Raising 11 Trillion Yen, the Birth of Humanity's First 'Trillionaire'—The Ambitions and Blind Spots of SpaceX's Largest IPO in History
On the morning of June 12 (US Eastern Time), shares of SpaceX (ticker: SPCX), which listed on the Nasdaq, opened at $150, 11% above the offering price of $135. During trading hours, the stock rose by over 30% at one point, with its market capitalization momentarily exceeding $2.25 trillion. The closing price was $160.95, up 19%, and the market capitalization surpassed $2 trillion (approximately 300 trillion yen at an exchange rate of 150 yen to the dollar). This means that a corporate value equivalent to half of Japan's nominal GDP was established in just one day of trading. The amount raised through the offering was approximately $75 billion (about 11 trillion yen). It is literally the "largest IPO in history," more than doubling the record set by Saudi Aramco in 2019 (approximately $29.4 billion). On this day, founder Elon Musk's estimated net worth reached approximately $1.1 trillion, marking the birth of the first "trillionaire" in human history. However, behind the frenzy, the financial situation revealed by the IPO prospectus is surprisingly raw. Why did a rocket company swallow an AI firm, and why did it enter the stock market now? This article organizes the details and blind spots of this massive IPO, as well as its impact on Japan.
This listing is significant not just because of its scale. First, as the main battlefield of the AI boom shifts from "model intelligence" to "securing computing resources and power," the first giant company to house space, telecommunications, and AI under one roof has appeared on the public market. Second, for the first time, individual investors around the world—including those with NISA accounts in Japan—can invest in "Mr. Musk's private empire," which was previously accessible only to venture capitalists and a select group of wealthy individuals. The structural changes in the tech industry and the story of individual asset formation are intersecting in this one company.
The "Largest in History" by the Numbers—What Makes It So Extraordinary
In this IPO, SpaceX sold 555.6 million shares at a fixed price of $135 per share. It is easy to understand how extraordinary the $75 billion raised is when compared to past records. The previous record was held by Saudi Arabia's state-owned oil company, Saudi Aramco (2019, approximately $29.4 billion), followed by China's Alibaba (2014, approximately $25 billion). SpaceX raised about 2.5 times that of Aramco in a single night. For Japanese readers, it is reminiscent of NTT, which listed in 1987 and became the world's largest company by market capitalization at the time, but the scale far exceeds that. As a capital raise for a single company, it is now at a level comparable to the national budget of a mid-sized nation.
The sales process was also full of anomalies. The roadshow for institutional investors, which began in early June, was attended by about 125 analysts from 21 financial institutions, and a separate briefing for individual investors was held for about 1,500 people. According to CNBC, the price was not determined by the usual book-building method that fluctuates based on demand, but was pushed through at a "fixed price" of $135. It was a seller-advantaged structure where they did not seek to gauge demand to set a price, but rather said, "Those who want to buy at this price can buy it." It was a hectic schedule, with the final price determined after the close of trading on June 11 and the listing on the 12th. Note that the result of an opening price of $150 and a closing price of around $161 means that the fixed price was about 20% lower than the market's valuation. One could say they "left about $14 billion on the table" in simple terms, but in the face of the largest capital raise in history, that is likely within the margin of error.
What cannot be overlooked is that Japanese individual investors were "participants" from day one. Three companies—SBI Securities, Rakuten Securities, and Mizuho Securities—handled applications from within Japan, and it was also eligible for the growth investment quota of the new NISA. SBI Securities began accepting applications on June 5 and announced the lottery results on the 11th. There are almost no precedents for Japanese household tax-exempt money participating in a US mega-IPO from the public offering stage.
When Did the Rocket Company Become an "AI Company"?
The key to understanding this listing is the absorption merger of the AI company xAI, which was completed on February 2, 2026. SpaceX acquired xAI through an all-stock swap, creating a massive conglomerate with a total value of approximately $1.25 trillion—SpaceX at about $1 trillion and xAI at about $250 billion. Since both companies were under Mr. Musk's control, the negotiations, which would normally take years, were completed with unusual speed. Mr. Musk has indicated a policy of dissolving xAI as an independent company and reorganizing its AI product lineup under the "SpaceXAI" brand.
