SpaceX IPO

[Psi]

Thread for discussing the SpaceX 2026 IPO

Posting in this thread automatically means you have accepted the following rules. Even if you never read them.

  Rule 1 - You agree that - This thread is about the SpaceX IPO. Do not turn this thread into another Musk thread. Do not post any rumors, accusations or facts about Musk or your view of him.
  Rule 2 - You agree that - Your post is considered to have broken rule 1 if 3 or more other forum members in this thread say it did. Your thoughts on if they're correct or not is irrelevant.
  Rule 3 - You agree that - If rule 2 is invoked on you (meaning you violated rule 1) then you have 2 options.  1) Never post in this thread again. Or 2) You must solder using wire-links on veroboard the text "Sorry, I was wrong" and then post the image for consideration. Only if your sorry image gets accepted by your peers are you then allowed to post again (except once for posting the image).  If veroboard is unavailable then other options maybe considered if in the same spirit and if using the same words.

[brucehoult]

Is there even any evidence for this? What I saw was posted by a parody account.

The company says they provide liquidity for employees and investors by having share "buy backs" twice a year.

[Psi]

Seems legit from what I could see.  (That they're planning to do it)
But I will retract thread if I'm wrong.

[Psi]

There's lots of downsides I can see to this idea of an IPO to raise money for AI datacenters in space.

- Shareholders want short term profits at the cost of innovation and long term progress or goals
- Shareholders have a tendency to ruin companies trying to get the share price up in the short term.
- There's a risk the IPO will get messed up (or people will throw that accusation around) and lawsuits will happen. *Cough* Firefly *Cough*.
- There's a risk new tech will appear to solve the issues limiting ground-based AI. If that happens it instantly makes the space based datacenters redundant after trillions have been spent.  (power & cooling are the only reason why space looks tempting).
- Putting AI/AGI in the sky over peoples heads may not go down so well with the room temp IQ public.
- Datacenters in space have HUGE solar arrays. Think how annoyed astronomers got with tiny Starlinks. Now think about km long solar arrays on AI datacenters.
- The AI bubble bursts

I guess any money spent making datacenters in space can pivot and turn the datacenter into a large space stations. But you still lose a lot of money in the GPUs

[tom66]

I'd buy some stock.  Unlike Tesla, I believe SpaceX (SPCX?) has a longer term future, and I've divested myself from direct investment in TSLA as of about 12 months ago.  With Tesla, I question how they will retain their market lead in the face of competition from China, Korea and Germany, at least outside of the US.

[BitBangBro]

SpaceX IPO is tied to AI bubble - now SpaceX wants part in it before it bursts.

Besides that fundamental problem - which share owner will want to fund the Mission To Mars on civilizational level, there are others:

• AI muscle ages FAST and is replaced rapidly. What is the turnover rate for datacenters in FRIGGIN SPACE ?!?
• Thes things are going to be MASSIVE. And they are obviously shooting for having MANY of them. How they intend to prevent The Kessler Syndrome ? https://en.wikipedia.org/wiki/Kessler_syndrome
• Datacenters in space will obviously be juicy targets. It's a practically a public secret that a few tons of gravel and sand, simply dispersed in the roughly right orbit would wreak havoc even in LEO, where things fall down over few years. In higher orbits this cloud could persist practically forever. How do they intend to protect them ?

If they seriously invest in this and any of listed SNAFU happens, they are toast.
And that's even ignoring AI bubble lifetime.
How many investors would be willing to follow Elon on such a long shot ?
Look what happened with his "Boring" or "Hyperloop" hype.
Or many others.

[Psi]

Yep, It just feels like a very risky distraction to me.

Lets hope the AI bubble busts before they get too far into wasting money on things.
Hopefully the IPO never goes ahead but I fear there is lots of pressure to do it.

[BitBangBro]

Yep, It just feels like a very risky distraction to me.

Here is Sabine Hossenfelder's take on it:
https://youtu.be/t8x09q1MjcM

[tom66]

Space datacentres are a properly stupid idea.  If the concern is over energy, put them in the Sahara Desert and run them off solar and batteries.  Cooling a datacenter in space will require radiators twice the size of the solar panels.  Maintenance is next to impossible without employing actual astronauts.  The whole thing screams of SV losing its shit over terrestrial prices and not actually thinking rationally here... maybe we don't need all this AI in our lives?  I like tools like Gemini and ChatGPT, but I only use them as an enhanced Google; I don't regard them as life changing and could easily live without them.

[Psi]

One "Carrington event" type solar storm and you might have the worlds most expensive constellation of paperweights instead of an AI datacenters.

[Psi]

Space datacentres are a properly stupid idea.

I don't think the idea is stupid, I think it's a "running before you can walk" type situation.

If we already had a big space/orbital construction industry it would probably be a reasonable idea.  Reasonably easy to take a spacecraft to go service it, infrastructure already up there for construction etc..

