Anthropic, the company behind the Claude AI models, has made its US IPO filing public, according to Reuters. The filing gives investors their first full look at the company's finances, and the growth is striking. Revenue climbed from $41 million in 2023 to $386 million in 2024, then to $4.59 billion in 2025, and reached about $16.3 billion in the first half of 2026 alone. Most of that comes from businesses paying per use of Anthropic's models, with the rest from paid subscriptions such as Pro and Max. Once the IPO closes, the company will trade on the Nasdaq under the ticker "ANT."

The costs are just as large. According to Reuters, computing power is Anthropic's biggest expense, and the company has committed more than $500 billion over a decade to six partners, much of it non-cancelable. That includes at least $110 billion with Amazon, about $114.5 billion with Google and $44.9 billion with Microsoft. Amazon and Google are also among Anthropic's largest investors, an example of the circular financing that has become common in AI. Anthropic posted a $693 million operating loss in the first half of 2026. Its net loss reached $80.7 billion in the same period, which Reuters attributes mostly to a jump in the value of bonds that convert into company shares.

The listing comes at a shaky moment for the market. Global stocks have sold off as investors question whether the trillions flowing into AI data centers and chips will pay off. Several companies have delayed their own IPOs. Reuters reports that some investors now value Anthropic at about $1.5 trillion, down from roughly $2 trillion a week ago. The filing also lands as Anthropic's own research flags risks from advanced AI, including code sabotage and information manipulation in controlled tests. Rival OpenAI has confidentially filed for its own IPO and is expected to list by early 2027.