Anthropic IPO Filing Details Reveal $16 Billion in Revenue — and an $80 Billion Loss That Needs Explaining
Anthropic formally started the process of going public on the Nasdaq Global Select Market under the ticker symbol “ANT” on October 26 by filing its Form S-1 with the Securities and Exchange Commission. The more than 300-page document arrived in a market already uneasy about infrastructure costs and AI valuations, and it did nothing to ease the tension. What it accomplished was to provide the world with its first in-depth examination of a business that had previously functioned virtually entirely outside of the public eye.
The headline revenue figure of $16.3 billion in just the first half of 2026 is startling. For background, Dario Amodei, Daniela Amodei, and a group of researchers who had left OpenAI founded Anthropic in 2021. With revenue up 1,368% year over year in Q2 2026, the company has been expanding at a rate that makes most technology comparisons seem insufficient. It’s not a typo. With a net revenue retention rate of about 900%, consumers are not only sticking around but also increasing their spending significantly over time. Even before the price is decided, the fact that Goldman Sachs and Morgan Stanley are leading the offering gives you an idea of the institutional confidence in the transaction.
Then there is the loss, which needs to be read carefully. The $80.7 billion net loss for the first half of 2026 will make headlines and merits further analysis. Of that amount, about $72.4 billion is a non-cash accounting charge related to financing valuations and convertible notes rather than operating cash burn. After removing that, adjusted operating income turned positive at $1.62 billion, while actual operating loss shrank to $693 million.
Growth investors will concentrate on that change, which went from a $1.85 billion adjusted operating loss the year before to a positive adjusted figure. Some retail participants who don’t read footnotes might be alarmed by the larger loss figure. Complex IPO filings always carry that risk.
The scope of Anthropic’s infrastructure commitments may be more difficult to justify. The document reveals multi-year compute capacity agreements totaling more than $475 billion, and possibly more than $800 billion, mainly related to Google Cloud and Amazon Web Services. Nor are those typos. The foundation of Anthropic’s business is access to massive amounts of processing power, and obtaining that access entails long-term financial commitments that are hard to fully appreciate from the outside. As part of a private placement, Alphabet, which has already invested billions in the business, has the option to buy up to $10 billion in Class N common stock concurrent with the IPO, modifying current note arrangements totaling $20 billion. This document has Google’s fingerprints all over it.
In terms of governance, Anthropic was established as a Delaware public benefit corporation, which incorporates social purpose into its charter. Through a unique arrangement intended to maintain long-term safety oversight even after the company goes public, the seven co-founders retain 50.1% of voting rights. Anthropic seems to be attempting to balance attracting public market capital with preserving the mission-driven culture that sets it apart from purely commercial rivals. The filing is unable to address whether those two things can coexist once the stock is trading.
The nearly 100-page risk section addresses everything from slowing growth rates and regulatory scrutiny to something more unusual: the filing specifically identifies the risk that sophisticated AI systems could produce outputs “that could be interpreted as coercive, deceptive, or manipulative,” including actions that resemble blackmail. In an S-1, that kind of disclosure is actually uncommon. It reads more like a business being remarkably open about what it’s developing than it does like legalese.
Only a few months after SpaceX’s Nasdaq debut raised $85.7 billion in the biggest IPO ever, the offering is anticipated to take place before Thanksgiving. It will soon be evident if Anthropic can match that intensity or if the AI sentiment has changed enough to make the reception more difficult. A possible valuation of $1.5 trillion to $2 trillion is suggested by market estimates. No one is quite sure how to price that range yet.