But trying to run before you can walk is a very risky idea.

[BitBangBro]

IMO if this thing - SpaceX IPO gets off the ground, it will be bad sign for everyone.
People have been suspicious about SpaceX ties to MIC from the begininning.

Not only that IPO could burry Mission To Mars thing and take SpaceSHip over for pure military projects.
Only legit use for massive datacenters in orbit that comes to mind is for global mass surveillance, fed by a fleet of sensors and cameras, not unlike current Starlink project. 🙄

Someone has said somewhere that first city on Mars could well be a military base...

[woofy]

Thread for discussing the SpaceX 2026 IPO

Posting in this thread automatically means you have accepted the following rules. Even if you never read them.

  Rule 1 - You agree that - This thread is about the SpaceX IPO. Do not turn this thread into another Musk thread. Do not post any rumors, accusations or facts about Musk or your view of him.
  Rule 2 - You agree that - Your post is considered to have broken rule 1 if 3 or more other forum members in this thread say it did. Your thoughts on if they're correct or not is irrelevant.
  Rule 3 - You agree that - If rule 2 is invoked on you (meaning you violated rule 1) then you have 2 options.  1) Never post in this thread again. Or 2) You must solder using wire-links on veroboard the text "Sorry, I was wrong" and then post the image for consideration. Only if your sorry image gets accepted by your peers are you then allowed to post again (except once for posting the image).  If veroboard is unavailable then other options maybe considered if in the same spirit and if using the same words.

Cambridge University Press says:
A rumor is an unofficial story or piece of information that is spread from person to person, often without verification of its truth. It can be about a person, event, or situation and may be true or invented.

So, with my screwball sense of humor, can we have 3 votes for the veroboard, please.
 

(I'll bet the rumor is true, though).

[magic]

By posting here I automatically declare that I don't give a damn about your rules and Elon sucks.

Space datacentres are a properly stupid idea.

I don't think the idea is stupid, I think it's a "running before you can walk" type situation.

It's not stupid if the goal is to rise money from idiots.

But it's stupid to believe that it will actually work or to invest in it.

[SiliconWizard]

Beyond not being realistic for now, people should seriously consider the relevance of the idea. It's bad enough to have your data hosted on remote servers you have no power over, but are you comfortable with your data being stored in space? But yeah, technically speaking, it's very, very far off.

If it's mainly for "AI", as some have suggested, then this is absolutely nuts. This thing has got to stop. It's becoming a literal cancer that's eating everything alive. If nothing changes short term, we're going to face a major semiconductor shortage within probably just a few months from now, much worse than what we had during covid. Prices of RAM and storage are already skyrocketing, and it's just the beginning.

Soon enough, the question may not be whether "AI" is going to take your job, but whether you'll be able to do your job at all - especially if you work in the electronics industry, like many of us do here. And whether the average Joe will be able to afford anything but very, very dumb technology barely powerful enough to access remote servers with constant internet connection. This is all getting mind-f*ckingly stupid. I don't think comparing that to cancer is fundamentally exaggerated.

[SeanB]

SpaceX already is putting up lots of Starlink units, so putting a small amount of AI in each new unit, increasing the power use by a little, and then using the built in comms between units to spread the compute load out over a cluster, might work well, and as they only have an orbital life of under 10 years, you just have to hope that early life failures do not land on you, but all go and put themselves into a few graveyard sites in the pacific, because you will need a second or third space dump site as not all orbits can make it to the existing one. Will make the cloud literally above you, and they likely will havelarge gateway complexes which spread out the uplink data and decrease latency, so that the inter unit links do have enough bandwidth to handle it. Just means you really have a hard version set, and it will have to be backwards compatible with upgraded hardware, because for at least 20 years you will have "legacy" units around doing work, and unless you plan to scrap units while they are young, very hard yo upgrade.

[SiliconWizard]

Anyway, the IPO seems like legit news: https://finance.yahoo.com/news/spacex-pursue-2026-ipo-raising-203725148.html

[Stray Electron]

SpaceX IPO is tied to AI bubble - now SpaceX wants part in it before it bursts.

Besides that fundamental problem - which share owner will want to fund the Mission To Mars on civilizational level, there are others:

• AI muscle ages FAST and is replaced rapidly. What is the turnover rate for datacenters in FRIGGIN SPACE ?!?
• Thes things are going to be MASSIVE. And they are obviously shooting for having MANY of them. How they intend to prevent The Kessler Syndrome ? https://en.wikipedia.org/wiki/Kessler_syndrome
• Datacenters in space will obviously be juicy targets. It's a practically a public secret that a few tons of gravel and sand, simply dispersed in the roughly right orbit would wreak havoc even in LEO, where things fall down over few years. In higher orbits this cloud could persist practically forever. How do they intend to protect them ?

