers might be the only option
Not sure about the cost perspective but, at least that makes more sense than a giant brick floating around.
[1]: https://lafibre.info/hetzner/over-200-000-servers-in-one-pla...
> Yes, they know.
> Starlink is already planned for a scale of tens of thousands of satellites.
Meanwhile Google installed that many TPUs yesterday afternoon. The idea is still stupid.
The availability of power is the constraint almost everywhere, no matter how much money you throw at it.
Gas turbine production has a many-year backlog. Everybody that can make the single-crystal superalloy turbine blades is fully booked for most of a decade and can't expand capacity for years (at least).
Meanwhile, putting a slightly larger solar panel onto a satellite is a trivial engineering excercise and has no blockers in 2026.
Disclaimer: Personally, I suspect all this AI-in-space "talk" from Elon is just cheap marketing to boost the IPO of xAI.
Is the sunlight millions of times brighter beyond the atmosphere? I don’t get it.
Also, there's a populist backlash on building datacenters, power transmission infra, and power generation in many areas on earth. Locally, we have a number of people complaining about solar arrays going up on farmland, even though it's the farmers choosing to do it. "It's an eyesore".
Example of a spec sheet: https://signaturesolar.imagerelay.com/share/ffc69ee2265b4613...
If you mean the farmers' arrays, those are meant for commercial generation, so a good bit bigger, but one nice thing about solar is it's extremely modular, and you can fit it to the land. I believe bigger panels are more common for commercial, but I think it's a lot nicer to handle 40-50 pound panels than 70 pound panels.
Because panel cap factor is about 10-20% to begin because day and night exists on earth. Say you wanted to power it on solar + batteries and picked Australia. You pick place that has decent port and most exposure, i.e. Port Hedland. In winter, daily average drops by 20%. Also because atmosphere - 30% less insolation when compared to space. Finally add 10-45% cooling losses.
Which effectively means you need something at least 10-20x more panels + batteries to match space.
Will we get to Mars soon? Hell no. But we may end up with a world-leading launch provider based in the US and that's a clear win for the country.
At least he has B.Sc. in physics and got admitted into Stanford.
I think what Elon says is better explained not as a promise what would happen, but rather as a goal which they're going to aspire to. It kinda supports the idea "we're in business of converting impossible into late". If Elon will start offering more "realistic" schedules, the pace of SpaceX will slow down, perhaps considerably. So, yes, it's "Elon time", which historically isn't particularly precise, but still useful.
And the Stanford admittance was for materials science, not physics as he lies about
I've got no opinion on the existence and legitimacy of any degrees Musk may or may not have, but whichever he does have you really can't infer much at all from whether a STEM degree in the US is a BA or a BS without looking at the specific requirements for the degrees at the particular school.
Some schools might give a BA for a program and other schools might give a BS for a nearly identical program. All of these happen in the US:
• BS is the only choice. (Caltech, for example. In fact, Caltech only offers BS for everything. Even English majors--and yes, there is an occasional English major at Caltech--end up with a BS).
• BA is the only choice. UC Berkeley is an example in this category for math and physics.
• Both are offered, with identical coursework and requirements. You can have whichever you want. Some will even for a small fee give you two diplomas, so you can use whichever seems appropriate for the situation.
• Both are offered, from the same department, with different in-major coursework and aims. One may be aimed toward students aiming to go into research, and one toward those aiming to go into teaching, for instance.
• Both are offered, from different departments. For example, UC Berkeley's College of Letters and Sciences offers a BA in chemistry, and the College of Chemistry offers a BS in chemistry. Computer science can be taken at Berkeley in the College of Letters and Science for a BA, or in the College of Engineering for a BS.
• Both are offered, with the same in-major coursework, but differ in out-of-major requirements. So, the BA and BS would require the exact same science and math courses, but the BA has specific breadth requirements to produce a well rounded education, whereas the BS lets you take pretty much what you want as long as you satisfy the math and science requirements and any general requirements of your school.
