Anthropic Files S-1 for Nasdaq IPO, Discloses Billions in Losses

Anthropic just put a number on paper that sounds almost made up: a net loss of roughly $80.7 billion in six months, filed the same week reports put its asking valuation north of $1.5 trillion. On October 26, the AI lab behind the Claude chatbot filed its Form S-1 with the US Securities and Exchange Commission (SEC), the paperwork a company must file before it can sell shares to the public. It plans to list on the Nasdaq Global Select Market under the ticker 'ANT.'

What the filing actually says

An S-1 is where a company has to stop talking in vague growth percentages and show real numbers, audited and under oath, more or less. Anthropic's version shows a business growing fast and burning cash just as fast. Reporting on the filing puts the company's revenue at roughly $16.3 billion for the first half of 2026, against a net loss of about $80.7 billion over the same stretch. That loss figure looks almost fictional next to the revenue, and the filing itself reportedly attributes most of it to non-cash accounting charges, things like the paper revaluation of employee stock awards and warrants, rather than money actually leaving the building. It is a real loss on paper, but not quite the same as a company spending $80 billion in cash it does not have.

• Exchange and ticker: Nasdaq Global Select Market, symbol 'ANT'
• Lead underwriters: Goldman Sachs and Morgan Stanley
• Share count and price range: not yet set, per the filing
• Reported valuation chatter: somewhere between $1.5 trillion and $2 trillion, well above the roughly $965 billion private valuation from Anthropic's May 2026 funding round
• Expected trading debut: reports point to mid-to-late November 2026, though nothing is locked in until the registration is declared effective

The fine print nobody wants to put on a billboard

Buried in the risk-factor section, the part of an S-1 where lawyers list every bad thing that could plausibly happen, is language acknowledging that companies building foundation models (the large, general-purpose AI systems that power chatbots and coding assistants) have already faced allegations of playing a role in cases of serious self-harm or violence, and that this kind of litigation and regulatory scrutiny is a real risk going forward. That is an unusually blunt thing to put in a prospectus meant to attract investors, and it says something about how seriously AI safety liability is now treated in boardrooms, not just in ethics papers.

A company with an $80 billion loss and a trillion-dollar asking price isn't really being judged on its balance sheet. It's being judged on a bet about where AI goes next.

Why this matters beyond Wall Street, especially in India

Anthropic opened an India office in 2025 and has been courting large Indian enterprises and developer communities, and Claude has become a genuine working tool inside Indian companies, not just a novelty. If you are a developer at an Indian bank or startup whose team now ships code with Claude's help, the financial health of the company behind that tool is not an abstract Wall Street story. A well-capitalized, publicly accountable Anthropic is more likely to keep investing in the product; a shaky one is a supply-chain risk for every team depending on it.

There is also a direct money angle. Indian retail investors cannot simply log into Zerodha and buy Nasdaq shares, but once ANT actually starts trading, those who want exposure can do it through the Reserve Bank of India's Liberalised Remittance Scheme (LRS), which lets a resident individual send up to $250,000 abroad each financial year, often routed through international broking apps that Indian investors already use for US stocks. It is the same path Indians have used to buy into Apple or Nvidia, not a special arrangement for Anthropic. Worth remembering too: nothing is actually for sale yet. An S-1 filing is a declaration of intent, not a ticket.

It is also a useful mirror for India's own IPO market. SEBI requires similarly extensive risk disclosures from companies going public here, as this site covered with Jio Platforms setting its price band for India's biggest-ever IPO. The instinct to bury the scary stuff in page 140 of a prospectus is universal; so is the obligation to put it there at all.

What happens next

The filing is public, but not effective, meaning no shares can change hands yet. Anthropic will likely take investors on a roadshow, a series of pitches to big institutional buyers, before pricing the offering and setting a final share count. Earlier reporting on this site covered investor pushback over Anthropic's valuation just a week before this filing landed, and that tension between an eye-watering asking price and a historic loss is exactly what will get argued over on the roadshow circuit in the coming weeks.

Whatever price Anthropic eventually lists at, this filing is the first time the company has had to show its actual numbers instead of letting valuation rumors do the talking. For an industry still arguing about whether AI spending is a bubble, that kind of forced transparency is worth paying attention to, even if you have no intention of ever buying the stock.