The $1.75 trillion price tag requires you to believe Starlink alone justifies three-quarters of the valuation - and that you are welcome to fund the rest.
The competitor headlines are already running: "Want to Buy SpaceX on IPO Day? Here's How." The tone is tutorial-meets-reverie, as if the question is merely logistical rather than economic. SpaceX filed its S-1 registration statement on May 20 and has set a Nasdaq listing price of $135 per share. Selling 555.6 million shares, the company will raise approximately $75 billion - the largest IPO in history - at a post-money valuation of $1.75 trillion. That is 40 percent above the $1.25 trillion private valuation from last year. The gap between what the stock prices in and what the S-1 actually contains is the entire story.
The revenue is real. The losses are the headline.
SpaceX's 2025 full-year revenue was $18.7 billion. Starlink contributed $11.4 billion - up 48 percent from 2024 - and now accounts for 61 percent of the company's total. The Space segment, which includes launch services and NASA crew contracts, brought in approximately $4 billion. The remainder comes from the newly merged AI segment, the product of SpaceX absorbing xAI earlier this year.
Now here is what the S-1 reveals that the PR machine has been working overtime to obscure: SpaceX flipped from a $791 million net income in 2024 to a $4.9 billion net loss in 2025.
The AI segment recorded a $6.35 billion operating loss in 2025. That number alone swallows the entire operating profit of Starlink - which generated $4.4 billion at the operating level - and then some. The Space segment generated $4 billion in revenue and spent nearly as much. Every segment outside Starlink is hemorrhaging.
The company that the IPO is selling you is not a profitable growth engine. It is a cash-generating Starlink business subsidizing two loss leaders, one of which lost more than the entire company's operating income.
Starlink is the only adult in the room
Build the per-unit framework and the $1.75 trillion valuation falls apart under its own weight.
Starlink generated $4.4 billion in operating profit on $11.4 billion in revenue - roughly a 38 percent operating margin. That is strong. It is also the entire profit pool. If you assign Starlink all of SpaceX's operating profit, the segment's contribution to an enterprise valuation - applying even a generous 30x multiple to operating income - comes to roughly $130 billion.
The Space segment is roughly breakeven operationally, which means it adds near-zero free cash flow to the equation. The AI segment is a $6.35 billion operating hole with no revenue disclosed at the segment level large enough to offset it. Starship, the platform supposed to revolutionize launch economics, has cost SpaceX more than $15 billion in development spend and remains years away from crewed orbital operations, let alone the recurring revenue that would justify calling it a growth option.
The xAI merger was pitched as "orbital data centers" - solar-powered satellites running AI inference from space. Any astute infrastructure analyst knows that orbital data centers are science fiction dressed up as a pitch deck. The physics of heat dissipation in vacuum, the latency penalty of uplink/downlink, and the cost per watt of solar-generated power at orbital altitude make this an engineering fantasy. Musk himself told CNBC in February that xAI needed SpaceX for money more than SpaceX needed xAI for technology.
The cross-currents are clear: Starlink is growing at 48 percent and generating real cash. The launch business is a government contract shop with thin margins. The AI segment is a money-losing acquisition that the S-1 does not yet know how to defend. Directionally, Starlink carries the business. Everything else is the bet.
85.1 percent. That is the number the S-1 will not let you forget.
Elon Musk will hold 12.3 percent of Class A shares and 93.6 percent of Class B shares, giving him a combined voting power of 85.1 percent. He gets to vote on every shareholder matter - board composition, related-party conflicts, capital allocation, related xAI infrastructure spending - with virtually zero countervailing power.
You are buying a 14.9 percent slice of the economic upside while surrendering 85.1 percent of the governance to the person who already owns Tesla, X, Neuralink, and The Boring Company.
This is not theoretical. The SEC has already filed a separate civil enforcement action against Musk for failing to timely disclose his beneficial ownership. The New York State Comptroller filed an IPO letter raising governance concerns before the prospectus even went public. The dual-class structure means public shareholders have no meaningful ability to question whether the $6.35 billion AI segment losses represent a strategic investment or an empire-building expense.
The government dependency you are not being asked to price
SpaceX won a $2.29 billion Space Force contract in May 2026 for a military data network. The April 2026 defense awards gave SpaceX contracts where its cost per launch ran $70 million to $130 million cheaper than competitors - but cheaper launch pricing does not mean high-margin revenue when the baseline spend is development-heavy and politically contingent.
Over the past two decades, SpaceX has received at least $22.6 billion in government contracts. That is not a footnote. That is the structural foundation of the Space segment. The Artemis lunar lander contract - $2.89 billion awarded in 2021 - is a fixed-price deal for a program that has already faced delays and performance shortfalls. Musk disclosed in 2024 that Starship was dealing with a 50 percent performance shortfall for lunar missions. The fixed-price nature of the contract means cost overruns land entirely on SpaceX's balance sheet.
Government contract revenue is predictable only when budgets hold and administrations don't pivot. In an election year with shifting defense priorities, $4 billion in launch revenue that barely breaks even is not a moat. It is exposure.
The valuation math does not work at $1.75 trillion
The $1.75 trillion valuation implies that SpaceX is worth more than Apple at its peak, more than NVIDIA today, more than any industrial company in history. But the S-1 shows a company with $18.7 billion in revenue, a $4.9 billion net loss, and one profitable segment.
Even if you apply an extremely generous 50x multiple to Starlink's $4.4 billion operating profit - far above what any satellite communications or connectivity business commands - you get $220 billion for Starlink. You still need $1.53 trillion to come from a Space segment that breaks even and an AI segment that lost $6.35 billion. At $135 per share, that is what the market is being asked to believe.
It is not as good as it looks.
The investor implication
The $75 billion raised from this IPO is not going into Starlink expansion, where the money is already being made. The S-1 risk factors, the xAI merger timing, and the Starship development runway all point in one direction: this capital raise funds the moonshots. Starlink's cash flow is already self-sustaining at 48 percent revenue growth. SpaceX does not need $75 billion to grow Starlink. It needs $75 billion to keep burning through capital on AI infrastructure, Starship development, and orbital fantasies.
The cross-currents are: Starlink is the only segment with real profitability and growth, the launch business is government-dependent with thin margins, the AI segment is a multi-billion-dollar loss with no credible engineering path to "orbital data centers," and Musk's 85.1 percent voting control means public shareholders cannot redirect capital allocation. Directionally, the IPO is a governance extraction event disguised as a growth offering.
The competitor headlines will tell you how to buy SpaceX on IPO day. They won't tell you that at $1.75 trillion, the pricing assumes every loss-making segment becomes profitable while Starlink's growth never slows and Musk never deploys capital toward his next vanity project. The S-1 says otherwise.
You decide which was marketing fluff and which one was analysis.