How to Read the SpaceX S-1: Key Numbers, Risks, and What SPCX Is Selling

Photo by SpaceX on Unsplash

SpaceX filed its IPO paperwork on May 20, 2026, and the document is 685 pages of consolidated launch, satellite, and AI accounting that does not resemble any version of SpaceX investors have analyzed before. The SpaceX IPO S-1 filing targets a $75 billion raise at a $1.75 trillion valuation, which would be the largest public offering in history by a factor of more than two. Reading it correctly means knowing which four numbers do the real work, and which ones are marketing.

This guide walks through the financials, the share structure, the risk factors, and the mechanics of the offering itself, in the order a working analyst would read them.

Start With Revenue, Then Immediately Distrust It

The top-line figure is $18.674 billion in 2025 revenue, with Q1 2026 alone hitting $4.694 billion, according to Stocktwits' breakdown of the filing. That number is not comparable to any prior SpaceX revenue estimate. As The VC Corner notes in its S-1 teardown, the $18.7 billion figure combines SpaceX launch, Starlink, xAI, and X under common-control accounting after the February 2026 merger, meaning every prior comparison point in circulation is wrong.

The segment split is where the story actually lives. TechCrunch reports that Starlink generated roughly $11 billion of the 2025 total, putting the connectivity business at more than half of company revenue before the AI segment is even counted. Launch services, the original business, is now a minority contributor by revenue.

The quarterly Starlink number is the one to anchor on. Stocktwits confirms the Connectivity segment delivered $1.188 billion in operating income in Q1 2026 against 10.3 million subscribers worldwide. That is a profitable, scaling consumer business. Everything else in the filing is funded by it or competes with it for capital.

How to Read the SpaceX S-1 Filing Without Getting Lost

If you only have an hour with the document, work in this order:

1. Segment reporting in the MD&A, not the consolidated income statement.
2. The capital expenditure line for the trailing twelve months.
3. The voting structure described in the share class section.
4. The risk factors specific to Starship and the xAI integration.
5. The use of proceeds.
6. The underwriter list and lock-up terms.

That sequence isolates the operating business from the moonshots, then forces you to confront how much cash the moonshots burn, then tells you who actually controls the company you are buying into. Anyone teaching how to read the SpaceX S-1 in a different order is selling the story rather than the structure.

The Capex Number That Changes the Bull Case

Stocktwits' filing summary puts Q1 2026 capital expenditure at $10.107 billion in a single quarter. That is more than twice the quarter's revenue. Operating losses came in at $2.589 billion for full-year 2025 and $1.943 billion in Q1 2026 alone, with the 2025 net loss reaching $4.9 billion according to HeyGoTrade's retail-investor breakdown.

The spending is concentrated in three places. TechCrunch reports the Space segment burned $3 billion on Starship R&D in 2025 and another $930 million in Q1 2026. Vested Finance's S-1 read makes the structural point that if you strip Starship out of the Space segment, the underlying launch business is profitable. The headline operating loss is therefore a choice, not a symptom.

The AI segment is a different story. BitMEX's IPO guide pegs xAI losses at roughly $2.5 billion per quarter inside the consolidated entity. That is not R&D that ends when a vehicle reaches orbit. It is a structural burn rate tied to a compute arms race against OpenAI, Google, and Anthropic, and it now sits inside the same P&L investors are being asked to value.

The Anthropic Contract Buried in the Footnotes

This is the line most retail summaries are missing. The VC Corner's teardown flags a multi-year compute contract with Anthropic for COLOSSUS II infrastructure at $1.25 billion per month through May 2029, which works out to roughly $15 billion per year and up to $45 billion over the full term.

That single contract reframes the xAI segment. SpaceX is not only building a frontier model competitor, it is also operating one of the largest commercial AI compute backbones in the world, with a named hyperscaler customer on a multi-year fixed commitment. The contract sits awkwardly next to xAI's own model training spend, because the same physical infrastructure presumably serves both. How the company allocates depreciation and energy costs between the internal AI workload and the Anthropic contract will determine whether the AI segment is a $10 billion revenue line or a $10 billion cost center within twenty-four months. The S-1 does not yet break that out cleanly.

For context on why this compute economics question matters more than the model itself, our piece on how orbital data centers would actually work covers the power and cost math that hyperscalers are running today, and the writeup of Cerebras' wafer-scale engine explains the silicon side of the same race.

Starlink Is the Whole Equity Story

Strip the consolidated numbers down and the SpaceX Starlink financials are what a public-market investor is actually buying. Vested Finance reports Starlink crossed 10.3 million subscribers across 164 countries as of March 31, 2026, more than double the 5.0 million one year earlier. HeyGoTrade notes the network was adding between 750,000 and 1.5 million new subscribers per month at the time of filing.

