SpaceX IPO: Hype or Real?
SpaceX filed for the first trillion-dollar IPO in history — a reported ~$1.75T target on a business that lost $2.6B last year. We run it through the evidence test: what is confirmed, what is speculation, and which public tickers actually move on the answer.
What a ~$1.75T target implies on SpaceX’s 2025 revenue
≈94× revenue
The S-1 is finally public, and it settles the rumors with audited numbers. Starlink is real: $11.4B of 2025 revenue, a 39% operating margin, 10.3M subscribers. But the same filing shows the launch segment and the newly folded-in xAI unit lose money — xAI alone lost $6.4B in 2025 — so the whole company posted a $2.6B operating loss and a $4.9B net loss. At a reported ~$1.75 trillion, SpaceX would price near 94x revenue: about four times Nvidia’s multiple, on a company that is not yet profitable. Add the segments up at sane multiples and you reach roughly $760B; the rest of the ask is option value on Starship and AI. You cannot buy private SpaceX shares, so this piece maps the public proxies that do move — RKLB and ASTS (competitive pressure), NVDA and GS (beneficiaries), HOOD (the 30% retail channel) — and grades the thesis on the reported ~June 11–12 pricing and the Nasdaq-100 fast-entry review in early July.
The contradiction
SpaceX’s S-1 is the most transparent document in the trillion-dollar IPO class — and that transparency is exactly what complicates the bull case. The numbers are audited and real: Starlink is a genuinely profitable satellite-broadband business throwing off $4.4 billion of operating income at a 39% margin. But it is bolted to a launch business that does not yet cover its costs and to xAI, the AI unit folded in this year, which lost $6.4 billion in 2025 alone. Add the parts up at defensible multiples and you reach roughly $760 billion. The reported ~$1.75 trillion target needs about another trillion dollars of value, and the only place it can come from is the option value of a loss-making AI bet and a Starship program with no commercial revenue yet. That is the line between hype and real: Starlink is real, the listing is real, the engineering is real. The trillion-dollar premium stacked on top of Starlink is, for now, a belief — and the IPO is the moment the market votes on whether to fund it.
What the headline says
The headline: the biggest IPO in history
SpaceX targets a reported ~$1.75 trillion listing — the first trillion-dollar IPO ever, Starlink at 10.3M subscribers, up to 30% of the float reserved for retail.
What the data says
The filing: only one of three businesses makes money
Starlink earned +$4.4B in 2025; the xAI segment lost −$6.4B; SpaceX posted a −$2.6B operating loss and a −$4.9B net loss — and ~$1.75T prices it near 94× revenue.
Chapter 01
The Price: What $1.75 Trillion Actually Asks You to Believe
At a reported ~$1.75 trillion on $18.7B of 2025 revenue, SpaceX would list at roughly 94× sales — above every public mega-cap and just past Palantir, the most expensive large-cap software name. Nvidia, the most valuable profitable company on earth, trades near 21×. The price is the whole story, so start there.
Before any debate about rockets or satellites, look at the multiple, because the multiple is what you are being asked to underwrite. A reported target of about $1.75 trillion against $18.7 billion of 2025 revenue is a price-to-sales of roughly 94×. To put that in company: Nvidia — the most valuable profitable business in the world, growing data-center revenue at triple digits — trades near 21× sales. Palantir, the richest large-cap software multiple in the public market, sits near 74×. Apple is around 8×, Microsoft near 11×, Amazon under 4×. SpaceX at 94× would be the most expensive large company in the market by this measure, and it would get there while losing money at the operating line. None of this says the business is bad — it says the price embeds a decade of flawless execution before you buy a share. A multiple like this is not a valuation; it is a forecast. The rest of this piece is about whether the filing supports the forecast.
Price-to-Sales: SpaceX Target vs Public Mega-Caps
Source: SpaceX Form S-1 (2025 revenue $18.7B) + reported ~$1.75T target (Bloomberg); public P/S from market data, 2026-06-01
Implied price-to-sales
≈94×Reported ~$1.75T target on $18.7B of 2025 revenue
Nvidia, for comparison
21.5×The most valuable profitable company on earth
The multiple gap
≈4× NvidiaNo profitable public mega-cap trades near 94×
Chapter 02
What’s Confirmed: Two Companies Wearing One Ticker
The S-1 breaks SpaceX into segments, and the split is stark. Starlink earned +$4.4B of operating income in 2025 on $11.4B of revenue — a 39% margin. The AI segment (xAI, folded in this year) lost −$6.4B. Net of everything, the consolidated company posted a −$2.6B operating loss. You are buying one profitable business stapled to one large cash furnace.
