La mayor salida a bolsa de la historia de Wall Street no es una empresa de cohetes

Key Takeaways

• The record is real: SpaceX raised $75 billion at $135 per share, the biggest Initial Public Offering (IPO) in history, more than two and a half times Saudi Aramco’s 2019 record of $29.4 billion.
• Rockets are the minority business: Launch generated about 22 percent of 2025 revenue. Starlink generated 61 percent and nearly all of the operating profit.
• The money is pointed at AI: The first specific item in the filing’s use-of-proceeds list is “the expansion of our AI compute infrastructure,” and the newly absorbed xAI segment out-spent the rest of the company on capital expenditure by more than three to one last quarter.
• You get shares, Musk keeps the votes: Elon Musk holds approximately 82.4% of voting power after the offering, and SpaceX will operate as a “controlled company” exempt from several Nasdaq governance rules.

The Ticker Says Rockets. The Filing Says Something Else.

SpaceX began its first day as a public company on June 12, 2026, trading on the Nasdaq under the ticker SPCX after pricing the largest IPO in financial history the evening before. The company sold 555,555,555 Class A shares at $135.00 each, raising $75 billion and valuing the 24-year-old company at roughly $1.77 trillion at the offer price, the seventh most valuable company in the United States, ahead of Tesla. When the stock finally opened on June 12, the first trade crossed at $150, an 11% pop.

For scale: the previous record holder, Saudi Aramco, netted $29.4 billion in its 2019 listing. SpaceX raised more than two and a half times that in a single offering, and its underwriters hold an option to buy an additional 83.3 million shares on top. Demand was never in question. The order book attracted bids for multiple times the shares available days before pricing, per Bloomberg.

The retail story is just as unusual. SpaceX targeted a retail allocation of roughly 30% of the deal, around $22.5 billion worth of stock, where a typical IPO hands individual investors 5% to 10%, according to CNBC. The prospectus names Charles Schwab, Fidelity, Robinhood, SoFi and E-Trade among the platforms making shares available to individuals. Small investors scrambled for allocations anyway, even as some told CNBC the valuation was, in a word, “stupid.”

Here is what most of those buyers probably did not do: read the prospectus. The Form 424B4 that SpaceX filed with the Securities and Exchange Commission (SEC) on June 12 describes a fundamentally different company than the one in the popular imagination. The rockets that made SpaceX famous produced about a fifth of its revenue last year. The company that actually listed is a satellite telecom funding an artificial intelligence (AI) buildout, with a launch business attached.

Three Companies in One Filing

SpaceX reports three segments, and the 2025 numbers tell the story cleanly.

Segment (FY2025)	Revenue	Operating income (loss)
Connectivity (Starlink)	$11,387M	$4,423M
Space (launch, Starship)	$4,086M	($657M)
AI (xAI: compute, Grok, X)	$3,201M	($6,355M)
Total revenue	$18,674M

Starlink pays the bills. The Connectivity segment grew revenue 49.8% year over year in 2025 and threw off $4.4 billion in operating income. As of March 31, 2026, roughly 9,600 Starlink satellites served approximately 10.3 million subscribers across 164 countries, territories and markets, and a separate constellation of about 650 satellite-to-mobile spacecraft served roughly 7.4 million monthly devices. If you want the backstory on that second constellation, see the earlier coverage of satellite-to-phone service killing mobile dead zones.

The rocket business is smaller than its reputation. Falcon launches rose from 134 in 2024 to 165 in 2025, yet the Space segment posted a $657 million operating loss for the year, dragged down by $3.0 billion in research and development spending on Starship. Launch remains a near monopoly in Western markets, but as a business it is a $4 billion line item inside an $18.7 billion company.

The AI segment is the cash furnace. In February 2026, SpaceX acquired xAI, which had itself absorbed X (the former Twitter) in March 2025. Because the deals were transactions between entities under common control, the filing retrospectively recasts SpaceX’s historical financials to include both. The combined AI segment, which the filing defines as “our AI compute, Grok, and X,” lost $6.4 billion at the operating line in 2025 on $3.2 billion of revenue.