The constellation behind that growth is itself a moat. Vested Finance puts the current fleet at approximately 9,600 satellites, roughly 75 percent of all active maneuverable satellites in orbit. Direct-to-cell service is live in 30 countries through partnerships with about 30 mobile operators, reaching 7.4 million unique devices a month.

The risk to the Starlink story is not competition, it is dependency. TechCrunch reports that Starship is meant to begin payload delivery in the second half of 2026 and carry the next-generation Starlink V2 mobile satellites starting in 2027. If Starship slips by a year, V2 deployment slips with it, and the direct-to-cell expansion that justifies the $1.75 trillion sticker slows in lockstep. The single largest hidden variable in the SpaceX IPO valuation breakdown is not subscriber growth, it is launch cadence on a vehicle that has not yet flown a commercial payload.

The Voting Structure Is the Real Risk Factor

This is where the S-1 stops being a finance document and becomes a governance document. HeyGoTrade reports that Elon Musk retains 85.1 percent voting control through the super-voting share class. Stocktwits confirms the dual-class structure, where Class A shares carry one vote and Class B shares carry ten votes and elect a majority of the board.

Public SPCX investors are buying economic exposure with effectively zero governance rights. That is not unusual for founder-led tech IPOs, but the magnitude of the gap, an 85.1 percent voting stake against a fraction of the economic interest, sits at the extreme end of the spectrum even compared to Meta or Alphabet at their listings. Any analyst comparing this offering to a traditional aerospace listing is looking at the wrong reference class.

Valuation: What 109x Revenue Actually Implies

The numbers do not flatter the price. HeyGoTrade calculates that at a $1.75 trillion valuation, SPCX prices at 109 to 116 times trailing 2025 revenue and 58 to 65 times forward 2026 revenue. TechCrunch notes that Nvidia, the largest public company by market capitalization at $5.4 trillion, trades on a profitable, scaled chip business with established gross margins. SPCX is asking for one third of Nvidia's market cap on a fraction of the revenue and with a $4.9 billion net loss.

The bull case requires three things to be true at once: Starship reaches commercial cadence on schedule, Starlink direct-to-cell becomes a primary phone service in dozens of markets, and the xAI segment turns the Anthropic compute contract into a recurring hyperscaler revenue line. Each is plausible in isolation. The valuation requires all three.

BitMEX notes that the target valuation has more than doubled since the December 2025 private tender offer, which is the closest thing to a recent market-clearing price the company has produced.

SPCX Stock on Nasdaq: How the Mechanics Actually Work

The SPCX stock Nasdaq listing has a specific structural quirk that retail investors should understand before placing orders. BitMEX reports that Goldman Sachs is the lead-left underwriter, with 21 banks involved in the deal. The VC Corner names Goldman Sachs, Morgan Stanley, BofA, Citi, and JPMorgan as the lead underwriters confirmed in the S-1 itself.

Retail allocation is unusually large. BitMEX reports that retail investors are earmarked for 30 percent of the float, three times the typical mega-cap IPO norm. That is a deliberate distribution choice, not a generosity, because broader retail ownership produces stickier shareholder behavior and tends to dampen post-lockup selling pressure.

For anyone asking how to buy SPCX IPO shares at the offering price rather than on the first day of trading, allocation at offer price flows through the syndicate banks and their retail brokerage partners. HeyGoTrade outlines an expected roadshow start of June 4, pricing on June 11, and trading beginning June 12, 2026, though the company has not confirmed exact dates. The S-1 price range is currently blank per BitMEX's read, meaning final terms will be set during the roadshow. Lockup expiration is expected in December 2026, with the first earnings call as a public company anticipated in September 2026.

For record context, The VC Corner notes Saudi Aramco held the previous largest IPO at $29.4 billion raised. SpaceX's $75 billion target more than doubles that benchmark.

The TAM Claim and Why It Will Not Survive First Earnings

The single most aggressive sentence in the document is the addressable market figure. Stocktwits reports that the S-1 claims SpaceX has identified the largest actionable total addressable market in human history, with a quantifiable TAM of $28.5 trillion. That figure rolls global telecommunications, defense launch, AI compute, and adjacent markets into a single number.

No public-company analyst model survives contact with a $28.5 trillion TAM claim. Sell-side coverage will normalize this to defensible segment-level addressable markets within the first two earnings cycles, and the gap between the S-1 framing and the eventual consensus model will be the source of much of the volatility in the first year of trading. That is the non-obvious risk in the offering: not that the businesses fail, but that the company is forced to reconcile its narrative arithmetic with quarterly disclosure.

The next data point that matters is the final price range when it lands during the roadshow window. Until then, the 685-page document is the only honest source, and the four numbers above, $11 billion in Starlink revenue, $10.1 billion in quarterly capex, $1.25 billion a month to Anthropic, and 85.1 percent voting control, are the ones to read first. For ongoing coverage of the offering and adjacent infrastructure stories, the Space Tech section and the Financial Technology category on AnIntent are the place to track it.