Chronic over-promise, underdelivery.
Where was the nearly 3T of fraud he said he'd uncover in the US government, again? Was that a clear win for the country?
But hey at least he's effective at getting people to give him money, I guess, which is an indistinguishable "skill" from that of someone who is able to convince people to buy an online course on how to make money online.
He just does it at a bigger scale so people are quick to suck him off. How we are still falling into the "money = smart/competent" trap in <<current year>> is beyond me.
I have immense appreciation for what SpaceX has done for humanity. I’m not being dramatic. Reusable rockets alone is an incredible achievement. But he’s lost the plot. He needs to drop his right wing bullshit and stardom chasing if he wants to be taken seriously again. The dude won’t even acknowledge his own kid because of his politics. I will never trust someone who makes that decision, personally. His judgment is beyond clouded.
The Elon bros will be mad but whatever. One day he’ll maybe remember why folks liked him. Hitching his wagon to Trump was a dumb move.
If you think musk hasn’t banned people for bullshit you’re not looking at all. The site has suspended literally millions of people since he took over. He banned the jet tracker by creating a curated doxxing policy specifically designed to cover his ass.
You need to spend 5min with a search engine. The myth that he has made it more open and free speech friendly is just that.
None of those are self sustaining.
You don't want to be there? Almost every other place on earth is better. So you send a skeleton crew along with what they need.
If it is to test an actual community living isolated, sure. But I think it'll always be different because you know that help is at most a few months away and probably a lot less. I don't think you can fake that, unless you're never told you're not alone
No it doesn't. "Permanent settlement" just means it's not temporary, only intended for a short-term mission.
https://en.wikipedia.org/wiki/Earth_Overshoot_Day
I'd keep the Moonraker film in mind as a metric for self sustaining colonies created by billionaires. They can't be trusted unless they are also working to fix what we already have.
Of course, there are still a lot of unknowns, any of which could prove fatal to the concept but I'm no longer comfortable just dismissing it as "obviously ridiculous."
And if he's actually capable of producing solar panels in the quantity that he's talking about in the time frame that he's talking about -- why doesn't he just put them on earth to solve our growing climate change problems and fuel shortages?
Well, yeah but that's just Elon being Elon. At this point I think even the most pro-Elon folks freely admit "The first rule of Elon is: 'Ignore everything he says about timeframes and scale.'"
I guess the focus is going to be on getting stuff up, rather than back down. Thus the Starlink and data center plays, not human space exploration.
I imagine at least some of the reason to chase the AI datacenters in space thing is because Starship is "too capable" if it succeeds. It makes available a technology that does not have a short-term utility that people will pay for. Starlink was something that's been useful as telecoms but perhaps that market is saturating. It makes sense to pursue what is currently high-utility but is not being met because of terrestrial constraints.
Well, good luck to him. A lot of smart people are chasing this idea and I can't seem how it could work, but I was honestly surprised that Tesla hit its production goals, and I was honest surprised that SpaceX hit success so fast, and I was honestly surprised by the rise of LLMs, so the truth is there are lots of paradigm shifts I just miss: BEVs, cheap space, AI.
Someone once tweeted something like:
> Less intelligent people perceive more intelligent people as incredibly lucky. They always make inscrutably stupid decisions, unjustified by visible information, and somehow fate rewards them for this.
But also, I'm just hoping that a new era of space exploration will open up in my lifetime. That sounds incredibly cool! And I dare say there are many people like me in the US at least judging by the popular baby names of this era, which have seen spikes in Aurora, Nova, and Luna - and in the one my daughter has: Astra.
The only problem that "data centers in space" solves is the problem of trying to scale a rocket company where the potential demand for rocket launches is simply not that big.
> Liftoff will occur at 6:30 p.m. ET on Monday (May 19)
https://www.space.com/space-exploration/launches-spacecraft/...