The most useful thing the filing does is force the segments into the open. Starlink, the satellite-broadband business, is the engine: $11.4 billion of 2025 revenue, $4.4 billion of operating income, a 39% operating margin and roughly 63% on an adjusted-EBITDA basis. By any standard that is a real, scaling, profitable business — the part of SpaceX that most resembles a public-market compounder. But Starlink does not get to list alone. The same consolidated entity now contains xAI, the artificial-intelligence unit brought in this year, which lost $6.4 billion in 2025 as it built the Colossus supercomputer and trained frontier models. Add the launch business (which carries its own losses as Starship scales) and the consolidated result flips negative: a $2.6 billion operating loss and a $4.9 billion net loss for 2025, against a $41.3 billion accumulated deficit. So when you buy the IPO you are not buying “the profitable satellite company.” You are buying a holding company in which one division earns $4.4 billion and another loses $6.4 billion, and the question is how much you will pay for the bet that the second division becomes the first.
2025 Operating Income by Segment ($B)
Source: SpaceX Form S-1, FY2025 segment disclosure (CONFIRMED)
Starlink operating income
+$4.4B$11.4B revenue, 39% operating margin (2025)
AI / xAI operating loss
−$6.4BOn $12.7B of AI capex — building Colossus
Consolidated result
−$2.6BOperating loss; −$4.9B net loss for 2025
Chapter 03
What’s Confirmed: The Starlink Engine Is Genuinely Accelerating
Starlink subscribers roughly doubled in a year — from about 5.0M in early 2025 to 9.2M at year-end to 10.3M by Q1 2026. That growth is the strongest fact in the bull case. The caveat is also confirmed: revenue per user has fallen from $99 in 2023 to $66, so the company is adding users faster than it is holding price.
If the previous section is the bear’s favorite chart, this is the bull’s. Starlink’s subscriber base went from roughly 5.0 million in early 2025 to 9.2 million at year-end and 10.3 million by the first quarter of 2026 — a near-doubling in twelve months, and the kind of curve that justifies a premium for a hardware-plus-network business. This is what “real” looks like: a confirmed, audited, fast-growing subscriber engine with a 39% operating margin behind it. But the same disclosures carry the tension. Average revenue per user has slipped from about $99 in 2023 to roughly $66 now, because the fastest growth is coming from lower-priced international and consumer tiers rather than premium enterprise and maritime. Subscribers up ~2×, ARPU down about a third. That is not a broken model — it is the classic trade-off of a network business scaling into mass-market price points, and it is exactly the variable that decides whether Starlink alone can ever grow into a meaningful fraction of a $1.75 trillion price. The growth is real. The pricing power is the thing to watch.
Starlink Subscribers (millions)
Source: SpaceX Form S-1 + reported subscriber disclosures (CONFIRMED / REPORTED)
Starlink subscribers
10.3MQ1 2026, up from ~5.0M a year earlier
Revenue per user
$66Down from $99 in 2023 — ARPU is falling
The tension
Subs 2×, ARPU −33%Growth is real; pricing power is eroding
Chapter 04
What’s Speculation: Add Up the Parts and You Get to $760B
Value the segments independently — Starlink on a broadband-network multiple, launch on a services multiple, xAI on a frontier-AI comp — and a reasonable sum-of-the-parts lands near $760B, with a stretch high case around $1.5T. The reported ~$1.75T target sits above even the high case. The gap, roughly a trillion dollars, is the part the market has to take on faith.
The cleanest way to separate hype from real is to build the valuation from the bottom up and see where it lands. Take the pieces on their own merits: Starlink as a profitable, fast-growing broadband network; the launch business as a dominant but lumpy services provider; xAI as a frontier-AI lab with no profits and enormous capital needs. Run those through defensible multiples and a low case sits near $158 billion, a mid case near $760 billion, and an aggressive high case around $1.5 trillion. Every one of those is below the reported ~$1.75 trillion target. The difference between the mid case and the ask — close to a trillion dollars — is not in any current cash flow. It is the option value of two things that do not yet make money: Starship at commercial scale and xAI as a winner in the most capital-intensive race in technology. That option value might pay off spectacularly; SpaceX has earned the benefit of the doubt more than once. But “might pay off” is the definition of speculation, and at this price the speculation is most of the market cap. The bull case is not that Starlink is real — it plainly is. The bull case is that the trillion-dollar premium on top of Starlink is real too.
Sum-of-the-Parts Scenarios vs Reported Target ($B)
Source: MarketDecode segment SOTP framework (Starlink / Launch / xAI at independent multiples) vs reported ~$1.75T target — ANALYSIS
Sum-of-the-parts (mid)
$760BSegments at independent, defensible multiples
The asking price
~$1.75TReported target (Bloomberg)
The gap to fill
≈$1TOption value on Starship + a unit that lost $6.4B
Chapter 05
What You Can Actually Trade: The Public Proxy Map
There are no retail-tradable SpaceX shares, so the exposure splits two ways. Suppliers and bankers benefit if the deal works: NVIDIA (GPUs into xAI’s Colossus), Goldman Sachs (lead underwriter), Robinhood (the reported 30% retail allocation channel), Vertiv (data-center power and cooling). Listed rivals get squeezed if SpaceX wins: Rocket Lab and AST SpaceMobile most directly, with Iridium and parts of Lockheed Martin behind them.
