SpaceX potential IPO

[RedLineTrain]

Note that Google is the only shareholder that alone could bring a shareholder suit, given that it would be the only entity with more than 3% of any SpaceX combination, with or without Tesla.

Unless someone buys more than that during the IPO, right?

Still ...3% ownership is required to bring a shareholder lawsuit? Ok, that probably pretty much eliminates my worry of an IPO opening up SpaceX to a bad faith (either politically motivated or competitor driven) shareholder lawsuit ... If valuation is something like $1T, 30 billion dollars is a lot to spend for that purpose!

A) I don't think that's true (the 3% rule), but also B) how can you worry so much about shareholder interference and not check about the basic rules governing that?  Not just percent ownership but also the types of possible claims, what needs to be argued, etc?

A) The 3% shareholder suit rule is true in Texas, where SpaceX was reincorporated recently.  Tesla included the 3% threshold in its bylaws after a similar reincorporation, so we can expect that SpaceX also did so.

Notice that in purchasing Echostar's spectrum, SpaceX was careful not to put Echostar at more than 3% of ownership.  It paid part in cash and part in equity, such that ownership was just below that threshold.

Of course, a group of shareholders at more than 3% ownership could join together to bring a lawsuit.  So it doesn't guard against any shareholder derivative suit.  Just the more frivolous ones, like those that are happening in Delaware.

Derivative Proceedings – Changes to Subchapter L
Beneficial Ownership Requirement: A new § (3) to § 21.552(a) would require that a shareholder desiring to bring a derivative action involving a corporation listed on a national securities exchange or a corporation that has made an affirmative election to be governed by the BJR statute and has 500 or more shareholders must at the time of bringing the proceeding beneficially own a number of shares sufficient to meet the required ownership threshold identified in the corporation’s certificate of formation or bylaws, provided that the required ownership threshold does not exceed 3% of the outstanding shares of the corporation.

https://corpgov.law.harvard.edu/2025/05/21/texas-corporate-law-changes-challenge-delawares-dominance/

[thespacecow]

https://www.bloomberg.com/news/articles/2026-02-02/elon-musk-s-spacex-said-to-combine-with-xai-ahead-of-mega-ipo

Elon Musk is combining SpaceX and xAI in a deal that values the enlarged entity at $1.25 trillion, as the world’s richest man looks to fuel his increasingly costly ambitions in artificial intelligence and space exploration.

...

The deal gives SpaceX a valuation of $1 trillion, and xAI a value of $250 billion, people familiar with the matter said. The combined company’s valuation was announced to employees in a memo on Monday, some of the people said earlier.

[XRZ.YZ]

The reason X.AI suddenly merge with SpX pre IPO is because X.AI needs lots of capital really badly.
And SpX has the cash flow and balance sheet so in a better position to raise capital plus internal funding with Starlink revenue.


It could also mean Mars dream will be at least a few years late as AI will suck out all resource for some time.

[meekGee]

(...moved)

[catdlr]

Moderation:

The Data Center Thread is where Aix discussion should go. Please don't side-track this thread.

Tony

[Tywin]

Some opinion of the high valuation:

https://twitter.com/peter_adderton/status/2018859958654374365

[Vultur]

Some opinion of the high valuation:

https://twitter.com/peter_adderton/status/2018859958654374365

It's future leaning, the valuation only makes sense with the expectation that extremely rapid growth will occur.

But there are a lot of tech stocks with ... Questionable valuations, I would say. It's not just them.

Ultimately if/when they IPO the market will decide.

[RedLineTrain]

50% EBITDA margin is considered extremely high.  As noted above, with operational leverage as the business grows, that percentage for launch/Starlink is at this moment higher and headed higher still.

Normally, private companies adjust their EBITDA.  So we will need to read the draft S-1 registration statement carefully to see what is included and excluded.

[meekGee]

Some opinion of the high valuation:

https://twitter.com/peter_adderton/status/2018859958654374365

Nvidia, for example, went up 16,000% over 5 years, and it's at the bottom of he food chain.

For this valuation to make sense, you need to believe that AI will become a part of industry and economics, not just a demo.  If you do, it's a perfectly good valuation.

Consider what happened to Intel, Microsoft, or Google since their very early days.

Your attempt to compare present-day revenue with valuation is not even credible - nobody ever does that in a situation like this. It's just irrelevant.

[RedLineTrain]

For this valuation to make sense, you need to believe that AI will become a part of industry and economics, not just a demo.  If you do, it's a perfectly good valuation.

I believe that this valuation is justified on the Starlink business alone.

