SpaceX Just Filed the Largest IPO in History. The Filing Itself Says the AI Growth Story May Never Work.
- $1.75TSpaceX target IPO valuation — would eclipse Saudi Aramco's $1.7T 2019 record as the largest IPO in history.
- $18.7BSpaceX 2025 revenue (33% YoY growth) — per the S-1 prospectus filed May 20, 2026.
- ($4.94B)SpaceX 2025 NET LOSS. Cumulative losses since inception: $37+ billion. Q1 2026 net loss: $4.27B (vs. $528M Q1 2025).
- 85.1%Elon Musk's voting power post-IPO via 93.6% of Class B super-voting stock — Class A retail buyers get a financial claim with essentially zero governance leverage.
- $40B+Anthropic compute commitment to xAI/SpaceX through May 2029 ($1.25B/month) — the AI-growth-pitch revenue line.
- $24.4BCumulative SpaceX federal contracts since 2008. 2024 unclassified government revenue: $3.3B. PLEO Starshield ceiling: $13B. Government-customer concentration risk now disclosed.
After the closing bell on Wednesday, May 20, 2026, Space Exploration Technologies Corp. (CIK 0001181412) filed a Form S-1 registration statement with the U.S. Securities and Exchange Commission. The ticker is SPCX. The exchange is Nasdaq. The target valuation is $1.75 trillion. If the deal prices at the target, SpaceX will surpass Saudi Aramco's 2019 listing ($1.7 trillion) as the largest IPO in the history of public markets.
The S-1 also discloses that SpaceX lost $4.94 billion in 2025 on $18.7 billion in revenue. Cumulative losses since the company's founding in 2002 now exceed $37 billion. Q1 2026 was worse: a net loss of $4.27 billion against $4.69 billion in revenue, versus a net loss of $528 million in Q1 2025. The company's business is growing, fast, and burning cash, faster. And the growth story being sold to the roughly 1,500 retail investors invited to the June 11 pricing event is built around a product line — orbital data centers training AI models in space — that the prospectus' own risk factors describe as “in early stages, involv[ing] significant technical complexity and unproven technologies” that “may not achieve commercial viability.”
The S-1 is the receipt. The story is the gap between what Wall Street is pricing — certainty, scale, the Musk halo — and what the prospectus itself discloses: dual-class voting control that locks 85.1 percent of decision-making with Elon Musk, government-customer concentration risk, an AI-orbital-compute pitch the company itself disclaims, and a Trump administration that publiclythreatened in June 2025 to terminate the federal contracts that produce roughly one-fifth of SpaceX's revenue. The Nasdaq roadshow launches June 4–8. Pricing is targeted for June 11. First trading day: June 12, 2026.
SpaceX organizes its business into three segments. The S-1 discloses each separately for the first time:
- Connectivity (Starlink): $11.387 billion 2025 revenue, $4.423 billion operating income, $7.168 billion adjusted EBITDA. 10.3 million subscribers, 9,600+ satellites on orbit. The profitable engine. Funds everything else.
- Space (Launch / Falcon / Starship): $4.086 billion 2025 revenue, $657 million operating loss, $653 million adjusted EBITDA. Starship R&D alone burned $930 million in Q1 2026, against $15+ billion cumulative dev spend.
- SpaceXAI: $818 million Q1 2026 revenue, $6.355 billion 2025 operating loss. xAI alone burns approximately $14 billion a year. The segment that's being sold to retail as the future.
Total 2025 capex: $20.74 billion. Of that, 61 percent— roughly $12.7 billion — went to AI compute and the orbital-data-center buildout. SpaceX spent more capex than revenue in 2025. The S-1 also discloses, for the first time, an 8,285 BTC Bitcoin treasury position (~$637 million at filing-week prices) held in Coinbase Prime custody since June 2022, and a $530 million expected legal costs reserve.
Dual-class share structures are not new. Meta, Alphabet, Snap, and Palantir all went public with founder super-voting stock. What is unusual about the SpaceX S-1 is the magnitude of the concentration:
Elon Musk: 93.6 percent of Class B super-voting stock (10 votes per share) plus 12.3 percent of Class A. Total post-IPO voting power: 85.1 percent. Economic stake: ~42 percent.
