SpaceX IPO: What the Only Disclosed Starlink Financial Filing Actually Shows
The SpaceX S-1 is expected in late May. It will be the first time SpaceX has been legally required to disclose its financial statements - for a company that has been privately setting its own valuation for 24 years. Before the roadshow framing dominates the conversation, here is what the available evidence actually shows.
**The valuation in context**
SpaceX is targeting $2 trillion on approximately $16 billion in 2025 revenue. That is 125 times revenue (not earnings). For context, Apple - one of the most profitable businesses ever constructed - trades at roughly 3 times revenue. Amazon at roughly 4 times. At 125 times revenue, a multiple that would make SpaceX the most expensively valued industrial company in public market history, justifying $2 trillion requires $200 billion in annual revenue. Current projections for 2026 are $23 to $24 billion.
**How $350 billion became $2 trillion**
In December 2024, insiders sold SpaceX shares at a $350 billion valuation. Fifteen months later the IPO target is $2 trillion. Revenue grew roughly 60 percent in that period. The valuation grew nearly 500 percent.
The gap was created primarily by a single transaction. In February 2026, SpaceX acquired Musk's AI company xAI in an all-stock deal. The combined entity was valued at $1.25 trillion: SpaceX at $1 trillion, xAI at $250 billion. Musk controlled both companies. The deal was completed without independent fairness opinions. Valuations were set by boards whose members Musk appointed. The D&O Diary, a specialist publication covering director and officer liability, described the transaction as a "fiduciary stress test."
What $250 billion bought is a chatbot generating approximately $250 million in revenue over six months while losing $2.5 billion. Within weeks of closing, Musk acknowledged the core technology needed to be rebuilt. By March 2026 every co-founder of xAI except Musk had departed.
A quarter of the combined company's pre-IPO valuation rests on a line item the seller priced himself, for a business he has since admitted needs to be rebuilt, staffed by a founding team that no longer exists.
From that $1.25 trillion starting point, the target moved to $1.5 trillion, then $1.75 trillion, then above $2 trillion - each step driven by testing-the-waters conversations guided by 21 banks that earn their fees only if the deal closes.
**The one disclosed Starlink financial filing**
Starlink is presented as a software-like subscription business with 54 percent EBITDA margins. There is one jurisdiction where Starlink was legally required to disclose its actual results: its European subsidiary, filed in the Netherlands.
Those statements show $2.7 billion in revenue and $72 million in net profit. A net margin of under 3 percent.
The gap between 54 percent and 3 percent has a specific explanation. SpaceX conducts approximately 120 Starlink satellite deployment missions per year. SpaceNews estimated that internal Starlink launches are charged at approximately $28 million each - well below the $62 million external commercial rate. At internal pricing, Starlink's annual launch bill is approximately $3.4 billion. At market rates, approximately $7.4 billion. The $4 billion annual difference is the subsidy that makes the margin story work. It turns a break-even satellite operator into a software business on paper.
Subtract that $4 billion and the 54 percent figure collapses. At market launch rates, Starlink as a standalone business almost certainly loses money.
This is not a criticism of the business model. Vertical integration is a genuine competitive advantage. The point is that the margin figure being used to justify $2 trillion is inseparable from an internal pricing arrangement that has never been publicly disclosed and that the S-1 will reveal for the first time.
Amazon is confronting this directly. Project Kuiper must purchase launches at market rates, in some cases from SpaceX itself. If Starlink, with a $4 billion annual internal subsidy, generates under 3 percent net margin in the one jurisdiction where disclosure was required, the arithmetic available to any competitor paying full market rates is not a forecasting problem. It is a structural wall.
**The Starship math**
At $2 trillion, attributing even a conservative 20 percent of the valuation to Starship implies $400 billion assigned to a vehicle that has never served a paying commercial customer.
At a 30 times revenue multiple - generous for any industrial business - $400 billion requires Starship to eventually generate roughly $13 billion in annual commercial launch revenue. At Musk's own stated target price of $10 million per launch, reaching $13 billion means approximately 1,300 commercial launches per year. The entire global launch market in 2025 was roughly 250 missions across all providers combined.
The bull case is that cheap launches create demand that expensive launches cannot, expanding the market by orders of magnitude - is coherent and may prove correct. The point is to show what the current price requires Starship to become, so an investor can decide whether that outcome is being priced as a possibility or as a certainty.
**Governance and the lock-up**
Bloomberg has reported SpaceX is weighing a dual-class share structure giving insiders 10 to 20 votes per share against one for public investors. The person who set the $2 trillion asking price, structured the xAI transaction, and controls the intercompany pricing that determines Starlink's margins will retain effective control of all three of those levers after the IPO. Public shareholders will own the economics. The decisions will remain with the founder.
The Financial Times has reported that bankers are considering eliminating the standard 180-day restriction on insider selling, or replacing it with a staggered release beginning immediately after listing. The retail allocation is being set at up to 30 percent of shares - roughly three times the typical level. More retail buyers than usual entering at IPO price, with insiders potentially able to sell into that demand from day one.
**Four tests to apply when the S-1 drops**
The prospectus will be the first document any investor can read without relying on the seller's representation of the numbers. These four tests separate the investment case from the IPO narrative.
*The margin reality test.* Strip out all intercompany transactions and see what each business unit earns at arm's-length prices. One number matters above all others: what Falcon 9 charges Starlink, and what Starlink's margins look like at the rate outside customers pay. The 54 percent headline and the 3 percent Netherlands filing are two ends of the same question. The S-1 has to answer it.
*The acquisition integrity test.* Who set the price on the most recent major acquisition, on what basis, with what independence, and what is the acquired business worth today? The xAI transaction - $250 billion, self-assessed, no fairness opinion, technology admitted to need rebuilding, founding team departed - is the central test. The S-1 will carry xAI's financials as a subsidiary. Compare what it shows to the implied value the transaction assigned.
*The optionality sizing test.* Identify unproven assets the valuation prices above their demonstrated value. Assign each a required revenue target at a reasonable public market multiple. Ask whether the timeline and capital required is coherent. Starship at $400 billion of implied value requires $13 billion in annual commercial launch revenue - approximately 1,300 launches per year at the target price. The S-1 will show development cost to date, forward capital requirements, and SpaceX's own commercial timeline. That is the first moment the optionality can be sized rather than assumed.
*The governance discount.* Establish what control rights the shares being sold actually confer. A dual-class structure with 10 to 20 votes per insider share isn't inherently disqualifying - Alphabet and Meta use similar structures. But the price paid should reflect the governance premium being surrendered, and that discount should be explicit in any investor's analysis.
**The bottom line**
The S-1 is expected in late May. It will be the first time SpaceX has been legally required to disclose its actual financials, governance structure, and the internal pricing between its rocket and satellite businesses.
Some version of the SpaceX investment case will make sense at some price. The S-1 is where that price gets anchored in disclosed reality rather than narrative. The incomplete information available today suggests the gap between the narrative and the numbers is large enough to warrant patience before day one.