[meekGee]

For this valuation to make sense, you need to believe that AI will become a part of industry and economics, not just a demo.  If you do, it's a perfectly good valuation.

I believe that this valuation is justified on the Starlink business alone.

You're right. It's almost complimentary.

AI COULD be much bigger, but is speculative. Starlink meanwhile is a demonstrated business, highly profitable, and will grow some 10x with Starship.

Despite Tywin's assertions, there's no meaningful competition on the horizon.

But it's also a known quantity. xAI is a sort of "the sky's the limit".

It's nice to have both.

[Tywin]

PER 300, very dangerous...

https://old.reddit.com/r/wallstreetbets/comments/1qwbj18/the_spacex_ipo_is_going_to_tank_the_market/

[Coastal Ron]

Some opinion of the high valuation:

{Tweet snipped}

Nvidia, for example, went up 16,000% over 5 years, and it's at the bottom of he food chain.

For this valuation to make sense, you need to believe that AI will become a part of industry and economics, not just a demo.  If you do, it's a perfectly good valuation.

However Nvidia is the preferred silicon that AI engines run on, which is why it was perfectly positioned to increase in value.

In contrast, xAI offers AI solutions in a very competitive market, and so far has not shown it has any advantages with regards to actually generating revenue - only ~$200M in 2025 vs $13B for OpenAI and $7B for Anthropic.

Consider what happened to Intel, Microsoft, or Google since their very early days.

I think you have survivor bias going on here, since you keep assuming that xAI has advantages and revenue that it really does not. For all we know, xAI may end up being the BeOS of the AI market wars. 

As I've said before, if the big goal for SpaceX is to build out data centers in space, it didn't need to buy xAI in order to do that. And even if data centers in space end up being a thing, it is more likely that SpaceX will rent/lease them to xAI competitors such as OpenAI or Anthropic who need to scale faster than what Google and other companies can do with their huge lead in data centers.

[russianhalo117]

Some opinion of the high valuation:

{Tweet snipped}

Nvidia, for example, went up 16,000% over 5 years, and it's at the bottom of he food chain.

For this valuation to make sense, you need to believe that AI will become a part of industry and economics, not just a demo.  If you do, it's a perfectly good valuation.

However Nvidia is the preferred silicon that AI engines run on, which is why it was perfectly positioned to increase in value.

In contrast, xAI offers AI solutions in a very competitive market, and so far has not shown it has any advantages with regards to actually generating revenue - only ~$200M in 2025 vs $13B for OpenAI and $7B for Anthropic.

Consider what happened to Intel, Microsoft, or Google since their very early days.

I think you have survivor bias going on here, since you keep assuming that xAI has advantages and revenue that it really does not. For all we know, xAI may end up being the BeOS of the AI market wars. 

As I've said before, if the big goal for SpaceX is to build out data centers in space, it didn't need to buy xAI in order to do that. And even if data centers in space end up being a thing, it is more likely that SpaceX will rent/lease them to xAI competitors such as OpenAI or Anthropic who need to scale faster than what Google and other companies can do with their huge lead in data centers.

Nvidia is in an M&A posture because it has hit a technological brick wall which it's currently band aiding by chucking more and more power at the cost of thermal waste whilst achieveing short term efficiency gains which ARM created a solution to. The SpaceX parent's business segments would benefit buying out Nvidia and other related companies to vertically integrate AI processing unit (APU) production. APU types consist of field programmable gate arrays (FPGAs), tensor processing units (TPUs), vision processing units (VPUs), quantum processing units (QPUs), neural processing units (NPUs), intelligence processing units (IPUs), and application-specific integrated circuits (ASICs) to name the most common examples.

[RedLineTrain]

In contrast, xAI offers AI solutions in a very competitive market, and so far has not shown it has any advantages with regards to actually generating revenue - only ~$200M in 2025 vs $13B for OpenAI and $7B for Anthropic.

xAI hasn't really tried to generate revenue, so you can't hold that against the company.

The problem with generating revenue is that you have to service that revenue with inference compute.  So far, xAI has been focused almost exclusively on training compute.  It outsources all of its inference compute to the hyperscalers like Microsoft.

[Vultur]

For this valuation to make sense, you need to believe that AI will become a part of industry and economics, not just a demo.  If you do, it's a perfectly good valuation.

I believe that this valuation is justified on the Starlink business alone.

Yeah. The potential of Starlink on Starship could be enormous. How many people in the world have devices but not good fiber connectivity? I could easily see a mature Starlink being 100x its current size, 1B people or so.

I remain concerned by the AI focus, but SpaceX has huge advantages which wasting money on AI probably won't destroy. Especially since the AI stuff requires Starship, and once Starship works their launch dominance goes from extreme to ridiculous.