That is why the bottom line looks the way it does. The recast company earned $791 million in 2024, then swung to a $4.94 billion net loss in 2025, and lost another $4.28 billion in the first quarter of 2026 alone. Accumulated deficit as of March 31, 2026: $41.3 billion.

Follow the Capex, Not the Branding

If you want to know what a company is becoming, ignore the logo and watch where the capital expenditure goes. In the first quarter of 2026, SpaceX spent $1.05 billion of capex on the Space segment and $1.33 billion on Connectivity. The AI segment spent $7.72 billion, more than three times the other two combined.

The use-of-proceeds section makes the priority explicit. SpaceX says it intends to use the roughly $74.4 billion of net proceeds “to fund our growth strategy, including the expansion of our AI compute infrastructure, enhancements to our launch infrastructure and launch vehicles, increases in the scale and capacity of our satellite constellations, and any remaining amounts for general corporate purposes.” AI compute is the first item on the list, ahead of the rockets.

The filing names the hardware behind that line: COLOSSUS, the company’s flagship data center in Memphis, Tennessee, and COLOSSUS II in Memphis and Southaven, Mississippi, which together it describes as a “coherent gigawatt-scale AI training cluster.” Beyond Earth, SpaceX says it expects to begin deploying orbital AI compute satellites, solar-powered data centers in orbit, “as early as 2028,” and pitches constellations of “potentially millions of satellites” for the purpose. Treat those as company projections, not facts. The physics and economics of computing in orbit are genuinely contested, and both the promise and the cooling constraints are covered in the deep dive on space-based AI data centers.

Buried in the risk factors is a glimpse of how far the ambitions run. The filing discloses an acquisition of spectrum assets and licenses from EchoStar for the Starlink Mobile push, a collaboration with Tesla and Intel on a project the filing calls Terafab, and a recent collaboration with Cursor along with “any potential acquisition of Cursor, if consummated.”

So the honest description of SPCX on debut day: a profitable satellite internet utility, welded to a break-even launch monopoly, both now financing an AI infrastructure bet of historic size.

The Math Behind a $1.77 Trillion Price

The valuation arithmetic is straightforward and uncomfortable. After the offering, SpaceX has 7,380,196,910 Class A shares and 5,695,668,265 Class B shares outstanding. At $135 per share:

$\$135 \times 13{,}075{,}865{,}175 \text{ shares} \approx \$1.77 \text{ trillion}$

Set that against the income statement:

$\frac{\$1{,}765 \text{ billion market cap}}{\$18.67 \text{ billion 2025 revenue}} \approx 95\times \text{ trailing revenue}$

Ninety-five times revenue, for a company that lost $4.9 billion last year. The filing itself is blunter than most coverage: SpaceX warns that it “may not achieve or, if achieved, sustain profitability in the future.” Rodney Comegys, chief investment officer for Vanguard Capital Management, put the retail risk plainly to NPR: “Investing in an IPO process can be highly speculative, and it’s really difficult to determine the path of an IPO on a given day.”

Professional skeptics went further. Morningstar’s analysts called the company “significantly overvalued” ahead of the debut, publishing a discounted cash flow valuation of $780 billion and warning that the xAI unit poses a “material threat of value destruction.” PitchBook’s analysis, for contrast, places fair value at $1.1 trillion to $1.7 trillion, a range whose top end still sits below the value implied by the $150 opening trade.

The bull case is real, and it deserves a fair statement: Starlink is compounding at nearly 50% a year with software-like operating margins in a market measured in hundreds of millions of households, launch is a structural monopoly, and no other company on Earth can manufacture and deploy satellites at SpaceX’s cost. The question $135 asks is not whether SpaceX is a great company. It is whether a great company is worth 95 times revenue while burning $4 billion a quarter.