Let's turn back the clock a little. Founded in 2002, SpaceX was a "cliff-edge venture" that nearly ran out of funds in 2008 and only succeeded in reaching orbit on its fourth launch. Since then, it has dramatically lowered launch costs through the idea of landing and reusing the first stage of its rockets, and in recent years, it has handled more than half of the world's orbital launches on its own. In 2019, it began building the low-Earth orbit satellite network "Starlink," and has grown to provide broadband to the world with a fleet of about 8,000 satellites. Meanwhile, xAI is an AI company founded by Mr. Musk in July 2023, which developed the conversational AI "Grok." It surprised the industry by building the "Colossus" supercomputer in Memphis, Tennessee, which brought 100,000 GPUs online in less than four months. In March 2025, it also brought the former Twitter, X, under its umbrella through a stock swap.
Why did a rocket company swallow an AI firm? The outward explanation is "complementarity." According to the prospectus, the company plans to expand Grok to a "multi-trillion parameter" scale. Parameters are an indicator of an AI model's scale, and their expansion requires vast computing resources and power. SpaceX has the stable revenue generated by Starlink and an ambitious exit strategy known as "in-orbit data centers," which will be discussed later. The communication data collected by the satellite network, the public discourse space held by X, the brain that is Grok, and the transportation means that is the rocket. In Mr. Musk's mind, these were likely one picture from the beginning.
However, another realistic reason was cash flow. As revealed by TechCrunch from the prospectus, xAI burned through $6.4 billion in cash in 2025. With the expansion of Colossus, xAI became one of Nvidia's largest customers, but as the competition for GPU procurement intensified, survival on its own was not easy. Covering the AI deficit with Starlink's profits and selling the combined "story" to the market—this integration and listing have that structure.
The Kitchen Table Talk Revealed by the S-1—2.8 Trillion Yen in Revenue, Yet Still in the Red
The IPO prospectus (S-1) released on May 20 brought the company's finances, which had been veiled for many years, into the light for the first time. It is the "household account book" of a company that has been running as a private entity for nearly a quarter of a century.
Consolidated revenue for 2025 was $18.7 billion (approximately 2.8 trillion yen at an exchange rate of 150 yen to the dollar, a 33% increase from the previous year). Looking at the breakdown, the "connectivity" division, centered on Starlink, accounted for $11.4 billion, or about 61% of the total, and is the only profitable division, earning an operating profit of $4.4 billion. The "space" division, which includes rocket launches that were the core business since its founding, remained at a scale of about $4 billion. In other words, in terms of revenue composition, SpaceX is already not a "rocket company" but a "satellite communications company," and launch and transport contracts for NASA and the US Department of Defense have shifted from being the backbone to being a source of credibility. And the "AI" division, derived from xAI, recorded an operating loss of $6.35 billion against $3.2 billion in revenue (as mentioned above, xAI has X under its umbrella, and it is believed that advertising revenue from X also contributes to the division's revenue). As a result, the company as a whole had a net loss of $4.9 billion. A company with 2.8 trillion yen in revenue went public with a deficit of over 700 billion yen instead of a profit—this is the true face of the numbers.
The expansion of investment is even more intense. Capital expenditure doubled from $5.6 billion in 2024 to $12.7 billion in 2025. In the January-March quarter of 2026 alone, it invested $10 billion, with a breakdown of $1 billion for space, $1.3 billion for connectivity, and $7.7 billion for AI; looking at the investment amount alone, it is now an "AI infrastructure company." Revenue for the same quarter was $4.69 billion, and the net loss was $4.3 billion. While Starlink subscribers nearly doubled from the previous year to 10.3 million, revenue per user (ARPU) is on a downward trend due to expansion into emerging countries and the expansion of low-price plans. Growth is real, but cash is going out at a pace that exceeds that growth. The 11 trillion yen from the IPO is also fuel to support this "burn rate."
Lending 300 Megawatts to a Rival—A Strange Alliance with Anthropic
The S-1 also contained a contract that surprised the market. It states that Anthropic (the developer of the conversational AI "Claude"), which competes in AI development, will lease almost the entire capacity of Colossus 1—about 220,000 Nvidia GPUs and 300 megawatts of power—for $1.25 billion per month until May 2029. 300 megawatts is enough power for hundreds of thousands of average households, and the total contract value exceeds $40 billion (approximately 6 trillion yen). It is structured as a "reservation-type" payment for capacity regardless of actual usage, and for SpaceX, it becomes a stable income of about $3.7 billion rolling in every quarter.