[phantomdj]

Cross-posting. I think this is relevant here.

Teslarati: SpaceX’s xAI merger keeps legal liability and debt at arm’s length [Feb 6]

Rather than fully combining the two companies, SpaceX retained xAI as a wholly owned subsidiary. The structure, commonly referred to as a triangular merger, allows xAI’s liabilities, contracts, and outstanding debt to remain isolated from SpaceX’s balance sheet.

As a result, SpaceX is not required to repay xAI’s existing debt, which includes at least $12 billion inherited from X and several billion dollars more raised since then. The structure also prevents the transaction from triggering a change-of-control clause that could have forced immediate repayment to bondholders.

This could be a very smart move and reminds me of the company I used to work for, United Technologies (UTX) before its merger with Raytheon. UTX was a conglomerate with wholly owned subsidiaries such as Pratt & Whitney, Otis Elevator and Carrier air conditioning just to name a few. xAI could just be a subsidiary of SpaceX. Benefits SpaceX but reduces the exposure and risk.

[Vultur]

Cross-posting. I think this is relevant here.

Teslarati: SpaceX’s xAI merger keeps legal liability and debt at arm’s length [Feb 6]

Rather than fully combining the two companies, SpaceX retained xAI as a wholly owned subsidiary. The structure, commonly referred to as a triangular merger, allows xAI’s liabilities, contracts, and outstanding debt to remain isolated from SpaceX’s balance sheet.

As a result, SpaceX is not required to repay xAI’s existing debt, which includes at least $12 billion inherited from X and several billion dollars more raised since then. The structure also prevents the transaction from triggering a change-of-control clause that could have forced immediate repayment to bondholders.

This could be a very smart move and reminds me of the company I used to work for, United Technologies (UTX) before its merger with Raytheon. UTX was a conglomerate with wholly owned subsidiaries such as Pratt & Whitney, Otis Elevator and Carrier air conditioning just to name a few. xAI could just be a subsidiary of SpaceX. Benefits SpaceX but reduces the exposure and risk.

Isolation from xAI's debt does significantly reduce my worry level about this.

[RedLineTrain]

I think we shouldn't believe anything reported about the nitty-gritty details of xAI's debt.  xAI most recently used an SPV for Nvidia chips that was perhaps at least partially owned by Nvidia, financed with loans from the private market (Apollo and the like), and with a 5-year lease of the chips to xAI.  Similar arrangements were made for the natural gas turbines.

Point is that any number of details on these transactions could impact the debt liabilities.  We may learn more in the draft S-1 registration statement.

[thespacecow]

From Payload Pro newsletter, not specific to SpaceX, but obviously SpaceX is the ultimate hardware investment opportunity:

The Space Investing Thesis During The Software Rout

The investor playbook has flipped: In the era of AI, the asset-light, high-margin good businesses start to look bad, while greasy, capital-intensive bad businesses now look good. That’s just as true in space as anywhere else.

Rethinking software’s moats: AI’s breakneck development pace has accelerated the enterprise SaaS sell-off this year. Software ETFs have plunged 20% in one month. The sector’s price-to-sales ratio down from 9x in September to 6x today. That’s a decade-low level.

Gold standard? Software has long been considered the gold standard business model—sticky, asset-light enterprise-focused, high switching costs, recurring revenue, and high margin. But now the investment philosophy is being questioned. When ChatGPT, Gemini, Claude, xAI (aka SpaceX) ship sector-upending applications seemingly every week, software’s once-impenetrable moats no longer hold.

Margin for the taking: Decades of maintaining strong moats helped form the backbone of software’s insanely good 80% gross margins. Years ago, channeling his inner Buffett, Jeff Bezos outlined Amazon’s strategy, saying: “Your margin is my opportunity.” But for the past 20 years, that undercutting playbook has generally not worked for enterprise SaaS. Why?

* High switching costs: Switching SaaS providers is a nightmare scenario. Data lock creates a six-month transfer process. And employees are always reluctant to learn a new trick.
* Price-inelastic customers: Plus, enterprise customers are generally loyal and willing to pay higher prices.
* Genuinely important service: Finally, SaaS businesses solve important problems for companies, increasing productivity and limiting headaches

As a result, SaaS pricing power and competitive edge have been god-tier. Investors flocked. And hardware was forgotten. Sad.

AI to hardware: But now, AI is coming for that margin, recurring revenue, and growth in a big way. Where do you find business moats when the AI raging bull is coming for every computer task, and every white-color job? The answer is hardware.