Gwynne Shotwell (President & COO): 7.1 million Class B shares; fifth-largest Class A holder per S-1.
Outside Class A buyers (retail + institutions): get a financial claim with effectively zero governance leverage. The S-1 verbatim: “Mr. Musk will have the power to control the outcome of matters requiring shareholder approval, including election of all our directors.”
On May 19, the day before the public filing, the NYC Comptroller (Mark Levine (D-NY)), the NYS Comptroller (Thomas DiNapoli (D-NY)), and CalPERS CEO Marcie Frost jointly sent a public letter to SpaceX asking the company to abandon dual-class structure before the offering. The letter was signed by managers of more than $700 billion in public-pension assets. SpaceX did not respond before filing the S-1 with the dual-class structure intact.
The growth story being sold to retail at the June 11 pricing event is SpaceXAI. The headline disclosure in the prospectus: on May 6, 2026, Anthropic committed to $1.25 billion per monthin compute payments to xAI/SpaceX through May 2029 — total potential value over $40 billion. Under the deal, Anthropic gets all of its compute at Colossus 1 (the Memphis, Tennessee data center hosting 220,000+ NVIDIA H100/H200/GB200 GPUs across 300+ megawatts of power) and also agreed to “consider” using future SpaceX orbitaldata center capacity once that's built.
That last part — the orbital piece — is the unprecedented swing of this offering. The pitch is that solar irradiance, radiative cooling to vacuum, and continuous uplink/downlink via Starlink make space data centers structurally cheaper than terrestrial. The risk factors in the prospectus disclose, in the company's own legal language, that the technology stack to make that work doesn't exist yet. Multiple risk-factor paragraphs read like investor warnings:
“Our space-based AI compute initiatives are in early stages, involve significant technical complexity and unproven technologies, and may not achieve commercial viability.”
SpaceX S-1 Registration Statement · Risk Factors · May 20, 2026
Jeff Bezos — who runs Blue Origin and Amazon's Project Kuiper, both direct SpaceX competitors — sat for an exclusive interview with CNBC's Andrew Ross Sorkin from Blue Origin's Rocket Factory in Merritt Island, Florida on the morning of the SpaceX filing. Asked about orbital data centers and the 2-3 year timeline being marketed to investors, Bezos answered:
“Space is going to be a gigantic industry... that timeline is a little ambitious.”
Jeff Bezos, CNBC Squawk Box, May 20, 2026, Merritt Island, FL
SpaceX's revenue isn't evenly distributed. The S-1 discloses cumulative federal contracts of $24.4 billion since 2008. 2024 unclassified government revenue alone was $3.3 billion. The company holds approximately 97 percentof awarded task orders under the Space Development Agency's Proliferated LEO program, whose ceiling has grown from $900 million to $13 billion. The National Reconnaissance Office's classified Starshield contract (the Pentagon's government version of Starlink) is worth $1.8 billion. The 2026 SDA Tranche 2 award added $739 million in January.
That concentration is the structural risk disclosed in the prospectus. It is also the risk President Donald Trump (R) personally and publiclythreatened to exercise during the June 2025 Trump-Musk feud — the period after Musk's 130-day DOGE tenure ended in late May 2025 and before the September 2025 public reconciliation between Trump and Musk. Trump's Truth Social posts from that window are the documentary receipt:
Elon may get more subsidy than any human being in history, by far, and without subsidies, Elon would probably have to close up shop and head back home to South Africa.
Paraphrased commentary · not a verbatim post
Verbatim quote text cross-referenced via Reuters, AP, NBC News, and Fox News coverage of the June 2025 Trump-Musk feud. PostId not embedded; the substance is corroborated and is now a documented government-customer-concentration risk in the SpaceX S-1.
The easiest way to save money in our Budget, Billions and Billions of Dollars, is to terminate Elon's Governmental Subsidies and Contracts. I was always surprised that Biden didn't do it!
Paraphrased commentary · not a verbatim post
Substance cross-referenced via national news coverage of the Trump-Musk feud. This is the explicit federal-contract-termination threat the S-1 risk factors now generically reference under ‘government customer concentration.’