This contract was announced by both companies in early May 2026, and details such as the amount were revealed with the release of the S-1 on May 20. It is likely no coincidence that the timing of the announcement was just before the IPO. Showing that "the AI division has guaranteed revenue" is directly linked to the pricing of the listing. Regarding the strange alliance of lending its prized assets to a rival, Mr. Musk joked, "No one triggered my 'evil detector'" (Tom's Hardware). Anthropic also announced the contract on its company blog, explaining that it would use the secured computing resources to expand usage of Claude. It was also reported that the two companies are considering the joint development of gigawatt-class "in-orbit AI computing power." A competitor's latest AI model being trained on one's own equipment—a lender willing to fill capacity even if it means allowing such a development, and a borrower wanting to secure power and computing resources from anyone. It is a situation where the interests of both parties aligned.
In the background is a scramble for computing resources that is engulfing the entire AI industry. OpenAI has launched a $500 billion data center plan called "Stargate," and Microsoft and Meta are also continuing to invest tens of billions of dollars annually. With power, land, and GPUs all in short supply, "how much power can be secured" is becoming the core of the competition rather than "whose chips are used for training." Anthropic is known for its multi-vendor strategy that also uses AI chips from Amazon and Google, but this time it even borrowed equipment from its arch-rival, Mr. Musk. The boundary between friend and foe is surprisingly fluid in the face of computing resources. And for SpaceX, this contract is both a way to cover the huge deficit in the AI division and a proof-of-concept for a new business model: "lender of AI infrastructure (the AI version of real estate)."
Is the 1-million-unit "Space Data Center" concept serious?
The company lists the use of the approximately $75 billion raised as: expanding ground data centers, completing the development of the super-heavy 'Starship' rocket, building the 'Moonbase Alpha' lunar base, conducting unmanned and manned missions to Mars, and deploying orbital AI data centers. These are items that would be dismissed as jokes in the prospectus of an ordinary company.
Among these, what most stimulates the imagination of investors is the concept of an orbital data center consisting of up to one million satellites. By generating solar power in space, performing AI calculations in orbit, and sending only the results back to Earth, the company aims to solve the energy crisis. Musk also unveiled the design for the prototype satellite 'AI1'. With the International Energy Agency (IEA) estimating that data center power consumption will double by 2030, terrestrial AI infrastructure faces a triple constraint of power, cooling, and land. In space, sunlight is uninterrupted 24 hours a day, cooling can be done via radiation, and there are no local residents to protest land acquisition—at least in theory.
In reality, there are many hurdles. These include semiconductor malfunctions and degradation caused by cosmic rays (radiation), the difficulty of repairing or replacing equipment in orbit, the increased risk of space debris from a massive number of satellites, the impact on astronomical observations caused by satellites crossing the night sky, and the environmental impact of satellites re-entering the atmosphere at the end of their lives. There are countless unresolved technical and social issues. Concerns have already been raised by astronomers and environmental researchers regarding the existing Starlink constellation of about 8,000 satellites, and 'one million' is a world two orders of magnitude larger. Above all, the success of this concept depends entirely on the flight frequency of Starship. According to the scientific journal Scientific American, Starship is expected to carry its first practical payload into orbit in late 2026, and the next-generation Starlink and orbital data centers are predicated on this vehicle flying stably and inexpensively. If the barriers of launch cost and frequency cannot be overcome, the one-million-satellite concept will remain a dream on paper. A significant portion of the $2 trillion market capitalization is being bet on rockets that have not yet flown and data centers that do not yet exist.
'One million people on Mars' as a compensation condition—The structure of Musk's control
The governance design is also unprecedented. SpaceX has adopted a dual-class structure using Class B shares with 10 times the voting rights, allowing Musk to retain approximately 85% of the voting power even after the IPO. It is virtually impossible for general shareholders to overturn Musk's management decisions, even if they act in unison.
Even more eye-catching is the compensation package granted in conjunction with the IPO. Musk will receive 1 billion Class B shares, but the vesting conditions are 'achieving a corporate value of $7.5 trillion' and 'building a permanent colony with 1 million residents on Mars.' While these conditions sound like science fiction, as TechCrunch points out, Musk can exercise voting rights with these shares before they vest and can also use them as collateral for loans. In other words, the design allows him to obtain control and fundraising power first, even if no one is living on Mars yet. A compensation package worth up to $1 trillion was approved by shareholders at Tesla in 2025, and the method of 'giving away a huge equity stake in exchange for grand goals' is becoming a staple of Musk-led companies.
There are also question marks regarding the independence of the board of directors. It was noted before the IPO that three of the four directors had close personal and economic ties to Musk, and how to monitor transactions and conflicts of interest with the group of companies Musk is involved in—such as Tesla, X, and Neuralink—remains an unresolved issue after the listing. As long as Musk serves as CEO and de facto controller of multiple companies, 'where his time and attention are directed' will remain the biggest variable for shareholders.
