The SpaceX IPO's Real Story Isn't Musk's Net Worth — It's the Balance Sheet Nobody's Pricing In
The headlines wrote themselves on June 12. SpaceX went public at $135 a share, raised roughly $75 billion in the largest IPO in stock market history, and instantly made Elon Musk one of the only people on Earth with a real shot at a trillion-dollar personal fortune. The stock popped about 19% on its debut day. Cable news ran the rocket launch footage. Financial Twitter — or X, since Musk owns that too now — treated it as a coronation.
None of that is the interesting part. The interesting part is buried about 40 pages into a prospectus most people will never open, in a section titled "Capital Allocation and Funding Strategy," where SpaceX disclosed that $20 billion of the $75 billion it just raised was already spoken for before the stock started trading — and not for rockets.
That's the story. Not whether Musk becomes a trillionaire. Whether the company he's taking public can fund the businesses it's promising to build, on the timeline it's promising to build them, without going back to the same retail investors who just bought in at the top for more money — or more dilution.
A company built on three very different engines
SpaceX's S-1 filing, published May 20 and amended June 1, broke the company into three reporting segments for the first time: Space (the rocket and launch business), Connectivity (Starlink), and AI (the recently folded-in xAI operation, including X and Grok). The company reported $18 billion in consolidated revenue for 2025, against a net loss of $4.9 billion, with EBITDA of roughly $6.6 billion for the year.
That gap between a healthy EBITDA number and a multibillion-dollar GAAP loss is the first thing worth sitting with, because it tells you these are three businesses at wildly different stages of maturity wearing one ticker.
Starlink is the one that actually makes money. It generated $11.4 billion in revenue in 2025, up 48% from the prior year, and threw off roughly $4.4 billion in operating profit at margins north of 60%. Subscriber count more than doubled in 2025 and crossed 10.3 million by the end of March 2026, with SpaceX operating across 164 countries and controlling roughly three-quarters of all active maneuverable satellites in low-Earth orbit. By any normal standard, this is the rare satellite-internet business that has actually scaled.
But Starlink's growth is decelerating on a per-user basis, and the prospectus is candid about why. Average revenue per subscriber fell from $99 a month in 2023 to $66 by the first quarter of 2026 — a one-third decline in three years, driven by international expansion into markets where the original price point was too high, plus cheaper hardware tiers like Starlink Mini pulling in lower-revenue customers. Growth is real. Growth at the old margins is not.
Then there's Space — the actual rocket business, the one most people assume is SpaceX. It generated $4.086 billion in revenue in 2025 but posted a loss from operations of $657 million for the year, and lost $662 million in just the first quarter of 2026 alone. Much of that is deliberate: SpaceX has spent more than $15 billion developing Starship, running roughly $3 billion a year in Starship-specific R&D with no Starship revenue yet offsetting it. That's a bet, not a failure — Starship is the vehicle SpaceX needs for next-generation Starlink satellites, lunar contracts, and any future Mars ambitions. But it's a bet funded entirely by Starlink's profits, and now it has company.
The AI segment is the one actually driving the loss
Without the xAI acquisition, completed in February 2026, SpaceX would have posted a $791 million net profit in 2024. With it folded in, the company posted a $4.94 billion net loss for full-year 2025 and a $4.28 billion loss in the first quarter of 2026 alone. SpaceX's accumulated deficit now sits at $41.3 billion.
The AI segment — now branded "SpaceXAI" — posted $818 million in revenue against a $2.469 billion loss from operations in the first quarter of 2026 alone, and SpaceX disclosed that AI operations posted losses exceeding $6 billion across 2025, burning another $2.5 billion in the first quarter of 2026. Morningstar's read on the filing put it plainly: Starlink's success is effectively subsidizing xAI's extensive expenditures.
This is the part that should give pause to anyone buying the SpaceX story as a space company with an AI side project. It's closer to the reverse. The rocket business is a capital-intensive bet funded by satellite internet, and the satellite internet business is now also funding an AI lab that's burning more than $6 billion a year competing against OpenAI, Anthropic, and Google in a market where none of the leaders are profitable yet either. SpaceX didn't diversify away from capital intensity by adding AI. It added a second, larger one on top of the first.
The $20 billion nobody priced into the valuation
Here's where the prospectus stops being a growth story and starts being a financing story. In March 2026 — one month after the xAI merger closed — SpaceX took on a $20 billion bridge loan from a syndicate of banks including Goldman Sachs, Bank of America, Citigroup, JPMorgan Chase, and Morgan Stanley, and that loan matures as early as September 2, 2027. The money wasn't used to build anything: it retired $17.5 billion in older, higher-interest debt that X and xAI had accumulated separately — including junk-rated notes carrying interest as high as 12.5% — plus $1.163 billion in prepayment penalties for closing those facilities early. That's 27% of the entire $75 billion IPO raise earmarked to repay debt with no connection to rockets, satellites, or anything resembling SpaceX's core business — it was leverage absorbed onto SpaceX's own balance sheet to clean up the books ahead of the listing.
So of the headline-grabbing $75 billion raise, more than a quarter walks straight back out the door to retire debt before a single dollar reaches Starship, Starlink's next-generation satellites, or new AI infrastructure. SpaceX's own capital allocation language in the prospectus states plainly that it plans to access "a range of debt and equity financing solutions" as a public company to fund future growth and maintain liquidity — corporate language for: there will be more raises, and they likely won't all be debt.
