The S-1: Who Is This IPO Actually For?

The Filing

SpaceX, officially Space Exploration Technologies Corp., filed its S-1 with the Securities and Exchange Commission on May 20, 2026. The company plans to list Class A common stock on the Nasdaq and Nasdaq Texas under the ticker SPCX. Goldman Sachs, Morgan Stanley, and Bank of America are leading the underwriting. The roadshow begins June 5. The target raise is up to $75 billion, which would make this the largest initial public offering in history by a wide margin.

This is the first time SpaceX has disclosed consolidated financials. Until yesterday, the company's numbers were known only through leaks, secondary market transactions, and the occasional Bloomberg report citing anonymous sources. That changed. Every number that follows comes from a document SpaceX filed with the SEC under penalty of law. These are not estimates. These are not projections from analysts. These are the company's own reported results.

The prospectus is 300+ pages of glossy rocket photos, pull-quote callouts of Musk's vision, and language about making humanity multiplanetary. It is also a financial document. And the financials tell a story the photos do not.

Starlink: The Real Business

Start with what works, because it matters. Starlink is a genuinely impressive business. As of March 31, 2026, the service had 10.3 million subscribers across 164 countries and territories, operating approximately 9,600 satellites in low-Earth orbit. The subscriber base doubled in 12 months, growing from 5.0 million in Q1 2025 to 10.3 million in Q1 2026.

Starlink generated $11.4 billion in revenue in 2025 at a 63% adjusted EBITDA margin, producing $7.17 billion in segment operating cash. It is one of the most profitable satellite businesses ever built. The connectivity segment is the reason SpaceX has a legitimate claim to being a world-class company. It is the anchor of the entire valuation.

But there is a number the glossy pages do not emphasize. Average revenue per user (ARPU) fell 23% year over year as Starlink expanded into lower-priced international and consumer markets. Subscriber volume is growing. Revenue per subscriber is declining. This is a common pattern in subscription businesses as they move from early adopters (who pay premium prices) to mass-market penetration (where price sensitivity increases). It does not mean the business is unhealthy. It means the margin profile will change as the subscriber base shifts, and any valuation model that assumes current margins at future scale is making an assumption the S-1 itself does not support.

The xAI Anchor

In February 2026, SpaceX acquired xAI in an all-stock deal. The S-1 restates all historical periods to consolidate the two companies. That restatement transforms the financial picture.

In 2024, before xAI consolidation, SpaceX was profitable: $791 million in net income on $14 billion in revenue. In 2025, with xAI consolidated, the company swung to a $4.94 billion net loss. In Q1 2026: a $4.28 billion net loss on $4.69 billion in revenue. The merger took a profitable company and made it a loss-maker. That is not a characterization. It is arithmetic.

The AI segment (xAI, Grok, and X) lost $6.4 billion from operations in 2025 on $3.2 billion in revenue. In Q1 2026, the losses accelerated: $2.47 billion operating loss on $818 million in revenue. At that quarterly run rate, xAI is losing nearly $10 billion per year.

For context on the revenue growth: xAI grew revenue 22% year over year. That is respectable for most companies. But in the frontier AI market, it is falling behind. OpenAI grew revenue more than 230% over the same period. Anthropic reported $4.8 billion in Q1 2026 revenue and expects $10.9 billion in Q2, with an operating profit of $559 million. Anthropic is approaching profitability. xAI is accelerating its losses.

The AI segment accounts for 17% of SpaceX's consolidated revenue but is responsible for the majority of its losses. The S-1 is asking public market investors to fund the ongoing cash needs of a frontier AI lab that is losing ground to its competitors, packaged inside a rocket company.

The Cash Furnace

The capital expenditure numbers in the S-1 are staggering. SpaceX spent $20.7 billion on capital expenditures in 2025. Of that, the AI segment consumed $12.7 billion, more than the Space and Connectivity segments combined. That money went to data center construction, GPU purchases, and power infrastructure for xAI's training clusters.

In Q1 2026 alone, total capex was $10.1 billion, with $7.72 billion attributed to AI. That is an annualized run rate of approximately $30.8 billion in AI capital expenditures. To put that in perspective, Meta's total 2025 capital expenditure budget (across all of Reality Labs, data centers, and infrastructure) was roughly $35 billion. xAI is spending at a comparable rate to one of the largest technology companies on Earth, while generating a fraction of the revenue.

The cash position reflects the burn. SpaceX ended 2025 with $24.75 billion in cash. By the end of Q1 2026, that had fallen to $15.85 billion. That is nearly $9 billion in cash consumed in a single quarter. Contractual commitments stood at $25.45 billion, with 95% due in 2026 and 2027. Debt load through Q1 was $29.1 billion.

This is why the IPO exists. SpaceX needs capital. The $75 billion raise is not a victory lap. It is a funding round disguised as a public offering. Without it, the company's current cash burn rate would exhaust its reserves in under two quarters. The filing itself acknowledges this: proceeds will fund "working capital, capex (Starship, Starlink expansion, AI infrastructure), and general corporate purposes."