In other words, what general shareholders of SPCX are buying is not 'a voice in management,' but 'a ticket to ride along with Musk's grand plan.' Governance structures that allow founders to run unchecked have been seen at US tech companies like Google and Meta, but the scale and extremity of 85% voting power and a 'Mars colonization' compensation condition are unprecedented. Whether one views this ticket as attractive or dangerous is the dividing line in the evaluation of this IPO.
Bullishness and skepticism—'Twice the value of Walmart, with sales less than Macy's'
Contrary to the surge on the first day, professional evaluations are sharply divided. The US research firm Morningstar calculated a fair value of $780 billion based on the Discounted Cash Flow (DCF) method. This is less than half of the $1.75 trillion valuation based on the offering price, and the firm warns that 'uncertainty is extremely high.'
The abnormality of the numbers is easy to see when compared. The offering price is about 107 times sales (according to US media estimates). Since the company is not profitable, the P/E ratio cannot even be calculated. While it has twice the corporate value of the major US retailer Walmart, its sales do not even reach those of the department store Macy's—US media used such comparisons to describe the frenzy. Analyst Ed Elson panned the S-1 as 'substance-less, hallucinatory, and bordering on dishonest,' and a US labor union representing about 2 million members submitted a warning letter to securities regulators calling it an 'overvalued, high-risk bet.' There is concern that the retirement funds of ordinary workers will be caught up in this bet through inclusion in pension funds and the like.
On the other hand, the arguments of the bulls are also clear: (1) overwhelming share in the rocket launch market and the superiority of reusable technology, (2) Starlink, which continues to grow at a doubling pace even after exceeding 10 million subscribers, and (3) a unique vertical integration that bundles AI, space, and telecommunications into one company. The bulls also see a next growth story for Starlink called 'Direct to Cell,' where off-the-shelf smartphones connect directly to satellites even in areas without base stations, and they believe that if the sale of computing resources, such as the Anthropic contract, expands, the losses in the AI division will shrink.
To add, special supply and demand circumstances may have also boosted the first-day surge. The ratio of shares available in the market (float) relative to the total shares issued is limited, and buying by individual investors concentrated on the stock with the highest name recognition in history. If inclusion in major stock indices is realized in the future, passive investment funds linked to the indices will flow in mechanically, while a large amount of selling pressure will await after the lock-up (restrictions on existing shareholders selling) expires. What will ultimately decide which side is correct is not debate, but the performance of Starship and the AI business over the next two to three years. During the dot-com era of the late 1990s, many stocks bought on 'stories' disappeared, but the handful that survived became companies that defined the era. The market will now grade SpaceX every quarter to see which side it falls on.
What it means for Japan
First, Japanese individual money has been directly incorporated into this 'bet.' As mentioned above, it was possible to participate in the public offering through SBI, Rakuten, and Mizuho, and it also became eligible for the growth investment quota (2.4 million yen per year) of the new NISA. It is groundbreaking for a massive US IPO to be opened to Japanese individuals from the public offering stage, and similar schemes may spread in the future. However, calmness is also required. It is a stock for which Morningstar estimates a fair value at less than half the offering price, and jumping in based solely on headlines like 'largest in history' and '19% gain on the first day' is dangerous. Since losses in a NISA account cannot be offset against other profits (tax loss harvesting), the significance of holding volatile individual stocks in a tax-exempt quota is a topic that should be considered carefully. It is also necessary to factor in that it is a stock whose price fluctuates significantly based on the words and actions of one individual, Musk.
Second, Starlink as a communication infrastructure is rapidly becoming 'the norm' in Japan. KDDI's 'au Starlink Direct' has expanded its compatible models to 50 as a service that connects smartphones directly to about 650 satellites, and on March 4, 2026, it launched the world's first international roaming for direct satellite communication in the US. From June, it will expand its coverage to Canada, the Philippines, and New Zealand. NTT Docomo has also announced the start of its direct satellite communication service for the beginning of fiscal year 2026, and communication in mountainous areas, remote islands, and during disasters such as earthquakes is shifting to a design based on satellites. This is a matter of convenience, but at the same time, it means that part of Japan's core communication infrastructure depends on a single overseas company—and a listed company where one individual holds 85% of the voting rights. In the war in Ukraine, there were actual instances where the availability of Starlink connections influenced the war situation and operations. Because satellite communication can be a lifeline for Japan, a disaster-prone country, the design of dependency and the securing of alternatives should be discussed head-on as an issue of economic security.
