The number that stopped me was not the $2 trillion valuation.
It was $791 million. That is what SpaceX made in net income in 2024. A profitable, growing aerospace company with a genuine moat in launch services and satellite internet.
Then I kept reading.
In 2025, after the xAI merger closed, that same company posted a $4.94 billion net loss. In the first quarter of 2026 alone, it lost $4.28 billion. The merger did not add a profitable AI business to a profitable rocket company. It did the opposite. And the IPO is asking you to buy the combined entity at up to $2 trillion.
I downloaded the full 300-page S1 and uploaded it to Claude Opus 4.8. Asked it to do one thing: find what the document is downplaying. Not the headline risks in the executive summary. The numbers buried in the comparatives, the footnotes, and the structural disclosures that change the picture if you actually read them.
Here are the five things Claude surfaced that the news coverage is not talking about.
1. The Valuation Is Being Sold Against a Business That Does Not Exist Yet
SpaceX generated $18.7 billion in revenue for full-year 2025. That number is real, growing at 33% year-over-year, and is anchored primarily by Starlink at $11.4 billion, roughly 61% of total revenue.
The valuation at $1.75 trillion is approximately 94 times that revenue. At $2 trillion it is 107 times.
That multiple is not being justified by what SpaceX is today. It is being justified by what the S1 claims the total addressable market will be.
Here is where it gets interesting. The S1 breaks the business into three segments: Space, Connectivity, and AI. The company's stated TAM by segment:
- Space: $370 billion
- Connectivity: $1.6 trillion
- AI: $26.5 trillion
The businesses generating actual revenue today, Space and Connectivity combined, account for under 7% of the total TAM number being used to justify the valuation. The other 93% is the AI segment. Enterprise Applications within AI alone is listed at $22.7 trillion.
You are being asked to pay 2026 dollars for a business that will need to capture a meaningful share of $26.5 trillion in a market that is actively contested by every major technology company on the planet.
Claude's specific flag: the S1 does not explain the mechanism by which SpaceX captures AI TAM that is currently dominated by Anthropic, OpenAI, Google, and Microsoft. The TAM is stated. The competitive moat is not.
2. The xAI Merger Turned a Profitable Company Into a Loss-Making One
This is the number most people miss because it requires reading the historical comparatives.
In 2024, before the xAI merger, SpaceX posted $791 million in net income. Profitable. Growing. A genuine business with a genuine moat in launch services and satellite internet.
In 2025, after the xAI merger closed, SpaceX posted a $4.94 billion net loss. In Q1 2026 alone, the company lost $4.28 billion in a single quarter. The accumulated deficit now stands at $41.3 billion.
The merger did not add a profitable AI business to a profitable rocket company. It added a loss-making AI infrastructure business, including X, to a rocket company and is now selling the combined entity at a valuation that prices the AI business as though it is the primary asset.
AI losses are running at approximately $2.5 billion per quarter. That is not R&D spending that will taper. AI infrastructure at the scale xAI is building requires sustained capital expenditure that compounds, not contracts, in the near term.
What Claude flagged: the transition from profit to loss happened in a single year and coincides precisely with the xAI consolidation. The S1 presents this as a strategic investment phase. What it does not model is the timeline on which that investment phase ends.
3. Starlink Is Growing by Volume. The Per-Subscriber Economics Are Deteriorating.
Starlink is the most investable part of this company and the only segment generating real scale revenue. Subscriber growth is genuinely impressive: from 5 million subscribers in Q1 2025 to 10.3 million in Q1 2026, more than doubling in a year.
Here is what the S1 buries in the segment breakdown.
ARPU, average revenue per user, fell 23% year over year. Revenue is growing because there are more subscribers. The revenue per subscriber is declining as Starlink pushes into lower-priced international and consumer markets to hit volume targets.
This matters because the Starlink bull case is built on the assumption that a large subscriber base at healthy margins produces durable cash flow. The actual trend in the document is a large subscriber base at shrinking margins. Those are different businesses.
Claude flagged it this way: the headline subscriber growth number and the ARPU decline are both in the document and both real. The executive summary features the first. The second appears only in the segment financials. If you value Starlink on subscriber growth alone, you are using half the picture.
4. The Retail Allocation Is Three Times Normal
Standard practice for a mega-cap IPO is to allocate approximately 10% of the float to retail investors. Institutions get the majority. Retail gets the remainder.
The SpaceX IPO is allocating 30% to retail. Three times the standard.
The selling group named in the S1 includes Schwab, Fidelity, Robinhood, SoFi, and ETRADE. These are retail platforms. The distribution infrastructure has been built specifically for individual investors to participate at the IPO price, not through secondary market purchases after listing.