One Share, One Vote Does Not Apply Here

Buyers of SPCX should be clear about what they purchased: economic exposure, not influence. After the offering, Musk holds approximately 82.4% of total voting power, overwhelmingly through super-voting Class B stock. That makes SpaceX a “controlled company” under Nasdaq rules, and the filing states it intends to rely on exemptions from certain corporate governance requirements that normally protect public shareholders.

The filing also carries a Tesla-style pay structure into the public market. Musk’s award entitles him to shares equal to 0.20% of fully diluted capitalization for each of 12 valuation milestones, running from $1.065 trillion to $6.565 trillion in $500 billion steps, conditioned on his continued employment. The first milestone was achieved before the xAI merger, settling 25,172,695 Class A shares on him. TechCrunch’s assessment of the debut pricing: the deal looks set to make Musk the world’s first trillionaire.

None of this is hidden. It is disclosed, in detail, in the same document retail buyers skipped. A $1.77 trillion public company where one person controls four out of every five votes, sits exempt from standard board independence rules, and is paid in valuation milestones is a governance experiment at a scale public markets have never run before.

From “Never Going Public” to Nasdaq Texas

For over a decade the company line was the opposite. In June 2013, Musk posted: “No near term plans to IPO @SpaceX. Only possible in very long term when Mars Colonial Transporter is flying regularly,” referring to the vehicle now called Starship. So when reports of a roughly $1.5 trillion listing surfaced in December 2025, the move read as a reversal; the site’s December preview of the SpaceX IPO walked through that earlier reporting. The filing explains what changed: the ambition got more expensive. Starship development consumed $3.0 billion in 2025, the AI segment burned $7.7 billion of capex in a single quarter, and private markets stop being deep enough at that scale.

The structure of the debut carries Musk’s fingerprints in smaller ways too. The shares list not just on Nasdaq but also on Nasdaq Texas, fitting for a company that reincorporated in Texas and runs its headquarters from Starbase, Texas. The share count, 555,555,555, is the kind of number you choose on purpose. And the fixed $135 price replaced the usual bookbuild theater with a take-it-or-leave-it offer that was oversubscribed anyway.

The debut itself was a lesson in expectations. Indications to trading desks started at $175, more than 16% above where the stock ended up opening, and the first trade crossed at $150, valuing SpaceX just under $2 trillion, the sixth-most valuable U.S. company, ahead of Meta and Tesla. Musk appeared from Starbase, Texas, while SpaceX executives rang the opening bell at the Nasdaq in New York.

What to Watch From Here

Four things will tell you whether this debut ages well.

The first earnings report. SpaceX has never had to explain a quarter to anyone. Public-market scrutiny of a quarterly net loss above $4 billion will be a new experience for a company that has operated as a private fortress since its 2002 incorporation.

Starlink’s growth curve. The valuation math above only works if Connectivity keeps compounding near its 2025 pace while holding margins. Any deceleration reprices everything beneath it.

Index inclusion. Morningstar flagged what it called an unprecedented path to inclusion in the Nasdaq 100 index just 15 trading days after the IPO, a mechanical bid from index funds that has nothing to do with fundamentals.

The orbital compute promise. The 2028 target for AI compute satellites is now a written, SEC-filed projection rather than a conference soundbite. If the physics or the economics fail, the central growth story of the filing fails with them.

The Bottom Line

The largest IPO in history closed its books with demand to spare, and the company it funded is genuinely extraordinary: a near-monopoly launch business, a constellation of roughly 9,600 satellites, and a gigawatt-scale AI training cluster under one ticker. But the prospectus, not the branding, is the contract. It describes a money-losing AI conglomerate controlled by one man, sustained by a brilliant satellite utility, asking public investors to fund the most expensive infrastructure bet ever written down in a single filing. The rocket on the logo is the smallest part of what traded under SPCX on June 12. Read the document before you buy the myth.