On June 12, SpaceX issued 555.6 million shares at $135 per share, raising $75 billion. This more than doubled Saudi Aramco's record IPO of $29.4 billion set in 2019, locking in the company's overall valuation at $1.77 trillion.

One figure immediately came to mind: founder Elon Musk's net worth surpassed $1 trillion. For the first time in human history, a trillionaire has emerged.

However, the S-1 filing also revealed another set of figures: the company's net loss in 2025 was $4.9 billion, with accumulated losses exceeding $37 billion; the xAI division was losing $6.4 billion while simultaneously collecting $26 billion annually in computing power rental fees from Anthropic and Google. With 4.2% of the stock actually circulating on the secondary market, how much will every dollar of net buying drive price increases? When will those holding vesting shares, including the 4,400 employees about to become millionaires, finally give in?

This is a math problem.

How is a trillion-dollar fortune "calculated"?

Musk became the first trillionaire, primarily driven by SpaceX's $1.77 trillion valuation. However, the exact composition of this "trillion" figure depends on a question not directly answered in the S-1 filing: what is his actual economic interest?

TechCrunch, citing S-1 filings, revealed that Musk holds approximately 85.1% of the voting rights, and will retain over 50% after the IPO. Voting rights do not equate to economic equity. Musk achieves control by holding Class B or Class C shares with super voting rights, typically 10 to 20 times that of common stock, while the actual economic equity percentage may be far lower than 85.1%.

Since Musk's exact economic stake in SpaceX is not disclosed in public reports in the S-1 filing, we can only make hypothetical assumptions. Assuming his economic stake is between 35% and 55%, the individual equity value corresponding to SpaceX's $1.77 trillion valuation is as follows:

If the economic stake is 35%, SpaceX's equity value is approximately $619.5 billion. Adding its approximately 13% stake in Tesla (equivalent to hundreds of billions of dollars based on Tesla's current market capitalization) and other unlisted assets (xAI, Neuralink, The Boring Company), its total net worth easily surpasses the $1 trillion mark.

If the economic equity is 45%, SpaceX's equity value is approximately $796.5 billion, and its total net assets fall between $1.2 trillion and $1.3 trillion.

If the economic stake is 55%, SpaceX's equity value is approximately $973.5 billion, and its total net assets approach $1.5 trillion.

In all three scenarios, the trillion-dollar threshold was surpassed. Musk's status as the "first trillionaire" is now officially established on paper.

However, trillions on paper and trillions in disposable assets are two different things. 85.1% of the voting rights means that Musk most likely holds super voting shares, which are extremely illiquid. Three factors combined lock up the ability to liquidate: lock-up clauses prohibit immediate sale after the IPO; large-scale reductions would directly threaten his absolute control over the company; and if the founder were to sell off a large number of shares, the resulting collapse in market confidence could lead to a stock price drop far exceeding the percentage of shares sold.

Let's do a comparison. In 2025, Bezos cashed out approximately $8.5 billion by gradually reducing his Amazon stock holdings, controlling the pace at no more than 2% to 3% of his stake annually. If Musk were to reduce his SpaceX holdings at the same pace, based on a 45% economic equity scenario, the maximum amount he could cash out annually would be approximately $16 billion to $24 billion. This figure seems enormous, but it only represents 1.6% to 2.4% of his trillion-dollar net worth. Of Musk's trillion-dollar net worth, the portion that can actually be converted into cash each year is likely less than 2%.

This is the core paradox of the "trillionaire" title: the numbers are real, but liquidity is illusory.

4,400 millionaires, and a conditional shareholding plan.

TechCrunch, citing an S-1 filing, revealed that approximately 4,400 employees are expected to become millionaires through the stock ownership plan. At the lowest possible threshold of $1 million per person, the total value of the employee stock ownership plan would be at least $4.4 billion. If we use the median level potentially implied in the S-1 filing (between $1.5 million and $3 million), the total value would be between $6.6 billion and $13.2 billion.

With 4,400 employees, SpaceX set a record for wealth creation in a tech company IPO. Facebook's 2012 IPO created approximately 1,000 millionaires, and Snowflake's 2020 IPO had about 3,000 employees and a market capitalization of about $70 billion. SpaceX's wealth creation has a broader reach, which is related to its organizational structure: software companies typically support a similar market capitalization with a few hundred employees, while SpaceX requires several times more manpower to maintain hardware manufacturing, launch operations, and satellite internet services. While the average wealth per employee may be lower than in pure software companies, its reach is closer to the wealth distribution logic of the manufacturing industry.

However, being a "millionaire" does not equate to having "millions in cash." Employee stock ownership plans are typically issued in the form of restricted stock units or stock options, and the transfer of paper wealth to a bank account must go through three hurdles.

