The rocket company closed its first trading day up 19% on Friday, but its own IPO filing says space enabled services and Starlink together make up under 7% of the addressable market it is selling to public investors.
SpaceX priced its initial public offering at $135 a share on Friday and closed at $160.95, a first-day gain of roughly 19% that briefly valued the company at close to $1.8 trillion and pushed Elon Musk's stake past $700 billion on paper. The debut was treated, in much of the financial press, as a vote of confidence in SpaceX the rocket company, the Starlink satellite-internet operator, and the long-running Mars and Moon brand that has defined the firm for two decades. The filing SpaceX submitted to take itself public tells a different story, and the gap between those two narratives is where the next year of this company will actually be decided.
The single most consequential number in the S-1, the registration document filed with regulators ahead of an IPO, is not the IPO price. It is a disclosure in the company's description of its own total addressable market: the share attributable to "space-enabled solutions" plus Starlink, combined, is under 7%. The rest of the addressable opportunity SpaceX is selling public investors on sits in adjacent categories, principally AI-related services delivered from or alongside its existing infrastructure. A reader who only saw the ticker tape would reasonably come away believing they had bought a piece of a space company. The document says otherwise, per the Ars Technica writeup of the debut.
That reframing does not mean the rocket business is a marketing prop. Starlink is the only SpaceX unit generating material revenue at the moment, and launch cadence continues to anchor a real backlog of commercial and government business. What it does mean is that the $1.8 trillion price tag is being justified, in the filing's own framing, by future AI services whose market does not yet exist at scale, rather than by the space business SpaceX has already built. This is the contradiction that animates the disagreement the IPO has already triggered: some analysts have called the valuation "fool's gold," while others characterize it as a rare chance to own a piece of a dominant space and AI infrastructure company, per the same Ars Technica reporting.
The governance structure is the second piece of the puzzle and is harder to undo than the valuation. SpaceX's dual-class share setup concentrates voting power with Musk regardless of how dispersed the public float becomes. Public shareholders get economic exposure to the AI thesis. They do not get corresponding control over how the company pursues it, how it allocates capital between Starlink cash flow and unproven AI services, or how it sequences the long-promised Mars and Moon programs against nearer-term data-center and enterprise contracts. The "first trillionaire" framing that followed Friday's close is, in this sense, a milestone that is also conditional: the paper wealth is real only as long as the share price holds, and the share price will hold only as long as the AI thesis the S-1 sells continues to attract buyers willing to underwrite a future rather than a present.
Both cases remain coherent. The bull case rests on Starlink's actual customer base and cash flow, the launch manifest's near-term revenue visibility, the government's apparent appetite for SpaceX-adjacent national-security work, and the optionality of a vertically integrated space-to-AI stack that no competitor can replicate quickly. The bear case rests on a total addressable market in which the proven segment is a sliver, an AI services category crowded with hyperscalers and well-funded startups, and a governance regime that lets the company pursue the bull case and the bear case simultaneously without shareholder recourse. Neither side has to be wrong yet. The S-1 made sure they can both be true.
What to watch over the next four quarters is narrow and specific. First, whether SpaceX breaks out Starlink revenue, launch revenue, and any new AI-services line items in its first quarterly filings, or continues to report them in a way that keeps the 7% number from being re-anchored in public discussion. Second, whether any of the AI-services contracts the S-1 alludes to are named, dated, and counterparty-identified before the lock-up expiry, the period after an IPO during which insiders are restricted from selling their shares. Third, whether the dual-class structure is used to approve a transaction, capital allocation, or program priority that a public shareholder vote would not have approved. None of those is a verdict. All of them are the places where the $1.8 trillion question will be answered, or at least narrowed.