SpaceX went public on June 12 at $135 a share in the largest IPO in history, raising $85.7 billion and valuing the company at $1.75 trillion. Within a week the stock had rocketed past $225 intraday. By July 7, it was fast-tracked into the Nasdaq-100 under rules the exchange rewrote specifically to let mega-IPOs skip the usual waiting period, forcing every index fund tracking the benchmark to buy shares whether they wanted to or not. That forced buying is normally the kind of mechanical tailwind that puts a floor under a stock. It didn't. SPCX fell 17% in the days after joining the index, dipping below its $135 IPO price intraday by mid-July before settling into the $122-to-$130 range by July 19 — a company that had briefly been worth roughly $2.66 trillion in public-float terms now trading closer to $1.6 trillion.
The easy read is that this is a standard post-hype comedown: a blockbuster debut, a wave of retail FOMO, and the inevitable reversion once the excitement burns off. That's the version most market coverage has settled on, and it isn't wrong so much as incomplete. It treats the sell-off as a mood swing. What actually happened is more specific and more interesting: the Nasdaq-100 inclusion was a real-time test of whether index mechanics could override fundamentals, and fundamentals won anyway. Roughly $4 billion in mandatory passive buying hit the stock in a single week and it kept falling. That's a natural experiment most IPOs never get, and it suggests the market isn't cooling off on SpaceX generally — it's pricing something specific that going public forced into the open for the first time: what SpaceX's vertical integration bet actually costs, not just what it's supposed to be worth.
To see why, it helps to go back to what SpaceX actually became in the run-up to its IPO. In February, Musk merged SpaceX with his AI company xAI, describing the combined entity in a statement as a "vertically-integrated innovation engine" spanning both Earth and space, folding in Grok, the Colossus data center, the X platform, and Starlink's satellite network alongside the rocket business. Then in April, days after that merger, SpaceX secured a call option to acquire the AI coding startup Cursor for $60 billion in stock, or pay a $10 billion breakup fee if it walked away. Four days after the IPO closed, with SPCX trading more than 50% above its offer price, SpaceX exercised that option and announced the formal acquisition. Because the entire purchase was paid in freshly inflated stock, Pitchbook analyst Franco Granda noted the deal showed how "the IPO gave SpaceX a valuation and a premium currency," effectively letting Musk buy one of the fastest-growing software companies on record for a fraction of its real cost to shareholders. The S-1 was explicit about why: Cursor's developer traffic was expected to feed real-world coding data back into Grok's training pipeline, tying the acquisition directly to the AI buildout rather than anything resembling the launch business investors originally associated with SpaceX.
The market's answer to that deal wasn't enthusiasm. Forbes reported that the stock shed roughly $600 billion in market value over the two trading days after the announcement, as SpaceX's market cap fell from its Tuesday peak to about $2.37 trillion by Thursday. That's the detail the "hype fading" narrative skips past: investors didn't wait for a slow build of disappointment. They moved within days once the scope of the AI ambitions inside the vertically integrated pitch became concrete and priced in stock, rather than described in a memo.
Now add what mandatory public disclosure has done since. As a private company, SpaceX could describe itself however it liked — a vertically integrated engine spanning rockets, satellites, AI, and social media, one balance sheet for all of it. As a public one, it has to report segments, and the segment math tells a more specific story than the pitch does. SpaceX's 2025 revenue was $18.7 billion; at its June 16 peak, the market was pricing the company at roughly 142 times that figure. After the pullback, it still sits near 87 times last year's sales — high even by the standards of a company growing revenue 33% a year. The reason the multiple hasn't compressed further, according to analysts tracking the disclosures, is that the AI and space segments — the businesses built from the xAI merger and now the Cursor acquisition — are running at losses large enough to erase the profits Starlink is generating on its own. In other words: the subscription satellite-internet business is the part of SpaceX that actually makes money today. The AI empire built to sit alongside it is, for now, a drag on that business rather than an amplifier of it. That's not the story "vertical integration" was supposed to tell. It's the story the disclosure forced into view once the company had to report it that way.
That gap between the narrative and the segment math is also visible in how far sell-side price targets still sit from where the stock is actually trading. In the days after the IPO and Nasdaq-100 inclusion, Morgan Stanley initiated coverage with a $300 target, Bernstein at $239, RBC at $225, and UBS at $210 — all built on the bull case that Starlink's expansion, Starship's flight cadence, and the AI infrastructure buildout would eventually justify the valuation together. Those targets haven't moved even as the stock has fallen closer to $120. Either the banks are right that the AI segment's losses are a temporary cost of scaling toward Musk's stated goal of $1 trillion in annual revenue by 2030, or the targets were set on the same story the market has since started to discount — that combining all of it into one balance sheet is worth a premium rather than a discount until the AI piece stops losing money. Cathie Wood's ARK funds have been buying into the decline, which at least confirms the bull case still has believers with real capital behind it, not just analyst notes.
The next test of that question arrives in August. SpaceX's IPO structure unlocks 20% of shares held by employees and pre-IPO investors on the second trading day after its second-quarter earnings report, guaranteed regardless of price. A further 10% would have unlocked early if the stock had closed at or above $175.50 for at least five of the ten trading days before that report — a threshold the stock was nowhere near meeting by mid-July. That detail matters beyond the mechanics of supply hitting the market: it means the company's own lockup structure was built assuming a share price trajectory the stock hasn't come close to sustaining since its first week of trading. When the guaranteed 20% unlocks, it will land at a moment when the market is still actively working out how much of SpaceX's valuation belongs to Starlink and how much belongs to an AI arm that hasn't yet shown it can pay for itself.
None of this means the vertical integration bet is wrong on its own terms — Musk's argument that AI's growth is bottlenecked by terrestrial power and cooling, and that space-based data centers eventually solve that constraint, is a real thesis with real engineering behind it, and the banks backing $200-plus targets are underwriting exactly that outcome. But the framing that's dominated coverage so far — a hot IPO cooling off, a stock "disappointing" after the hype wore thin — misses what actually changed. Going public didn't just give SpaceX a stock price. It gave the market a segment report. And what that report shows, for now, is that the empire Musk built to be worth more than the sum of its parts is trading like the market believes the opposite: that until Grok, Cursor, and the rest of the AI buildout stop costing more than Starlink earns, vertical integration is a discount, not a premium.