The first hurdle is the lock-up period. In US IPO practice, employee stock ownership plans are typically locked for 180 days, during which trading is prohibited. After Facebook's 2012 IPO, some early employees were unable to cash out when the stock price briefly fell below the offering price because they were prohibited from selling during the lock-up period. The second hurdle is the exercise price. If the stock is in the form of options, employees need to purchase the shares themselves at the exercise price; the difference between the exercise price and the offering price is the actual return. The third hurdle is tax obligations. Exercising options creates a taxable event, with federal tax, state tax, and even the Alternative Minimum Tax (AMT) all added together, meaning the actual amount received may be far lower than the paper figure.

Snowflake's stock price doubled on its first day of trading after its 2020 IPO, but employees could only begin selling in batches after 180 days, during which time the stock price fluctuated by more than 30%. SpaceX's 4,400 millionaires faced the same time misalignment: the IPO pricing day marked the peak of their paper wealth, with the real cash only available after the lock-up period ended, and the stock price at that time depending on market dynamics over the next six months.

The specific form of employee stock ownership, exercise price, and lock-up period details in the S-1 filing have not been fully disclosed in public reports. The fastest window for the 4,400 employees to cash out and the actual after-tax amount received remain to be seen.

Of the total 13.11 billion shares outstanding, only 4.2% are traded on the secondary market.

SpaceX's shareholding structure is key to understanding its stock price fluctuations.

The IPO issued 555.6 million shares at $135 each, raising $75 billion and valuing the company at $1.77 trillion. Based on this valuation, the total number of shares outstanding is approximately 13.11 billion (1.77 trillion divided by 135). The 555.6 million newly issued shares represent approximately 4.2% of the total 13.11 billion shares outstanding.

What does a 4.2% free float mean? The price of any stock is ultimately determined by the buying and selling in the secondary market. When the pool of shares available for trading represents only 4.2% of the total outstanding shares, even a small amount of net buying can drive a significant price increase. Conversely, when the lock-up period ends and 96% of the shares are gradually released, selling pressure will also be amplified.

Let's make a comparison on a scale. Apple's average daily trading volume is approximately $12 billion, corresponding to a turnover rate of 0.36% of its total market capitalization of approximately $3.3 trillion. If SpaceX, with a circulating share capital of $75 billion, were to operate at the same 0.36% turnover rate, its daily trading volume would be approximately $270 million. At a daily turnover rate of 0.36%, the net buying pressure of $270 million in a single day could theoretically drive stock price fluctuations. However, in the actual market, the involvement of market makers, high-frequency trading, and arbitrage funds complicates this relationship.

It cannot be definitively stated that "a net purchase of $7.5 billion will cause the stock price to rise by 10%." However, it is certain that SpaceX's shareholding structure gives the stock price a natural high elasticity, while also making it vulnerable to huge selling pressure after the lock-up period ends.

Here's a diagram illustrating the increase in market capitalization under different price increase scenarios. If the stock price rises by 10% to $148.5, the market capitalization increases from $75 billion to $82.5 billion, an increase of $7.5 billion. If the stock price rises by 20% to $162, the market capitalization increases to $90 billion, an increase of $15 billion. If the stock price rises by 30% to $175.5, the market capitalization increases to $97.5 billion, an increase of $22.5 billion.

These incremental figures only represent changes in circulating market capitalization and do not equate to the net funds required for purchase. The actual stock price is determined by the matching of buy and sell orders; market makers' inventory adjustments, high-frequency trading arbitrage strategies, and institutional investors' portfolio allocation all affect capital efficiency. However, the 4.2% circulating share ratio provides a clear sense of scale: SpaceX's stock price is far more sensitive to capital flows than that of ordinary large-cap stocks.

The real test will come after the lock-up period ends. When 96% of the 13.11 billion shares begin to be gradually released, including shares held by employee stock ownership plans, early investors, and Musk himself, can the secondary market absorb such a flood? The S-1 filing warns investors of the potential for further equity dilution after the IPO, a risk that will be concentrated when the lock-up period expires.

How many millionaires could be created if Anthropic and OpenAI went public?

SpaceX's 4,400 millionaires provide a baseline. By comparison, how would the total value of employee stock ownership plans at Anthropic and OpenAI, if they went public at their current private equity valuations, be comparable?

CNBC and Morningstar have confirmed that Anthropic has completed its Series H funding round at a valuation of $965 billion, bringing its ARR to $30 billion. Forbes and Sacra report that OpenAI is valued at $852 billion, with annualized revenue of approximately $25 billion. Neither company is publicly listed, and the percentage of employee stock option pools and the exact number of employees have not been disclosed.

