Anthropic Completes $107.3 Billion IPO as Nearly 90 Scholars Jointly Question Its AI Safety Stance
Anthropic began trading on Nasdaq under the ticker ANT, raising $107.3 billion in its IPO and closing with a market capitalization corresponding to a $1.5 trillion valuation. The night before the opening-bell ceremony, nearly 90 scholars co-signed “The MisAnthropic Statement,” accusing the company of racing to develop self-improving AI capabilities while warning of their risks, and of relaxing its safety commitments. The article also says that about 598 non-founder employees hold equity worth at least $100 million, even as the company faces pressures from long-term purchasing and lease commitments, rising computing costs, and revenue that depends on successive releases of new models.
Why read this
SelectedAI score 87
The article juxtaposes the IPO's scale and employees' equity gains with computing-power commitments and the company's revenue structure, highlighting the wealth effect and operating pressures behind its lofty valuation.
Topics
Body · AI translation
It closed at $97.17, with its gain narrowing to 17.07%, while full-day trading volume reached a staggering $86.63 billion. The IPO raised a total of $107.3 billion—$81.3 billion through the public offering plus $26 billion in a private placement—officially setting a new record for the largest IPO in history.
Yet on the other side, nearly 90 people signed a letter on the eve of the bell-ringing ceremony. The signatories included Nobel laureate in economics Oliver Hart, Nobel laureate in physics Sheldon Glashow, Turing Award winner Judea Pearl, three Fields Medalists, MIT physicist Max Tegmark, who has called for a pause on superintelligence, and hacker George Hotz, who treats AI safety as a joke. Tegmark and Hotz, who ordinarily see eye to eye on little when it comes to the direction of AI, put their names on the same document.
The letter was titled “The MisAnthropic Statement”—a play on Anthropic’s name that turns it into the English word for “anti-human.”
Image | The opposition letter
Their allegations were specific: while warning that “self-improving AI” could escape human control, the company is itself racing at the front of the pack to be the first to build it; and while using its “AI safety” persona to raise funds, recruit talent, and win goodwill, it has quietly loosened its own safety commitments.
Seemingly in response to the opposition, Dario repeatedly stressed in interviews that AI safety issues remain far from resolved and that the company would not lower its safety standards in pursuit of commercial growth.
Silicon Valley was thus confronted with a rare spectacle: the largest sum of money in human history and an opposition letter featuring a star-studded roster were both directed at the same company.
Hundreds of Potential Billionaires Created in Five Years
Claude developer Anthropic has just begun trading on Nasdaq under the ticker ANT. The IPO generated a remarkable wealth effect:
Based on the closing price that day, the combined value of the seven founders’ holdings had risen from approximately $140 billion at the offering price to around $162.6 billion. CEO Dario’s stake was worth about $23.06 billion, while that of his sister, company president Daniela, was worth approximately $23.13 billion. Each of the other five co-founders’ stakes was also worth more than $23 billion.
This means that nearly every one of Anthropic’s seven founders owns more than $23 billion in company stock, giving them combined paper wealth exceeding RMB 1 trillion.
But the founders’ wealth is not even the most astonishing part.
Anthropic is distributing the spoils of an enormous wealth bonanza among its rank-and-file employees.
According to estimates by private equity trading platform Hill, following Anthropic’s listing, roughly 598 non-founder employees hold equity worth at least $100 million each. Another research firm, Sacra, estimates that around 2,800 current and former employees own at least $1 million in vested equity.
Anthropic’s wealth-creation legend stems first and foremost from its astonishing growth rate.
In 2021, Dario Amodei and his sister Daniela Amodei founded Anthropic with several former OpenAI colleagues. At the time, it was merely a startup focused on AI safety research, with no products and no revenue.
Image | Dario and Daniela
When FTX collapsed at the end of 2022, the $500 million stake of Anthropic’s largest early shareholder, SBF—representing about 13.56% of the company—was frozen overnight in bankruptcy proceedings. The company was burning cash, revenue was close to zero, and its funding could have dried up at any time. Fortunately, cloud giants Amazon and Google picked up the pieces, subsequently investing more than $30 billion between them.
The commercial turning point, however, came when Claude was transformed from a “chat assistant” into “something that gets work done,” enabling the company to find use cases that customers would pay for. Developers, programming, and enterprise workflows became key markets, while Claude Code emerged as a growth engine. The number of enterprise customers rose from 1,500 at the end of 2025 to 11,900 in September 2026, while revenue climbed from $41 million in 2023 and $4.585 billion in 2025 to $33.2 billion in the first three quarters of 2026.
Claude’s rapid development has enabled Anthropic to build increasingly strong commercial influence in fields including programming, enterprise services, and AI agents. The capital markets also assigned it an extraordinarily generous valuation: $1.5 trillion.
The rapid surge in the company’s valuation directly magnified the value of equity held by early employees.
Anthropic disclosed in its IPO filing that, as of September 30 this year, the paper gains on employees’ stock options had reached $119 billion, of which approximately $73 billion had already vested.
Even more noteworthy is the cost of the employees’ equity.
For options granted before the end of 2023 that remain unexercised, the average exercise price is just $0.52. This means that some employees who joined the company early may need to pay only a very low exercise cost to obtain shares worth millions or even hundreds of millions of dollars.