Third, it is necessary to face the difference in financial power. The Japanese government launched a 1 trillion yen support program over 10 years as a space strategy fund, but SpaceX raised 11 trillion yen in one night. We are in an era where one company raises 11 times the amount of a national 10-year plan in a single day. While the challenges of Japanese players continue, such as Mitsubishi Heavy Industries and JAXA with the H3 rocket, ispace challenging lunar transport, and Interstellar Technologies funded by the Toyota Group, a strategy of competing on 'volume' on the same playing field is not realistic. Rather, a realistic path is to earn money 'on top of' and in the 'gaps of' giant platforms. There are many areas where Japanese companies can leverage their accumulated technology, such as breaking into supply chains for satellite parts and ground station equipment, disaster prevention, agriculture, and logistics services using satellite data, and orbital services (satellite inspection, repair, and debris removal). Japanese startups are globally ahead in debris countermeasures, and a world where satellites increase by orders of magnitude is, from another perspective, a tailwind for Japan's 'cleaner and maintenance' business. Furthermore, if the orbital data center concept becomes realistic in the long term, it will not be someone else's problem for Japan's AI infrastructure strategy, which is struggling to secure power for data centers.
And what must not be forgotten is the context of security. Launch capability and satellite networks are no longer purely civilian infrastructure. The US has a 'private company with national-level capabilities' in SpaceX, and China is rushing to build its own satellite constellation as a nation. How much Japan maintains and strengthens its own launch means and satellite networks is a matter of industrial policy as well as a matter of diplomatic and defense policy.
Summary—3 points to watch next
This IPO is a grand experiment that bundles two capital-intensive industries, space and AI, under the control of one individual and puts their entire future to the public market. I would like to conclude by listing three points to watch for the time being. First, will the orbital insertion of a practical payload by Starship, scheduled for late 2026, be realized as planned? If this is delayed, the foundation of the growth story, from next-generation Starlink to orbital data centers, will be shaken. Second, when and with what specifications will the prototype satellite 'AI1' be launched? This will be the first yardstick to measure the 'seriousness' of orbital AI computing. Third, how much will the profit and loss of the AI division, disclosed in quarterly financial results, improve through external sales revenue such as the Anthropic contract?
What must not be forgotten is that 24 years ago, this company was on the brink of bankruptcy, with the mantra 'if the fourth launch fails, we go under.' It is precisely because they achieved the feat of reusable rockets—once laughed at as absurd—that the market has even put a price on 'Mars' and '1 million satellites.' The 19% rise in the stock price on the first day shows that the market has, for now, bought into that 'story.' But from here on out, it is the turn of quarterly earnings to verify that story with numbers. Whether the birth of humanity's first trillionaire was a turning point in history or the peak of overheating—the answer is already beginning to unfold.
*This article is not investment advice. The figures stated are based on reports and public documents as of June 13, 2026. Yen conversions are standardized at $1 = 150 yen for convenience.
Sources
NPR - SpaceX IPO makes history as largest ever. Stock gains 19% on first day
CNBC - SpaceX stock jumps 19%, closing near $161 after record IPO
CNBC - SpaceX targets fixed $135 IPO price for roadshow
CNN Business - SpaceX's record market debut makes Musk a trillionaire
TechCrunch - xAI burned $6.4B last year
TechCrunch - How Elon Musk will increase his power through the SpaceX IPO
TechCrunch - Anthropic will pay xAI $1.25B per month for compute
Anthropic - Higher usage limits for Claude and a compute deal with SpaceX
Tom's Hardware - SpaceX rents 220,000 GPUs / 300MW to Anthropic
Hargreaves Lansdown - Inside SpaceX's IPO filing
Morningstar - 6 Charts on SpaceX's Pre-IPO Financials
CNBC - SpaceX is worth less than half of its $1.75 trillion IPO target, Morningstar says
Scientific American - SpaceX IPO valuation depends on Starship and orbital AI data centers
Space.com - SpaceX goes public with a historic IPO
Yahoo Finance - Analysts shred SpaceX's IPO filing
Al Jazeera - Musk's $1.8 trillion SpaceX IPO could be 'highly undesirable' for some
Variety - Elon Musk Becomes First Trillionaire With SpaceX IPO
KDDI News Room - au Starlink Direct to Expand International Satellite Roaming
Light Reading - NTT Docomo to roll out direct-to-cell service in early FY2026
FinTech Journal - SpaceX IPO to be handled by domestic Japanese securities firms, also eligible for new NISA