The core assumption is that if both companies IPO at their current private equity valuations, the employee stock option pool will represent 10% to 15% of the total share capital (a common practice for Silicon Valley unicorn IPOs). Under this assumption:

Anthropic's total employee stock ownership plan is valued at approximately $96.5 billion under a 10% option pool assumption and approximately $144.8 billion under a 15% assumption.

Under the assumption of a 10% option pool, OpenAI's total employee stock ownership is worth approximately $85.2 billion; under the assumption of 15%, it is worth approximately $127.8 billion.

Even under the most conservative assumption of a 10% option pool, the total value of employee stock ownership in Anthropic and OpenAI is 22 times and 19 times that of SpaceX's lowest known value ($4.4 billion), respectively. This gap stems from a structural reason: in SpaceX's $1.77 trillion valuation, Musk's personal share is extremely high (85.1% voting rights mean that even with a lower economic stake, he still holds a huge share), naturally limiting the proportion of the pie allocated to employees. In contrast, Anthropic and OpenAI, as AI-native companies, have smaller employee numbers (OpenAI approximately 1,200, Anthropic approximately 1,500 to 2,000, both estimates based on publicly reported figures) and a higher degree of equity dispersion.

However, this is a horizontal comparison on paper. Anthropic's $965 billion valuation corresponds to an ARR of approximately $30 billion, with a price-to-sales ratio of approximately 32; OpenAI's $852 billion valuation corresponds to an annualized revenue of $25 billion, with a price-to-sales ratio of approximately 34. This ratio is far higher than SpaceX's approximately 9.8 times ($1.77 trillion divided by $18 billion in revenue).

Whether the secondary market is willing to pay the same or even higher price-to-sales (PS) ratios for AI companies is an unproven question. CoreWeave's IPO serves as a warning: its valuation, projected to shrink to $23 billion to $30 billion by its 2025 IPO, is far below the expectations of some investors. This is because the contradiction between the high capital expenditures and low profit margins of AI computing power leasing is being publicly scrutinized.

The financial data of the xAI division exposed in SpaceX S-1 reveals a similar contradiction within SpaceX's AI computing power business. OmniTools previously reported that the xAI division's revenue in 2025 was $3.2 billion, with a loss of $6.4 billion and estimated annualized capital expenditures of approximately $30.8 billion. This data means that the $26 billion in annualized computing power rental revenue that SpaceX receives from Anthropic and Google is backed by a business unit that is still incurring huge losses and whose capital expenditures far exceed its revenue.

If Anthropic and OpenAI are on the verge of IPO, the valuation multiplied by their shareholding percentage will be substantial for employees. However, whether they can be converted into cash in the secondary market depends on whether their price-to-sales ratio can be maintained after listing. The story of SpaceX's 4.2% free float will likely be repeated with these AI companies, only the protagonist will be a large model instead of a rocket.

The other side of computing power revenue

The long-term support for SpaceX's stock price ultimately comes down to one question: can AI computing power leasing revenue cover xAI's losses and capital expenditures?

According to the S-1 filing, xAI has secured $1.25 billion per month in computing power lease contracts with Anthropic and $920 million per month with Google, resulting in an annualized revenue of approximately $26.04 billion. Meanwhile, xAI's revenue in 2025 is projected to be $3.2 billion, with a loss of $6.4 billion. Looking at this alone, the $26 billion annualized revenue is more than enough to cover the $6.4 billion loss; xAI's operating losses are not its biggest problem.

The problem lies in capital expenditures. OmniTools estimates annualized capital expenditures at approximately $30.8 billion, far exceeding the $26 billion annualized revenue. The difference of approximately $4.8 billion requires external financing or funding from SpaceX. More importantly, the $26 billion annualized revenue is based on a static calculation of "$1.25 billion plus $920 million per month," while the contract terms, renewal clauses, and early termination conditions between Anthropic and Google are not disclosed in the S-1 filing.

SpaceX listed water resources alongside chips and electricity as core risks in its S-1 filing. The $30.8 billion in capital expenditure means xAI must continuously purchase GPUs, build data centers, and consume electricity and water. If Anthropic and Google's computing power contracts include clauses to shorten the cycle or allow for early termination, the $26 billion annualized revenue is not a guaranteed sum.

For secondary market investors, this is a knot that must be untied. SpaceX's stock price narrative relies on the high growth of its AI computing power leasing, but the scale of capital expenditure supporting this growth means continuous capital consumption. When 96% of the locked-up shares are released, if xAI is still in a financial state of "annual revenue of 26 billion, annual loss of 6.4 billion, and annual expenditure of 30.8 billion," will institutional investors continue to pay for this cycle?

SpaceX's IPO is a math problem, and the answer isn't on the pricing date, but in the first quarter after the lock-up period ends.