If they seriously invest in this and any of listed SNAFU happens, they are toast.
And that's even ignoring AI bubble lifetime.
How many investors would be willing to follow Elon on such a long shot ?
Look what happened with his "Boring" or "Hyperloop" hype.
Or many others.

      because a simple thanks isn't enough.  You read my mind!

[Stray Electron]

By posting here I automatically declare that I don't give a damn about your rules and Elon sucks.

Space datacentres are a properly stupid idea.

I don't think the idea is stupid, I think it's a "running before you can walk" type situation.

It's not stupid if the goal is to rise money from idiots.

But it's stupid to believe that it will actually work or to invest in it.

    I agree with all of this. I think that anyone that supports this idea is completely naive about the practical considerations involved.  I've been down this road before; for a time I was an engineer on the Brilliant Pebbles program (look it up) and was in charge of trying to deal with the logistics of making it happen.  It was fascinating work but I knew that it would never be a reality so I wasn't upset when they moved me to a new job   You can make projects like this work on a small scale (we absolutely did! :-) but not at the scale needed to make it effective.  And anything less than 100% effective is a failure.

[tom66]

Let's not forget the added risk of cosmic rays damaging non radhard GPUs (can you imagine trying to radhard qualify a GPU like the Nvidia A100?) and the issues with memory/flash reliability.  You'd have to ECC everything beyond the normal level that a terrestrial datacentre runs at.  Component lifespans will also be far worse if they're in orbit, just because the environment is that much more hostile, so you'll have yet more e-waste and need to carry hot spares on board the spacecraft because presumably swapping a server over isn't something that is trivially done any more.

There is no world in which AI datacentres in space make sense.

[mawyatt]

With the acres of silicon AI processors flying around in space unprotected, seems they will be seeing constant radiation induced SEUs occurring, often massive bursts of SEUs, and an occasional complete blasting of SEUs (solar flairs). Sure redundancy and/or error correction can be employed but with the potential massive SEUs bursts occurring often, this correction must take it's toll on overall performance and thru-put.

Not to mention by the time the arrays are deployed and fully functional the acres of AI chips will be somewhat outdated, and over the expected lifespan completely obsolete.

Edit: Tom's post above ^^

Think of the MW/MMW/Laser Communication Data Pipes to and from the orbiting AI centers, how's that going to be achieved with the required data flow to make these orbiting AI centers competitive with earthbound AI centers!!

An analysis of the cost to develop/deploy/repair/risk an orbiting AI Center vs the earthbound equivalent, and the power per search cost, considering over the lifespan of the orbiting AI center the earthbound can be upgraded significantly reducing power cost per search, should make the orbiting AI center impractical and not financially viable.

Most of the data routing stuff in space is what we used to call "Bent Pipes", not much data computational and storage going on up their, certainly not in the AI context!!

Anyway, this is an interesting topic for discussion even if it never materializes!!

Best     

[magic]

I think simply logistics of lifting tonnes of computers to the orbit and heatsinking them (sorry, you can't cool with air or water over there, only radiation) and the cost of lifting tonnes of heatsinks will effectively kill any such project.

[SiliconWizard]

SpaceX already is putting up lots of Starlink units, so putting a small amount of AI in each new unit, increasing the power use by a little, and then using the built in comms between units to spread the compute load out over a cluster, might work well, and as they only have an orbital life of under 10 years, you just have to hope that early life failures do not land on you, but all go and put themselves into a few graveyard sites in the pacific, because you will need a second or third space dump site as not all orbits can make it to the existing one. Will make the cloud literally above you, and they likely will havelarge gateway complexes which spread out the uplink data and decrease latency, so that the inter unit links do have enough bandwidth to handle it. Just means you really have a hard version set, and it will have to be backwards compatible with upgraded hardware, because for at least 20 years you will have "legacy" units around doing work, and unless you plan to scrap units while they are young, very hard yo upgrade.

Increasing the power use "by a little"? Do you have any idea how much additional power it would draw to be the beginning of economically worthwhile?
Do you know what a typical Starlink satellite draws on average and how much it's able to generate with its solar panels?

[NE666]

There is no world in which AI datacentres in space make sense.

And 'they' must know that. So the question should be, what's the real game here?

The cynic in me says, SpaceX has realised that it needs more external investment in order to continue development of Starship and any future Mars missions, and an IPO has been chosen as the vehicle. So why not find a tenuous link between the company and the current AI frenzy, to pump the issue price as high as possible by association, before that particular bag of tulip bulbs hits its shelf life.

Later down the road, you can quietly 'rest' the idea, citing feasibility studies.

[mawyatt]

I've been down this road before; for a time I was an engineer on the Brilliant Pebbles program (look it up) and was in charge of trying to deal with the logistics of making it happen.

Think the Homing Overlay Experiment (HOE) was the predecessor to Brilliant Pebbles. Very interesting stories about this development. Maybe should start another thread on this not to violate the rules here

Best