This also explains why Anthropic’s IPO has produced such an extraordinary employee wealth-creation effect.
At the end of 2024, Anthropic had around 940 employees; by the end of September this year, that figure had grown to approximately 5,300.
Phoenix New Media Technology noticed a job posting from May 2024, when Anthropic was hiring an executive assistant responsible for managing the schedules of company executives, coordinating meetings and travel, processing expense reimbursements, and other duties. The annual salary ranged from $140,000 to $190,000. The posting specifically noted: “Equity will be a significant component of total compensation.”
Image | Executive assistant job posting
This means that even some early employees who did not hold core technical positions may see their wealth leap to an entirely new level in this IPO because of the stock options they received at the time.
Interestingly, the prices of detached homes in San Francisco surged almost in tandem with the wealth-creation effect of Anthropic’s IPO. Their median price jumped nearly 30% year over year in October to $2.39 million, the largest monthly increase in the area’s median home price since February 2018.
At a traditional technology company, an ordinary employee seeking to accumulate wealth of more than $100 million would often need to remain with the company for a long time, earn continual promotions, and perhaps even wait through several cycles of corporate growth. In the AI era, a successful startup and a single leap in valuation can compress that process to around five years.
Anthropic is not an isolated case.
A previous employee share sale at OpenAI generated enormous wealth for hundreds of current and former employees. According to disclosures related to the transaction, more than 600 employees cashed out approximately $6.6 billion through share sales, averaging around $11 million per person.
Anysphere, the parent company of the AI coding tool Cursor, was founded by four young people from MIT who initially just wanted to make coding with AI more convenient. After Cursor became popular, the company’s valuation soared, and it was later acquired by SpaceX for $60 billion. According to estimates by foreign media, each of the first 50 employees to join the company earned equity returns ranging from tens of millions to more than $100 million.
Perplexity, another AI search company founded just four years ago, has likewise seen its valuation climb from approximately $500 million in its early days to more than $30 billion, rapidly magnifying the value of equity held by early investors and employees.
On the one hand, developing large models requires enormous capital investment, but once a company establishes technological and product advantages, it has an opportunity to enter the global enterprise market rapidly. On the other hand, competition for AI talent is exceptionally fierce, and companies increasingly rely on equity incentives to attract top researchers and engineers.
Anthropic’s case is particularly representative: the company expanded from a team of fewer than 1,000 people to 5,300, while the value of employee equity incentives rose sharply alongside the company’s rapidly growing valuation.
Image | Founder Dario
Beyond paper wealth, how long can the revelry last?
However, going public does not mean that all this wealth can be cashed out immediately.
Anthropic employees face restrictions on selling their shares. Non-founder employees may sell no more than 10% of their holdings on the first trading day, followed by another 35% after the company releases its fiscal 2026 results.
Therefore, the reported hundreds of billionaires and thousands of millionaires represent potential wealth calculated from the offering price and their shareholdings, rather than cash already in their accounts.
More importantly, an increasingly taut rope is hidden on the balance sheet.
Publicly available information shows that, as of the end of September, Anthropic’s signed long-term purchase and lease commitments had reached nearly $895.4 billion, including approximately $69.8 billion due within the next 12 months. These contracts are highly rigid, and many contain minimum payment obligations requiring payment regardless of actual usage.
This means Anthropic’s expenses are fixed while its revenue is variable. Enterprise customers can reduce API calls billed by usage, and industry price wars can drive down unit prices, but its computing bills will not be discounted accordingly.
Moreover, profitability pressures have already emerged. Anthropic had only just achieved an operating profit of $974 million in the second quarter of this year before immediately swinging to a $567 million loss in the third quarter. The reason is harsh: computing costs surged 77% quarter over quarter in the third quarter, while the revenue generated for every $1 spent on computing fell from $2.08 in the second quarter to $1.72.
The quality of its revenue is likewise less stable than it appears. Around 80% of the company’s revenue comes from usage-based APIs, meaning fluctuations in customer budgets feed directly into revenue.
Even more notably, new model versions released during the six months ended September 2026 contributed more than 80% of that quarter’s revenue. In other words, Anthropic must continually launch more powerful models to preserve its revenue trajectory. If the pace of iteration slows, both customer usage and willingness to pay could decline.
Image | New models released by Anthropic in the past six months; data from the company’s official website
The governance structure presents challenges as well. Cloud giants such as Google and Amazon are simultaneously shareholders, suppliers, and competitors with their own large models; meanwhile, the seven co-founders control the company through shares carrying enhanced voting rights, leaving public shareholders with limited influence over major decisions.
In addition, more than 20 lockup-expiration events will occur during the first year after the listing, potentially releasing large volumes of shares for sale around financial-reporting windows. If the share price underperforms expectations, the potential selling pressure will be further amplified.
A $1.5 trillion valuation is not buying Anthropic as it stands today, but the probability that it will continue to outpace its computing bills, model-iteration demands, and competitive landscape over the next five years.
Whether the company can continue expanding revenue, maintain its technological lead, and convert its enormous investment in computing into stable profits will determine whether this wealth bonanza can continue.
Regardless of how its share price subsequently changes, however, Anthropic has already demonstrated a new form of wealth creation in the AI era: technological breakthroughs can give rise to a corporate giant in a short time, while equity incentives spread the resulting wealth beyond founders and investment institutions to thousands of employees.
AI Trends