On June 12, the largest initial public offering (IPO) in history is set to hit the tape, and almost nobody is asking whether the price is right, because almost everybody already wants in.
The Elon Musk-founded SpaceX (SPCX) is going public at a targeted $1.75 trillion market cap, raising $75 billion in equity at $135 a share, and the reception will be a stampede regardless of what the filing says. That is exactly the problem. When demand runs this hot, the questions that normally get asked at the IPO table simply do not get asked, and the May 20 prospectus is full of reasons they should be.
Priced for an age of abundance
Before the skepticism, give the bull case its due: It is being made in the trillions, and the market is nodding along. Lead underwriter Goldman Sachs has told clients SpaceX’s AI revenue could grow 100-fold by 2030, and bankers are already floating a stretch to $2 trillion on the stock’s debut. The S-1 filing itself argues less in multiples and more in destiny, promising an age of abundance, the harnessing of the Sun’s full output, and a constellation that could one day run to a million satellites.
This is what matters for June 12. SpaceX does not need a discounted cash flow to clear. It needs the same reflex that drove GameStop (GME) and Tesla (TSLA), a brand so familiar it has already become a meme, and Musk’s retail base that buys the story long before it reads the statements. Management is not coy about it: Chief Financial Officer Bret Johnsen has called retail “a bigger part than any IPO in history,” casting the unprecedented allocation as a thank you to the people who have backed Musk for years. On that score the deal is done before it opens.
The number came before the price discovery
Most mega-cap deals spend weeks grinding an indicative range down to a price as bookrunners feel out demand. SpaceX skipped all of it. The original S-1 went out with the price fields blank, and the company then handed the market a flat $135 a week ahead of pricing rather than let a bookbuilder find the level. Cooley’s Richard Segal read the message bluntly: you are either in or you are out. One syndicate banker told PitchBook that real price discovery could have pushed the stock well above $135, which tells you the fixed number is a floor dressed up as fair value.
The appetite is not in doubt. The recent Cerebras float ran roughly 20 times oversubscribed; SpaceX has reserved an unprecedented 30% of the deal for retail, three times the usual sliver; the Japanese tranche was upsized 25% on demand before the roadshow had even wrapped. Retail will buy at whatever the open prints, sight unseen. None of that is price discovery. It is a queue.
A broadband company in a rocket’s coat
Strip the romance out and read the segment table. The only part of SpaceX that reliably makes money is Starlink: $11.4 billion of revenue and $7.2 billion of adjusted EBITDA in 2025. The Space segment (the rockets, the part everyone is buying the story for) lost money from operations last year once you count the $3 billion sunk into Starship, SpaceX’s designated ride to Mars and the most expensive fireworks program in Texas, engineered for full reusability but mostly achieving full disassembly instead.
Consolidated, the company lost $4.9 billion in 2025 and another $4.3 billion in the first quarter of 2026 alone. And here is the wrinkle the bulls skate past. Starlink’s average revenue per user is sliding, from $99 a month in 2023 to $91, then $81, and down to $66 in the first quarter of this year. Subscribers are climbing toward 10.3 million, but the unit economics of the one business that works are compressing as it scales. The floor under the valuation is softer than the headline suggests.
The premium rides on hardware nobody has flown
So why $1.75 trillion, double the $800 billion the company fetched in a tender barely six months ago? Because the number is not priced on rockets or broadband. It is priced on AI. xAI, folded into SpaceX in February, lost $6.4 billion from operations in 2025. The thesis that carries the premium is orbital AI compute, and on this the filing is disarmingly honest: neither SpaceX nor anyone else has ever operated, or even attempted, it. Either way, deployment is not expected before 2028. The company pegs its addressable market at $28.5 trillion, of which $26.5 trillion is AI, a figure that conveniently leaves out China and Russia.
The revenue meant to validate all this is thinner than it looks. The marquee third-party compute contract, $1.25 billion a month, runs to a single counterparty, the AI lab Anthropic, that can walk on 90 days’ notice. The Cursor tie-up is a $60 billion all-stock option, not a closed sale. Terafab, the chip venture with Tesla (TSLA) and Intel (INTC), is by the filing’s own admission a framework that binds no one. Morningstar’s Nicolas Owens calls the company significantly overvalued at a fair value of $780 billion; the Danish pension fund AkademikerPension has branded the valuation pure fantasy and excluded it outright.
The furnace is fed by the thing that works
Follow the cash and it gets starker. AI swallowed $12.7 billion of capital spending in 2025, and roughly three-quarters of the total in the first quarter, more than the rest of the company combined. Net cash used in investing was $16.7 billion across those three months, and the cash pile fell from $24.7 billion at year-end to $15.9 billion by the end of March. And those figures come up against $29.1 billion of debt.
In plain terms, the profitable broadband business is being used to feed an AI build-out that does not yet earn its keep, and you are being asked to pay the premium for the part doing the burning, not the part doing the earning. That can work if the orbital bet lands. It is simply not what most buyers think they are buying.
The exit liquidity is your 401(k)
The most underappreciated source of demand is not conviction at all. It is plumbing, and the plumbing runs straight through your retirement account. S&P Dow Jones held the line this week, retaining its profitability rules and shutting SpaceX out of the S&P 500 for now.
But Nasdaq bent its rules to wave the stock into the NASDAQ-100 about 15 trading days after listing, with the Russell indices close behind, forcing every fund that tracks them to buy. Those funds are not making a call on value. They are the index trackers inside millions of 401(k)s and pension plans, more than $30 trillion of largely passive money that must buy SpaceX at over 90 times sales because the benchmark says so, not because anyone judged it cheap. Estimates of the forced buying run from $22 billion to as high as $60 billion.
Now line that up against who gets to sell. Most IPOs lock insiders up for 180 days. SpaceX wrote a staggered schedule instead, letting insiders start cashing out as early as the second trading day after its first quarterly earnings, likely in August, while the friends-and-family allocation carries no lockup at all and can sell on day one.
The early unlock does double duty: it widens the float, which lifts SpaceX’s index weighting, which forces the passive funds to buy even more. The calendar is built so that insiders can step out at roughly the moment your retirement fund is mechanically obliged to step in.
The biggest bagholder exercise of all time
Bloomberg’s Eric Balchunas has noted how neatly the unlock and passive-buying schedules line up, which is the point at which the word passive stops meaning much. And the sellers answer to no one: the dual-class vote leaves Musk in control, and the filing flags live Federal Trade Commission (FTC) and European inquiries into the X and Grok side.
The Financial Times’s Robin Wigglesworth called the broader pattern the biggest bagholder exercise of all time; the Wall Street Journal’s Jason Zweig called the fast-entry rule arbitrary, unfair and potentially risky. The Tesla template from 2020 is the tell: front-running money drives the inclusion run-up, then exhausts, and the stock drifts once the forced bid is filled, leaving the index holder long at the top. You are buying a ticket on the rocket, not a seat in the cockpit.
None of this will dent June 12. The book is covered many times over, retail is lined up around the block, and the index machinery will do its forced buying on schedule. The demand is the one thing here that is not in question.
What is in question is what that demand is pricing. A company that loses billions and funds its boldest bet by burning the cash from its only profitable line? Or just the near-certainty that whoever buys at the open can sell to someone further back in the queue who wants in even more? Those are not the same trade, and only one survives the moment the queue stops growing.
About The Author