The Anthropic Paradox

Buried in the "Recent Developments" section of the S-1 is a disclosure that reframes the entire xAI narrative. In May 2026, SpaceX entered into Cloud Services Agreements with Anthropic for access to compute capacity across the Colossus and Colossus II data center clusters. The terms: Anthropic will pay SpaceX $1.25 billion per month through May 2029, with capacity ramping at a reduced fee in May and June 2026. Either party can terminate on 90 days' notice.

At full rate, that is $15 billion per year in revenue. Over the full three-year term, the deal could generate up to $45 billion. It is, by a wide margin, the largest single customer contract in the S-1.

Now consider what this means. Musk built the Colossus data centers to power xAI and Grok. He described them as the engine of artificial general intelligence. He claimed AGI would arrive by the end of 2025, then 2026. He recruited 11 cofounders (all of whom have since left) to build the models that would run on this hardware. And now the company that was supposed to achieve AGI on these GPUs is renting them to a direct competitor because, as we covered in The Perpetual Promise Machine, its own fleet was running at 11% utilization.

The financial logic is sound. Anthropic is willing to pay $1.25 billion per month for compute that xAI was using at 11% efficiency. That is a better return than running Grok on it. But the strategic implication is extraordinary: SpaceX's most valuable AI asset is not its own models. It is the hardware it is renting to the company building competing models. The S-1 literally lists Anthropic as both a "key competitor" in AI and one of its largest revenue sources.

The Governance Fortress

The dual-class share structure in the S-1 is not unusual for Silicon Valley. Alphabet, Meta, and Snap all have versions of it. But SpaceX's implementation goes further than most.

Musk holds approximately 42% of equity but controls 85.1% of voting power through Class B shares that carry 10 votes each. He owns 12.3% of Class A shares and 93.6% of Class B shares. He serves simultaneously as CEO, Chief Technology Officer, and Chairman of the Board. The S-1 states explicitly that Musk "can only be removed from our board or these positions by the vote of Class B holders," the same super-voting shares he himself controls. Public shareholders buying Class A shares at the IPO will have no meaningful ability to influence the company's direction, replace its leadership, or challenge any strategic decision.

SpaceX has designated itself a "controlled company" under Nasdaq rules. This allows it to opt out of requirements for an independent board majority, an independent compensation committee, and an independent nominating committee. The company's charter gives Musk the freedom to "engage in businesses that compete directly" with SpaceX. Read that again: the CEO's own employment agreement allows him to run competing companies.

This structure has already drawn opposition from the largest institutional investors in the country. CalPERS, the New York City Comptroller, and the New York State Comptroller issued a joint letter demanding the dual-class structure be removed or modified before the IPO. These three institutions collectively manage trillions in retirement assets. Their opposition is not ideological. It is fiduciary. They are telling their beneficiaries: we cannot responsibly invest in a company where one person holds permanent, irrevocable control and no external accountability mechanism exists.

For context, Gwynne Shotwell, SpaceX's president and the executive most credited with the company's operational excellence, received total compensation of $85.8 million in 2025. She holds 5.46 million Class A and 7.1 million Class B shares. Even Shotwell, the person who actually runs the day-to-day operations, holds a fraction of Musk's voting power.

The Cross-Entity Architecture

The S-1 confirms, in SEC-filed detail, the cross-entity financial relationships we analyzed in The Musk Shell Game. What was previously assembled from leaks and reporting is now disclosed under oath.

SpaceX owns xAI. xAI owns X (formerly Twitter). SpaceX, the company that lands rockets on barges, now owns the social media platform formerly known as Twitter. In January 2026, Tesla (TSLA) invested $2 billion in xAI via Series E Preferred Stock. Following the merger, that converted into 3.8 million SpaceX Class A shares delivered to Tesla on March 12, 2026. Tesla also sold $430 million in Megapack battery storage to xAI in 2025.

The S-1 further discloses that SpaceX owns aircraft used by Musk "in his capacity as CEO of Tesla." The Boring Company, another Musk entity, is not part of the S-1 but shares operational overlap through Musk's leadership. The interconnections between Tesla, SpaceX, xAI, X, and The Boring Company are no longer a matter of speculation. They are a matter of SEC disclosure.

For public market investors, this creates a specific governance risk. Musk simultaneously serves as CEO of Tesla (a public company with its own shareholders) and CEO/CTO/Chairman of SpaceX (soon to be a public company with its own shareholders). The charter allows him to pursue competing interests. Capital flows between the entities at his discretion. Tesla shareholders are funding xAI infrastructure through a $2 billion investment. SpaceX shareholders will be funding the same infrastructure through the IPO proceeds. The question is not whether these transactions are legal (they are disclosed and presumably board-approved). The question is whether any individual investor in any single entity can evaluate whether their capital is being allocated in their interest or in the interest of the broader Musk ecosystem.

