--- title: "From refusing to go public to the largest IPO in history: SpaceX's trillion-dollar gamble." type: "News" locale: "en" url: "https://longbridge.com/en/news/287228706.md" description: "SpaceX has officially filed for its IPO, marking a significant shift from Elon Musk's previous stance against going public. The company, now positioning itself as a 'space AI infrastructure company,' faces substantial financial challenges, including a $4.94 billion net loss in 2025. Starlink remains profitable, while the AI segment incurs heavy losses. The IPO could potentially value SpaceX at $2 trillion, making it one of the largest in history, driven by the need for funding to support its ambitious Mars colonization plans." datetime: "2026-05-21T13:16:08.000Z" locales: - [zh-CN](https://longbridge.com/zh-CN/news/287228706.md) - [en](https://longbridge.com/en/news/287228706.md) - [zh-HK](https://longbridge.com/zh-HK/news/287228706.md) --- # From refusing to go public to the largest IPO in history: SpaceX's trillion-dollar gamble. Shan Ouba, Golden Finance ## **Introduction** On May 20, 2026, in the evening of Eastern Time, the SEC website quietly released an S-1 filing that shook the world—SpaceX officially submitted its IPO application. It's only been a little over a year since Musk vowed "I will never go public before colonizing Mars." And the story he's telling this time is unprecedented: this isn't the listing of a rocket company, but the first public offering of a "space AI infrastructure company" in human history. ## **I. The Reversal of "Never Going Public"** SpaceX was founded with the mission of "making humanity a multi-planetary species." Musk has long refused to go public, arguing that the short-term profit pressure of the public market would erode the long-term vision of colonizing Mars. For the past two decades, SpaceX has only opened private placements to a limited number of institutions; Google, Scottish Mortgage Investment Trust, and Ark Capital are among its most well-known early investors. When Saudi Aramco set a record IPO of $29.4 billion in 2019, SpaceX showed no signs of going public. However, the situation changed dramatically in early 2026. The prospectus revealed that in the first quarter of 2026 alone, SpaceX burned through $9 billion in cash, with reserves plummeting from $15.852 billion to $11.385 billion. Musk estimated that the launch cost alone for colonizing Mars would be approximately one trillion dollars. Three factors contributed to this historic shift: first, the massive funding requirements of the Mars program itself; second, the acquisition of xAI in February 2026, bringing in an AI company with annual capital expenditures exceeding $30 billion, which the private equity market could no longer support; and third, Musk's strategic failure in the general-purpose large-scale model race. On February 2, 2026, xAI was officially integrated into SpaceX, becoming its AI business segment. On the same day, SpaceX reached a major agreement with Anthropic: leasing approximately 320,000 NVIDIA GPUs (including H100, GB200, and GB300) and a total computing power of approximately 560 megawatts from the Colossus 1 and Colossus 2 supercomputing clusters to Anthropic for $1.25 billion per month until May 2029, totaling approximately $45 billion. This contract is non-exclusive, and SpaceX retains the right to reallocate the computing power. This reveals the essence of Musk's strategic shift: from a competitor in the AI ​​model layer to a general contractor for space computing power. "Making money off those who build models" is the true starting point of SpaceX's "space AI" narrative. ## **II. Starlink is a money-printing machine, AI is a money-devouring beast** The prospectus reveals SpaceX's extremely unbalanced financial landscape: In 2025, consolidated revenue was $18.674 billion, a year-on-year increase of 33.2%, and adjusted EBITDA was $6.58 billion, but the net loss was as high as $4.94 billion. The loss further worsened to $4.29 billion in the first quarter of 2026, compared to only $528 million in the same period last year. The three major business segments play very different financial roles: Starlink (connectivity business): the only cash cow. In 2025, Starlink generated $11.387 billion in revenue, a 49.8% year-over-year increase, with operating profit of $4.423 billion and an EBITDA margin as high as 63%. As of March 2026, Starlink had 10.3 million global users, covering 164 countries, with 9,600 satellites in orbit. Space Launch and Starship Development: A strategic loss-making division. In 2025, it generated $4.086 billion in revenue and a loss of $657 million; in Q1 2026, revenue was only $619 million, and the loss widened to $662 million. The Starship project has accumulated over $15 billion in R&D investment. AI Business: A capital black hole. In 2025, Starlink's revenue was $3.2 billion, with an operating loss of $6.355 billion; in Q1 of 2026, the operating loss alone was $2.469 billion, with capital expenditures of $7.723 billion. Capital expenditures for the AI ​​segment in 2025 reached $12.727 billion, more than three times the total for the space and communications segments combined. 