It's a difficult engineering challenge but physically possible, and Elon is no stranger to engineering challenges.
Some numbers: assume an emissivity of 0.85, assume no absorption from the sun, assume heat rejected from both sides of a panel, a 1m^2 panel will reject 1.45kW/m^2 @ 350K.
At 900K its 62 kW/m^2. Not a trivial amount of heat.
The rise of moralization of everything is really killing online discourse. It's gotten to the point where people will now mostly criticize and support ideas based on who proposed them, and not based on their merits. Tribalism at its worst.
I think religion helped reduce tribalism, at a societal level, by making evil/demons/bad acts as the "them" and everyone that went to church on sunday (it was the whole town previously) was the "us". Now, without religion, and the physical/social bringing together it brought, that hardware in our brain still tries to segment a clear "us"/"them", but with much less guidance.
I too would like it to just be about the content, but nothing exists in a vacuum.
A lot of Buddhist practice is basically trying to train against immediately collapsing reality into self/other, right/wrong, craving/aversion.
Practicing this with Elon Musk is effectively ultra hard mode.
--
Though I do think there’s a subtle irony here too — the original commenter may simply be describing their own emotional reaction/disillusionment, while your response risks collapsing them into "part of the problem."
Feels like everybody in the thread is pointing at the same tendency from different angles.
Personally I am looking forward to the post-IPO world where a lot of very smart people with hard-won knowledge will have their golden handcuffs off.
The sum of these merits adds up to something. SpaceX is a political venture, and just like the uncomfortable questions that Microsoft/Google/Apple all pose, it's worth asking what the consequences will be in the long term. Lawful intercept sounded like a great plan, before it was leveraged by America's adversaries in Salt Typhoon as a prepackaged surveillance network.
"People" were always like that and will be so..stupid. Let me quote Agent K from MIB for you.
> A person is smart. People are dumb, panicky, dangerous animals and you know it...
The funny thing is that these are the same people who applauded obvious scams because Musk proposed it when they liked him...
Which, at this point, has already been beaten to death and is just... tiresome. While discussing the broad concept of space-based compute in general (outside of SpaceX, Elon, etc) can still actually be interesting.
Whatever good Musk has accomplished with SpaceX will be offset by the harm he has done to biomedical research in the final accounting.
NASA has been propaganda since Operation Paperclip, sadly. It's hard to politicize something that's always been political, even if Musk gives Peenemünde optics a run for it's money.
The problem is the recent shift away from science towards a more performative roadmap – getting to the Moon (again) is about showing off US might, not about science this time around, at least that's how it's being messaged. Many pure science endeavours have been canned. And the Artemis missions have a strong vibe of propaganda to them with slick marketing designed to emphasise America.
I guess to sum it up: doing good stuff and being seen to be good because of it, is fine, but making a show of doing good stuff explicitly for show, while behind the scenes doing as little as you can get away with, is not.
The first time around it was also about showing off US might. I don't think that something has changed much. Maybe wild Musk's lies are the only thing that was added.
What SpaceX has accomplished is just phenomenal.
Do you actually believe this data centers in space nonsense?
I'm a big space fan, don't get me wrong. But your exuberance uh, needs tempering.
More or less those things you mentioned have solutions and they are getting better.
https://www.health.harvard.edu/blog/measles-is-making-a-come...
Right now, the greatest threat to our survival and prosperity is humanity itself.
For their conventional space launch operations they also want multiple... to target different orbits, and to parallelize the high volume operations they anticipate.
There's already two Starship launch sites. The one in use in Texas, and one (LC-39A) in development at Kennedy Space Center, Florida. And there's good reason to believe they've begun planning a third in Louisiana. https://forum.nasaspaceflight.com/index.php?topic=64900.0
On a more serious note, the Cape Canaveral area / Kennedy Space Center has a large amount of empty land to build space infrastructure. The island has been dedicated to space facilities since the 1960s. Both SpaceX and Blue Origin have facilities there.