The Valuation Test

The target IPO valuation is $1.75 trillion on consolidated 2025 revenue of $18.67 billion. That is approximately 94 times revenue.

For comparison: Apple, the most valuable company in the world, trades at roughly 8 times revenue. Microsoft trades at about 13 times. Amazon, a company investing aggressively in AI and cloud infrastructure, trades at roughly 4 times. Even NVIDIA, which has experienced explosive AI-driven growth, peaked at approximately 35 times revenue. No company in the S&P 500 trades at 90 times revenue. The closest comparisons are pre-revenue biotech companies and speculative small caps.

The bull argument is that you are not buying today's revenue. You are buying the future of Starlink (potential 30+ million subscribers), Starship (the Mars colonization vehicle and next-generation satellite launcher), orbital data centers, and the AI infrastructure business. These are real optionalities with large addressable markets. SpaceX's S-1 claims a total addressable market of $28.5 trillion.

The bear argument is that every speculative premium in history has been justified by the same logic: "you are buying the future." The question is what multiple you pay for a future that includes $41.3 billion in accumulated deficit, accelerating AI losses, declining ARPU in the core business, $29.1 billion in debt, and a governance structure that gives public shareholders no voice. Paying 94 times revenue for a company losing $47 million per day, where you cannot vote the CEO out and he is legally permitted to run competing businesses with your capital, is a bet on faith, not on financial analysis.

Who Is This IPO Actually For?

Every IPO has a target buyer. The structure of the offering tells you who it is.

An IPO designed for institutional investors would include independent board oversight, a path to governance accountability, a clear timeline to profitability, and a valuation anchored to comparable companies. It would not include a CEO who holds 85% of the vote, serves in three C-suite roles simultaneously, can never be removed by public shareholders, and whose charter permits him to compete with the company he runs. Institutional investors at CalPERS, the NYC Comptroller's office, and the NYS Comptroller's office have said exactly this, in writing, before a single share has traded.

An IPO designed for retail investors looks different. It leads with the vision. It emphasizes Mars colonization, orbital data centers, and AGI. It includes glossy photos and pull-quote callouts from the founder. It prices at a multiple that assumes not just success but dominance across every addressable market simultaneously. And it structures voting rights so that the buyers have economic exposure but no control. The IPO buyer gets the upside if the vision works and the downside if it does not, but no ability to course-correct if management goes off track.

SpaceX's S-1 is structured for the second buyer. Every element of the filing, from the governance terms to the valuation target to the marketing language, is optimized for investors who buy the Musk narrative. It is not optimized for investors who buy balance sheets.

This is not inherently wrong. Retail investors have every right to buy what they believe in. Tesla (TSLA) was built on the same model: a visionary founder, a loyal retail base, a premium multiple justified by the future, and governance that insulates leadership from shareholder pressure. It worked spectacularly for early Tesla investors and catastrophically for those who bought at the wrong multiple. The SpaceX IPO is the same bet. The question for each individual investor is whether the current multiple prices in the right amount of future, or whether it prices in a future that, as we have documented across The Perpetual Promise Machine, Tesla: A Trillion-Dollar Bet on Promises, and One Million vs. One Billion, has a long track record of being promised and not delivered.

The Bottom Line

SpaceX has built remarkable things. Starlink is a genuinely transformative satellite internet business generating $11.4 billion in revenue at 63% margins. The Falcon 9 is the most reliable and frequently launched rocket in history. The company has landed and reflown a single booster 34 times. These are real engineering achievements with real commercial value. Nobody credible disputes this.

But the S-1 is not asking you to buy Starlink at a Starlink valuation. It is asking you to buy Starlink plus an AI lab losing $10 billion per year, plus a social media platform, plus a Starship program still in flight testing, plus orbital data center concepts described with the disclaimer "significant technical complexity and unproven technologies," plus a chip fabrication partnership where "neither Tesla nor Intel are obligated to remain," all wrapped in a governance structure where one person holds 85% of the vote and cannot be removed.

The price for this package: 94 times revenue. The accumulated deficit: $41.3 billion. The Q1 net loss: $4.28 billion. The cash burn: $9 billion in a single quarter. The contractual obligations: $25.45 billion, 95% due within two years. And the governance: no independent board requirement, no way to remove the CEO, and charter language that permits the CEO to compete with the company using your capital.

At Wealth Engine Pro, we evaluate companies based on what the numbers say, not what the marketing says. The SpaceX S-1 contains both. The marketing says this is a once-in-a-generation opportunity to own the future of space, AI, and connectivity. The numbers say this is a satellite internet company with a profitable core business, an AI division burning cash at an accelerating rate, declining per-subscriber revenue, and a governance structure that makes public shareholders passengers, not participants. Both of those things are true. Which one you pay attention to determines which kind of investor you are.