63% of Starlink's staggering profits were entirely used to fill the bottomless pit of Starship and AI. This extremely unbalanced financial structure is the core narrative tension of the prospectus. However, Starship's commercialization is imminent. The prospectus reveals that Starship has conducted 11 test flights, with a 12th test flight planned, and orbital payload delivery expected to begin in the second half of 2026. Its hundreds-ton-class low-Earth orbit payload capacity and fully reusable design will completely rewrite the on-orbit cost model and lay the foundation for on-orbit AI data centers. III. A $2 Trillion Valuation: What Makes It So? SpaceX's S-1 prospectus is a preliminary document; the specific offering price and fundraising scale have not yet been determined. However, based on the market's generally expected valuation range, SpaceX is poised to become one of the largest IPOs in history. Even at the lower end of the valuation range, the price-to-sales ratio is as high as 93 times, 13 times the average of the top 15 US technology companies. Wall Street is deeply divided on this. Valuation expert Damodaran gives a reasonable valuation of $1.22 trillion, 30% lower than the target price. He acknowledges that SpaceX has a stronger "option value" than Tesla, but explicitly states that he will not buy at a $2 trillion price. However, the enthusiasm of the capital market is not without basis: First, the Starlink business has seen certain growth, with users doubling in two years and profits surging by 120%; second, Anthropic's approximately $45 billion computing power contract has transformed space AI from a "blueprint" into "revenue"; and third, the prospectus depicts a total addressable market of $28.5 trillion, of which 93% comes from AI computing power demand. More importantly, there is FOMO – the fear of missing out on the next Tesla. Tesla's stock price has risen by over 2700% in the past decade, causing countless skeptics to miss out. Now, the same logic is being replicated with SpaceX, and the stakes are even higher. The prospectus mentions that a portion of the IPO quota has been reserved for retail investors, and the market expects extremely enthusiastic retail subscriptions, exceeding twice the net retail purchases of Tesla over the past year. JPMorgan Chase's chief strategist bluntly stated that this is a "bubble at the top of the market," but also admitted that "no one dares to really miss this IPO of the century." The prospectus reveals that SpaceX has completed 11 Starship test flights, and the 12th test flight is planned, which will debut the next-generation Starship and Super Heavy booster, powered by the next-generation Raptor engines, and launch from the new Starbase launch pad. The Super Heavy booster is powered by 33 Raptor engines, and Starship is designed to carry over 100 tons to low Earth orbit. The true value of Starship lies not in the numbers on paper, but in its validation of three core technologies: full reusability, orbital propellant transfer, and mass satellite deployment. This sends a clear signal to the capital market: Starship is not a PowerPoint presentation, but a real project about to be commercialized. The timing of the test flight plan and the IPO is no coincidence. If Starship delivers its orbital payload as scheduled, it will provide the most crucial "empirical support" for its trillion-dollar valuation, which is also the psychological premise for large institutions like BlackRock to invest billions of dollars. At the same time, Starship's performance will directly determine SpaceX's competitive position in NASA's Artemis lunar landing program. NASA has clearly stated that it will send humans back to the moon in 2028; this contract, worth hundreds of billions of dollars, is another major highlight for SpaceX's future growth. V. An Unfalsifiable Vision The most eye-catching part of the prospectus is not the financial data, but the governance structure. Under the dual-class share structure, Musk holds 42% of the equity but possesses 85.1% of the absolute voting rights. SpaceX is designated a "controlled company," exempting it from several Nasdaq governance requirements. For retail investors, what they are essentially buying is "a ticket without voting rights." "Investing in SpaceX is investing in Musk himself"—this subtext encapsulates the entire meaning of this governance structure. SpaceX's IPO was a pivotal moment in the history of commercial spaceflight. It wasn't the listing of a mature, profitable company, but rather the first public offering of a cash-burning "dream engine." Its finances were extremely unbalanced, its valuation relied on the still-early space AI narrative, and its success or failure rested entirely on the shoulders of one man. But if successful—if Starlink continues to be profitable, if Starship is commercialized as scheduled, if on-orbit AI data centers become a reality—this IPO will be remembered as the beginning of a new era. As the prospectus states, "Space is inherently a hostile environment," which perhaps best encapsulates this IPO: extremely risky, but once the threshold is crossed, a vast, untapped market lies ahead. When SPCX's ticker finally flashes on Nasdaq, the global capital markets will witness a rare moment: the largest IPO in history, from a company with an annual loss of $4.9 billion. It doesn't validate a profit model, but rather the pricing power of dreams. "You want to wake up in the morning with a vision of a bright future—and that's the essence of spacefaring civilization," Musk wrote in the prospectus. This time, he's given this belief the highest price tag in human history. The significance of SpaceX's IPO isn't just that the capital market accepted a company, but that the capital market began pricing in "orbital economics" for the first time. ### Related Stocks - [ITA.US](https://longbridge.com/en/quote/ITA.US.md) - [ROKT.US](https://longbridge.com/en/quote/ROKT.US.md) - [UFO.US](https://longbridge.com/en/quote/UFO.US.md) - [ARKX.US](https://longbridge.com/en/quote/ARKX.US.md) - [GOOGL.US](https://longbridge.com/en/quote/GOOGL.US.md) - [GOOG.US](https://longbridge.com/en/quote/GOOG.US.md) - [SMT.UK](https://longbridge.com/en/quote/SMT.UK.md) ## Related News & Research - [Mark your calendar -- the SpaceX IPO involves these 3 key dates](https://longbridge.com/en/news/287240248.md) - [SpaceX's investor pitch reads like a sci-fi manifesto](https://longbridge.com/en/news/287181510.md) - [SpaceX details how retail investors can buy into its IPO](https://longbridge.com/en/news/287161988.md) - [SpaceX's historic IPO filing is here. Here's what investors should know.](https://longbridge.com/en/news/287125288.md) - [Can you really buy SpaceX stock on Hyperliquid before the IPO?](https://longbridge.com/en/news/287213104.md)