Anthropic IPO Filing Shows $16.3bn in Six-Month Revenue and an $80.7bn Loss

The Anthropic IPO moved from rumor to paperwork on Monday, when the maker of the Claude chatbot publicly filed its prospectus with US regulators. The document gives the first audited look at the finances of one of the most closely watched companies in tech. It shows revenue growing at a pace few businesses have matched, alongside losses large enough to fuel the debate over whether the AI boom can pay for itself.

Revenue growth that is hard to overstate

The company’s annual revenue was $41m in 2023 and $386m in 2024. In 2025 it reached $4.585bn, an increase of more than 1,000%. In the first six months of 2026 alone it booked $16.286bn, against roughly $1.1bn in the same period a year earlier, according to City AM’s reading of the filing.

Unlike OpenAI, which has built its brand on a mass consumer product, Anthropic has pitched itself mainly at businesses. As of 30 September it counted more than 14,000 enterprise customers. Of these, 11,900 had generated $100,000 or more in revenue, more than 1,900 had passed $1m, and about 250 had topped $10m. “The speed and depth of our enterprise customer adoption are powerful evidence of Claude’s position at the frontier of AI,” executives wrote.

An $80.7bn loss, with caveats

Against that growth sits a net loss of roughly $80bn for the first half of 2026. City AM puts the precise figure at $80.7bn. It also reports that about $72bn of that came from non-cash accounting adjustments tied to convertible debt and other financing instruments. Those adjustments mean the headline loss overstates the cash the company burned in the period, though the filing still shows heavy spending. Anthropic lost $42bn in 2025, and it spent $9bn on AI infrastructure in the first half of this year, compared with $7.3bn across all of last year.

The company said it had raised $141.1bn as of the end of June and has taken on $6.2bn in debt to build out its infrastructure. It also spends billions each year on computing power from third parties and has committed hundreds of billions of dollars to future capacity, including cloud agreements with Amazon and Google.

Valuation, timing and the bubble question

The company plans to list on the Nasdaq under the ticker ANT, with Goldman Sachs and Morgan Stanley leading the offering. It confidentially filed in June, and a debut could come as early as November, though no date or price has been set. Analysts and investors have valued Anthropic at more than $1.5tn. Bankers have reportedly aimed for $1.5tn to $2tn, even as recent market volatility has raised doubts about appetite for richly priced AI stocks.

Supporters say the numbers justify the ambition. Jared Carmel of Manhattan Venture Partners, an early investor, argued that a $2tn valuation is not unreasonable: “The market isn’t pricing a chatbot. It’s pricing an infrastructure layer for how every business is going to run.”

Risks the prospectus spells out

The filing runs to more than 90 pages of risk disclosures. One concerns concentration: a quarter of revenue comes from the top five customers, although none accounts for more than 10% of the business. Losing a few of them could still dent results.

Anthropic also lays out four categories of AI risk: misuse by bad actors, loss of control, authoritarian abuse and economic disruption. The company has faced scrutiny over its safety practices, and a former researcher has accused it and OpenAI of “racing straight to self-improving superintelligence.” The prospectus nonetheless carries the grand language typical of the sector, describing AI agents as “a country of geniuses in a datacenter, available to every organization and sector of the economy.”

For investors, the Anthropic IPO will test a simple question. Will customers keep paying at this pace for long enough to cover the cost of the computing that produces the revenue? The filing does not answer that, but it gives the market real numbers to argue over.

Explore more on this topic with Vexel Search.