That matters because of who actually bought into this IPO. Retail investors were allocated as much as $22.5 billion of the $75 billion raised — not a side pool, but close to a third of the entire offering. Unlike the venture funds and institutional holders who got in at $350 billion and $1 trillion valuations years ago, retail investors bought in at the top, at a fixed $135 price with no traditional range process, on a company now valued near $1.75 trillion. At that valuation, SpaceX trades at roughly 94 times 2025 revenue — a multiple that, per BitMEX's analysis of the filing, "assumes flawless execution," and one that contracts violently if Starlink's per-user revenue keeps compressing or AI losses keep accelerating.
A lockup schedule built for speed, not stability
The mechanism by which all of this could hit retail holders fastest is the lockup structure — and SpaceX's is unusual. A conventional IPO lockup holds insider shares for 180 days. SpaceX instead built a staggered schedule: 20% of insiders' locked shares become sellable right after the company reports earnings for the quarter ending June 30, with an additional 10% unlocking if shares trade at least 30% above the IPO price for five of the following ten days. Five more tranches of 7% each unlock at 70, 90, 105, 120, and 135 days, another 28% unlocks after Q3 earnings, and the remainder clears entirely at 180 days.
One detail stands out from the rest: a 5% friends-and-family carve-out carries no lockup at all, meaning roughly $3.75 billion in shares could hit the market on day one — alongside the retail shares bought at IPO price. The structure begins releasing meaningful supply just six to eight weeks after listing, far sooner than the December timeline a conventional lockup would have enforced, and was not primarily designed to protect retail holders, according to Investing.com's analysis of the filing terms.
With Musk himself controlling roughly 42% of equity and 85% of voting power under a separate 366-day restriction, but early employees, early investors, and the underwriting bank syndicate all becoming eligible sellers within the same compressed window, market analysts have flagged the period following the first lockup tranches as a potential supply shock — a moment when a large bloc of shares acquired at a fraction of $135 could hit a market where retail investors are holding in at the top.
The thesis confirmed itself within a week
None of the above was theoretical for long. On June 16 — four days after the IPO closed — SpaceX confirmed it would acquire Anysphere, the company behind the AI coding assistant Cursor, for $60 billion in an all-stock deal. The transaction converts an option SpaceX secured back in April: buy Cursor outright, or pay roughly $10 billion to keep a compute-and-training partnership running instead. SpaceX chose to buy. Cursor brings real revenue, an estimated $2.6 billion annualized as of June 2026, but at a price representing close to 60 times that figure, and the deal dilutes existing SpaceX shareholders by approximately 3.4%, according to Morningstar's analysis of the filing.
Because the deal is paid entirely in stock rather than cash, it cost SpaceX nothing on its balance sheet today — but it commits the company to a fourth major bet, layered on top of Starship, Starlink's next-generation buildout, and xAI's existing burn rate, financed by minting new shares rather than new revenue. The market's response arrived almost immediately. SpaceX stock had already run from its $135 offer price to an all-time high of $225.64 on June 16, the day the Cursor deal was announced. By June 18 it had fallen to roughly $182, a pullback of nearly 20% from that peak, as Morningstar cut its fair value estimate on the stock to $62 a share — citing the dilution from the deal — and flagged SPCX as one of the most expensive stocks in its entire coverage universe even after the drop.
Days later, the financing story caught up with the narrative entirely: bankers began preparing a new bond offering of at least $20 billion, intended to refinance the original bridge loan before its September 2027 deadline with longer-term, permanent debt. SpaceX secured investment-grade ratings from Moody's, Fitch, and S&P ahead of that offering — a vote of confidence from credit markets even as equity investors were still digesting the Cursor dilution. The sequence is the clearest possible confirmation of what the prospectus flagged from the start: a company funding multiple frontier bets simultaneously, financed through a rotating mix of debt refinancing and equity issuance, less than two weeks after telling retail buyers it had just raised $75 billion.
What this actually means
None of this means SpaceX is a bad company, or that Starship will fail, or that Starlink's growth is fake. It isn't and it doesn't. Starlink is a genuinely dominant, genuinely profitable infrastructure business that took two decades to build and has no real competitor at its scale. That's real.
But the IPO coverage has mostly collapsed three different businesses — one mature and profitable, one capital-intensive and pre-revenue, one newly acquired and burning more than $6 billion a year — into a single trillion-dollar number and a story about Musk's wealth. The prospectus itself told a more complicated story: a quarter of the raise already committed to old debt, a valuation multiple that assumed nothing went wrong across any of the three segments simultaneously, and a lockup structure that releases insider supply on a faster clock than retail buyers may have assumed when they read "180-day lockup" in the headlines and stopped reading.
The rocket didn't need to be perfect for the IPO to work. The financing did. And it didn't even take a quarter for that to start showing — it took six days, a $60 billion stock-funded acquisition, and a 20% pullback from the post-IPO high. The first real test, the lockup tranches and the bond offering meant to retire the bridge loan, is still ahead. But the balance sheet nobody priced in on June 12 is already the story by June 19.
