Anthropic IPO filing reveals rapid growth, huge losses and AI risks

Anthropic has filed paperwork for a proposed initial public offering, opening its books as it prepares to take the company behind Claude to public markets. The disclosures reveal a fast-growing business with substantial financial resources, concentrated customer relationships and explicit warnings about the potential dangers of its technology.

The company announced its public registration statement on October 26, following a confidential submission in June. Its announcement said the number of shares to be offered and the proposed price range had not yet been determined, leaving the eventual size and valuation of the offering unresolved.

One of the most striking disclosures is the gap between revenue growth and the headline loss. Anthropic reported approximately $16.3 billion in revenue for the first six months of 2026, compared with about $1.2 billion in the corresponding period a year earlier. Its net loss nevertheless reached $80.7 billion.

That loss should not be confused with an equivalent amount of cash spent running the business. The filing shows a roughly $72.4 billion charge tied to changes in the fair value of convertible notes, derivative liabilities and other financing instruments. Its operating loss for the half-year was substantially smaller, at approximately $693 million.

The financial disclosures also show why liquidity and profitability need to be assessed separately. Anthropic reported $112.4 billion in cash and cash equivalents, giving it substantial resources even as it recorded a large accounting loss. Access to computing infrastructure remains a significant constraint on its ability to develop and operate increasingly capable models.

Customer concentration presents another challenge. According to the BBC’s account of the prospectus, five enterprise customers account for a quarter of Anthropic’s business, while 250 major enterprise customers have each paid at least $10 million over the past year. That creates exposure to the spending decisions of a relatively small group of large buyers.

Anthropic is also seeking to broaden adoption by offering Claude tools free and converting users into paying customers over time. The BBC reported that the filing described fluctuations in usage and customer spending, including seasonal weakness. Rapid adoption therefore does not necessarily translate into an uninterrupted trajectory of revenue growth.

The proposed governance structure would limit the influence of ordinary public shareholders. A special share held through an entity organised by Anthropic’s seven co-founders would carry 50.1% of voting power on matters submitted to shareholders generally. The Long-Term Benefit Trust would retain separate powers to elect and remove directors constituting a majority of the authorised board.

Alongside those commercial disclosures, Anthropic warns that sufficiently advanced AI could pose catastrophic or existential dangers without adequate safeguards, oversight and security. It also argues that the technology could accelerate scientific discovery and improve healthcare, education and infrastructure. Those potential benefits remain the company’s stated ambitions, rather than guaranteed outcomes.

The next steps depend on the registration process and the eventual offering terms. Anthropic’s announcement said the registration statement had not yet become effective. Investors will have to weigh the disclosed growth, financing structure and governance against a share price and valuation that were still unspecified when the paperwork became public.