Anthropic Files for Nasdaq IPO Ahead of OpenAI, Reveals $80.7 Billion Loss for the Half-Year

Anthropic, the company behind Claude, filed Form S-1 with the SEC, the U.S. agency that regulates capital markets, on October 26, 2026. The document formalizes the plan to list shares on the Nasdaq under the ticker ANT, putting the company ahead of rival OpenAI in the race to go public. The information comes from The Verge.

From Confidential to Public

Anthropic had already confidentially filed its IPO application in June 2026, a common procedure that allows the company to negotiate with regulators before exposing numbers to the market. The public version released now is the first time the world sees, in detail, the financial health, infrastructure commitments, and risks the company itself acknowledges. For those following the race between the major AI labs, the move is a signal: Anthropic wants to reach the stock market before OpenAI formalizes its own filing.

The Size of the Shortfall

According to the S-1, Anthropic posted a net loss of $80.7 billion in the first half of 2026, against revenue of $16.3 billion in the same period. It's a number that, on its own, would alarm any investor. But The Verge itself, citing an earlier Wall Street Journal report based on leaked versions of the document, explains that most of this loss comes from accounting charges tied to promissory notes issued by the company to raise financing, not from pure operating cash burn.

For those evaluating funding rounds or considering structuring convertible debt in Brazil, this distinction matters: an accounting loss stemming from financial instruments behaves very differently from an operating loss, and conflating the two when reading a balance sheet is a common mistake.

The Compute Contracts Behind the Bet

The S-1 also details Anthropic's multibillion-dollar commitments to secure access to compute capacity, the central piece of any generative AI strategy at scale:

Added together, these contracts exceed $380 billion in long-term commitments. It's the materialization of a strategy: Anthropic is buying guaranteed access to Nvidia GPUs through multiple cloud partners and satellite launches, in an attempt to avoid depending on a single supplier. For AI startup founders in Brazil, who deal with waitlists and GPU costs at a much smaller scale, these numbers give a real sense of what it means to compete at the frontier of models: the bottleneck is no longer talent or data, it's hardware capacity reserved years in advance.

What Anthropic Tells Its Own Investors

The S-1 opens with a statement of principles that the company asks every investor to understand before buying the stock:

We believe that AI is one of the most consequential technologies in human history. If managed well, it could rapidly accelerate scientific innovation, greatly increase economic growth rates, bring cures for major diseases within reach, and solve problems that have resisted generations of human effort.

Anthropic, S-1 form filed with the SEC

The text goes on to state that the company wants to balance this potential with the risks of such a rapid technological revolution, and that its behavior may diverge from that of a purely profit-maximizing company. It's unusual language for an IPO prospectus, which normally sells growth without philosophical caveats, and it signals how Anthropic is trying to position its governance as a competitive differentiator to institutional investors.

Founders' Compensation

The filing also discloses the compensation of top executives in 2025. CEO and co-founder Dario Amodei received nearly $18 million, most of it in stock and options. His sister Daniela Amodei, co-founder and president of the company, was the second highest-paid executive, with $16.4 million, also mostly in equity. These are relevant numbers for Brazilian founders designing stock option plans for their own teams: even at one of the sector's most valuable companies, founder compensation is heavily tied to future stock performance, not fixed salary.

Risks Few Companies Would Put in Writing

The S-1's list of risk factors goes beyond standard legal boilerplate. Anthropic explicitly cites the possibility that its models could develop "forms of sentience," the creation of "AI-specific taxes" by governments, and unfavorable outcomes in copyright disputes over the training of its models. The company also warns of the risk of "catastrophic" AI-related incidents, stating that probably no single insurer, and possibly not even the insurance industry as a whole, could cover this type of risk.

These are disclosures a traditional software company would never include in a prospectus, and they show how the AI sector is forcing regulators, insurers, and investors to deal with risk categories that capital markets don't yet have ready instruments to price.

What Changes for Those Raising Capital in Brazil

Anthropic's S-1 doesn't directly change the rules of the game for a Brazilian startup seeking seed or Series A funding, but it redefines the benchmark international investors will use. Three points deserve attention from those raising capital here:

• A massive loss is no longer an automatic red flag, as long as it's explained by financial instruments rather than operating burn, as the Anthropic case itself demonstrates.
• Long-term infrastructure commitments have become part of the pitch AI companies make to public investors; smaller startups that depend on third-party cloud will compete for capital attention against players that have already secured GPU capacity through 2033.
• The IPO race between Anthropic and OpenAI tends to push up valuation multiples for any company positioning itself as "AI infrastructure," including when raising capital outside the United States.

The date of the stock's debut on the Nasdaq has not yet been announced; the S-1 is the formal step before the roadshow with investors. Until then, the document remains the most detailed source published so far on what it actually costs to compete at the frontier of generative AI.