Trump and Musk publicly reconciled in September 2025. The reconciliation does not remove the underlying risk: a single executive-branch decision could unilaterally reduce SpaceX's revenue by approximately the $3.3 billion 2024 unclassified-government line, and place additional billions of Starshield, NSSL, and SDA contract ceiling under review. The S-1 risk factors disclose this in anodyne language. The actual political history is on Truth Social.
The underwriting syndicate disclosed in the S-1 is the largest IPO bank lineup in history: 21 institutions. Lead-left is Goldman Sachs. Other US bookrunners: Morgan Stanley, Bank of America, Citigroup, JPMorgan Chase. European book: Deutsche Bank, UBS. UK book: Barclays. The roadshow launches the week of June 4 and culminates in pricing on June 11, with first day of trading on June 12 under the SPCX ticker on Nasdaq.
What's structurally new about this offering is the retail allocation. CFO Bret Johnsen told the underwriters that up to 30 percentof the float will be allocated to retail buyers — an unprecedented retail tilt for a mega-IPO. The company is hosting roughly 1,500 retail investors at a June 11 in-person pricing event. Johnsen, verbatim:
“Retail is going to be a critical part of this and a bigger part than any IPO in history.”
Bret Johnsen, CFO, SpaceX · per Reuters and CNBC reporting of the IPO pitch to banks
The retail tilt is editorially significant because the institutional buyers in this offering have access to the same S-1 risk factors as retail does, but unlike retail, they have the trading liquidity and post-IPO M&A optionality to exit quickly if the orbital-data-center pitch falls apart. Retail buying at the June 11 price is, by structure, the longest-hold cohort. Bloomberg's May 18 markets-desk framing called the $1.75T number “a price tag on the fear of missing out.” The 1,500 in the room are who pay it.
Musk has not been quiet about the IPO. His X account in the weeks leading up to filing carried the maximalist version of the pitch:
If SpaceX succeeds in this absurdly difficult goal [making life multiplanetary], it will be worth many orders of magnitude more than the economy of Earth.
I would be there in person, but we gotta get this @SpaceX IPO going pretty soon.
On the AI growth pitch: “Our space-based AI compute initiatives are in early stages, involve significant technical complexity and unproven technologies, and may not achieve commercial viability.”
On Musk: “Mr. Musk will have the power to control the outcome of matters requiring shareholder approval, including election of all our directors.”
On the government concentration: standard language disclosing federal-contract dependence and the risk of contract termination or non-renewal — the substantive risk now memorialized by Trump's June 2025 Truth Social posts and the documented willingness of the executive branch to deploy that lever as a political tool.
The S-1 is asking $1.75 trillion in market capitalization for a business whose own legal counsel discloses, in writing, that the AI growth thesis may never work and the voting structure makes the buyer of Class A stock a passive holder of someone else's decisions.
Three reasons SpaceX's S-1 lands as a stop-the-presses accountability story rather than a financial-press novelty.
- The retail tilt makes this a consumer-protection story, not a Wall Street story. 1,500 individuals at a June 11 event, buying into a $1.75 trillion valuation backed by a $37 billion cumulative loss and an AI growth line that the prospectus itself disclaims.
- The government-concentration risk is documented executive-branch behavior, not theoretical. Trump already exercised the threat verbally in June 2025; the S-1 risk factors now have to disclose the structural exposure.
- The dual-class governance gap is the cleanest accountability question on the page.Three of America's largest public-pension funds publicly asked SpaceX to drop the structure before filing. SpaceX filed anyway. The Class A shares retail will buy are, by design, non-voting in any matter that matters.
$1.75 trillion target. $18.7 billion 2025 revenue. $4.94 billion 2025 net loss. $37+ billion cumulative losses since founding.The orbital AI compute pitch is disclaimed in the company's own risk factors. 85.1 percent of voting power concentrated with the founder. $24.4 billion in cumulative federal contracts subject to a presidential lever Trump has already threatened to pull. 1,500 retail investors at the June 11 pricing event.
SpaceX is, on the publicly available record, the most ambitious private space company in history. It is also the one asking the largest IPO check ever written, on a balance sheet that books losses faster than revenue, with a growth story its own SEC filing labels speculative, under a governance structure that gives outside buyers no meaningful vote.
Wall Street is pricing certainty. The prospectus is disclosing uncertainty. Retail is the bag.