This is being marketed as democratisation. An opportunity for ordinary investors to get in at the same price as institutions.
The less promotional framing: when venture capital and private investors who have held SpaceX equity for years at much lower valuations are exiting, someone has to buy their shares. That someone, at three times the normal allocation, is retail.
Claude's observation: the S1 does not hide this. It discloses the 30% allocation clearly. The disclosure is in the document. The framing is not.
5. The Voting Structure Makes Accountability Structurally Impossible
This is the section of the S1 that requires the most attention and receives the least.
Elon Musk holds approximately 42% of SpaceX equity. His voting power is approximately 85%. That is not a rounding error or a minor imbalance. It means every shareholder who is not Elon Musk, collectively, controls 15% of the voting power in a company being sold at up to $2 trillion.
The S1 also includes an arbitration clause and class action waiver. Shareholders who have a grievance cannot bring a class action lawsuit. They must pursue individual arbitration.
The combination: a voting structure where Musk cannot be outvoted on any material decision, and a legal structure where shareholders cannot pursue coordinated legal action if something goes wrong.
Claude's specific note: this governance structure is disclosed in the S1 in the risk factors section. It is presented as a feature of how the company operates, not as a risk to investor returns. The actual risk, that no mechanism exists for shareholders to hold management accountable if capital allocation decisions destroy value, is not framed that way in the document.
What the Prediction Markets Are Pricing
Your audience thinks in positions, not just opinions. Here is what Polymarket shows as of this week.
The leading outcome on Polymarket's SpaceX market cap contract is $2 trillion plus, sitting at 47% probability. The $1.8 trillion to $2 trillion range is at 18%.
The market is pricing the AI TAM story. The prediction markets believe the IPO will close at the top of the range or above it.
The financial data in the S1 tells a different story. The businesses generating actual revenue today, Falcon launch services and Starlink, would support a valuation significantly below $1.75 trillion if priced in isolation on current earnings multiples. The premium above that is entirely the xAI and AI segment TAM claim.
Two ways to think about this: either the market is right that SpaceX will capture a meaningful share of a $26.5 trillion AI market, in which case the premium is justified. Or the market is pricing a narrative that the S1 itself cannot substantiate with a competitive differentiation argument.
Polymarket has the roadshow and IPO pricing as a separate set of markets. Both are worth watching this week as the actual institutional book-building data starts to surface.
What Claude Said When I Asked for the Summary
After reading the full document, I asked Claude one more question: if you had to describe what is being sold here in one paragraph, what would you say?
Its response:
Prompt
SpaceX is selling three businesses under one ticker: a profitable launch services company (Falcon 9, Falcon Heavy), a growing but margin-compressing satellite internet business (Starlink), and a loss-making AI and social media infrastructure business (xAI, including X). The valuation prices the third business as the primary asset using a $26.5 trillion TAM for which no competitive differentiation is explained. The first two businesses, which generate actual revenue and have genuine competitive moats, would trade at a significantly lower valuation in isolation. The IPO pricing is asking buyers to pay for the third business at scale before that scale exists.
That is the document, compressed to what it actually says.
What to Do With This
This is not a recommendation to buy or avoid the SpaceX IPO. The launch services business is genuinely remarkable. Starlink's subscriber growth, doubling from 5 million to 10.3 million in a single year, is real. The reusable rocket moat is defensible in a way few technology businesses can match.
The question is not whether SpaceX is a great company. It is whether $1.75 to $2 trillion is the right price for what is being sold today, under this governance structure, at this stage of the xAI integration.
Those are different questions. The S1 answers the first. It is less useful on the second.
The document is public and free. Download it from SEC EDGAR under registration number 333-296070. Upload it as a file to Claude Opus 4.8 inside a Project. Then use this prompt:
Prompt
I have uploaded the SpaceX S1 filing. Read the full document before producing any output.Then produce the following:1. The three financial disclosures that appear in the document but are not in the executive summary and that materially change the investment picture
2. Any metric where the headline number and the underlying trend point in opposite directions
3. The single risk factor that is disclosed in the document but framed in a way that understates its significance
4. What the valuation implies about xAI's future revenue contribution, stated as a specific number derived from the current multipleDo not summarise what the document says about SpaceX's strengths.
Focus entirely on what the document reveals when read against itself.
The roadshow started June 4. Pricing is expected around June 11. You have a week.
Decide before the institutional book closes. After that you are buying on secondary market momentum, not fundamental analysis.
This article is financial commentary, not financial advice. Do your own research before making any investment decision.
Follow @damidefi on X for daily Claude AI tools, crypto analysis, and the full journey to 100K. Bookmark this. Share it with one person about to buy SPCX without reading the S1.