This is the section that matters most for anyone reading from a brokerage account, because the honest answer to “how do I buy SpaceX?” is: you cannot, not as private stock. What you can do is map the public names whose cash flows or competitive position move on SpaceX headlines, and the map has two halves. On the beneficiary side sit the picks-and-shovels: NVIDIA, which sells the GPUs that xAI is buying by the billions for Colossus; Goldman Sachs, positioned as a lead underwriter that earns fees whether or not the stock holds; Robinhood, the reported conduit for the up-to-30% retail allocation; and Vertiv, levered to the power and cooling that every AI data center — including xAI’s — has to build. On the pressured side sit the listed competitors: Rocket Lab, whose launch economics compress if Starship delivers cheap heavy lift, and AST SpaceMobile, in direct line of fire from Starlink’s direct-to-cell push; Iridium and pieces of Lockheed Martin face thinner but real overlap. The chart plots each name by how strongly it is exposed and by whether it benefits from or is pressured by a SpaceX win. The point is not a recommendation — it is a watchlist: these are the tickers that will actually print a reaction when the IPO prices.
SpaceX Public Proxies: Exposure vs Direction
Source: MarketDecode proxy framework — exposure scored 1–5; direction = benefits (+) vs competitive pressure (−), 2026-06-01 — ANALYSIS
Most exposed rivals
RKLB · ASTSStarship and direct-to-cell compress their TAM
Most exposed beneficiaries
NVDA · GSGPU supply into xAI; lead underwriter fees
The retail channel
HOODReported up-to-30% of the float reserved for retail
Chapter 06
What Would Break the Story — and What to Watch
The clearest bear signal is observable right now: SpaceX spent $20.7B building in 2025 against $18.7B of revenue — it is outspending its entire top line, 111% capex-to-sales, versus ~22% at Microsoft and ~9% at Nvidia. The thesis grades on dated catalysts: reported pricing around June 11–12, then the Nasdaq-100 fast-entry review in early July. Founder voting control near 85% is the structural asterisk on the whole deal.
End where the risk is most concrete. The single most telling number in the filing is not the loss — it is the capital intensity. SpaceX spent $20.7 billion on capital expenditure in 2025, more than its entire $18.7 billion of revenue, a capex-to-sales ratio of 111%. For scale, Microsoft runs near 22% and Nvidia near 9%. A company outspending its own revenue is not unusual for an infrastructure build, but it means the business consumes cash to grow and will keep tapping markets to fund Starship and xAI; the IPO is partly that fundraise. What would break the story is straightforward: the AI losses widen faster than Starlink’s profits grow, ARPU keeps falling, or the IPO prices well below the floor and the listed space complex re-rates with it. What would prove it: Starlink margins hold as subscribers compound, Starship reaches commercial cadence, and the stock prices near the top of the range and holds. The grade dates are real and close. Reported pricing lands around June 11, with first trade around June 12; Nasdaq’s fast-entry rule then forces index funds to buy within about fifteen trading days, putting a Nasdaq-100 review in early July with a reported $27–$60B of mechanical demand behind it. And underneath all of it sits governance: founder voting control of roughly 85% means public shareholders are buying the economics, not the steering wheel. Hype or real? Starlink is real, the engineering is real, the listing is real. The trillion-dollar premium is a bet — and the next six weeks start settling it.
Capex as a Share of Revenue (2025)
Source: SpaceX Form S-1 ($20.7B capex / $18.7B revenue) vs reported company capex (REPORTED / CONFIRMED), 2026-06-01
Capex vs revenue
111%$20.7B capex on $18.7B of 2025 sales
Next dated test
Pricing ~Jun 11First trade ~Jun 12; Nasdaq-100 review ~early Jul
Founder voting control
~85%Public holders get economics, not control
Resolution window — 3 months
What would confirm or invalidate this read
Confirmation
SpaceX prices its IPO at or above the reported ~$1.75T floor around June 11–12 and holds the level into the Nasdaq-100 fast-entry review in early July, with index inclusion forcing the reported $27–$60B of mechanical demand. If the listing prices firm and Starlink’s next disclosed subscriber/ARPU update shows margins holding, the AI-era multiple is validated and the beneficiary proxies (NVDA, GS, HOOD, VRT) carry the read-through; competitive proxies (RKLB, ASTS) stay pressured.
Invalidation
SpaceX prices materially below the reported ~$1.75T floor, or the stock trades down 20–30% in its first weeks, signaling the market will not fund the trillion-dollar premium on a loss-making consolidated entity. In that case the listed space complex (RKLB, ASTS, IRDM) re-rates lower with it, and the read flips from “price the bet” to “the bet was too rich.” Widening xAI losses or further ARPU erosion in the next disclosure would reinforce the invalidation.