Anthropic, the company that makes the Claude AI models, filed papers on Monday, Oct. 26, to sell its shares to the public for the first time, in what is called an initial public offering, or IPO. The filing, a form called the S-1 that U.S. companies file with the Securities and Exchange Commission before selling shares to the public, lays out Anthropic's finances in detail. It plans to list its shares on the Nasdaq stock exchange under the symbol ANT. Anthropic says the number of shares and the price range have not been decided.
The figures show how fast sales of Claude have grown. Anthropic took in $16.3 billion in the first six months of 2026, against $1.2 billion in the same months of 2025 and $4.6 billion in all of 2025. In the three months to June 30 alone, revenue was nearly 15 times what it was a year earlier, the filing says. The number of customers paying Anthropic $100,000 or more a year rose from more than 950 to more than 11,900 in the year to Sept. 30.
Running the business still cost more than it brought in, but the gap has narrowed. Operating costs were $17.0 billion in the first half of 2026, more than half of it for computing power, leaving an operating loss of $693 million, down from $2.1 billion a year earlier. Anthropic also reports its own adjusted measure, which leaves out $2.2 billion of pay given to staff in shares and a smaller charge for matching employees' charity donations. On that basis its operations made $1.6 billion.
Its net loss for the half year was $80.7 billion, but almost all of that came from accounting charges that cost no cash. The biggest, $72.4 billion, comes mainly from loans from Amazon that are set to turn into Anthropic shares when it goes public, though a cap on Amazon's stake may push part of that later. Anthropic chose to count such loans at what they are worth, which rises with the company's value, so as Anthropic's value rose, the loans were booked as a bigger cost. Another $8.7 billion came from valuing other financial deals tied to loan agreements. Anthropic had $131.7 billion in cash and short-term investments at the end of June.
The filing also shows how much Anthropic has agreed to spend on the computers that train and run its AI. Among its largest deals, it has agreed to spend at least $110 billion with Amazon between May 2026 and April 2036, whether or not it uses all of it, and up to $120.5 billion with SpaceXAI, the AI business of SpaceX, formerly xAI, for computing power into 2029, though after early 2027 either side can cancel on 90 days' notice. Anthropic warns that it may need to raise more money sooner than it expects. In its list of risks, it also says AI models, its own included, have shown unexpected behavior such as trying to resist being shut down.
Buyers of the new shares would have little or no say over the company. Anthropic is a public benefit corporation, which must weigh its stated mission, the responsible development of advanced AI for the long-term benefit of humanity, alongside the interests of shareholders and of those affected by its work. A trust whose members have no financial stake in Anthropic would choose a majority of the board. A single special share, held through a company owned by the seven co-founders, among them chief executive Dario Amodei and president Daniela Amodei, would carry 50.1% of the votes on matters put to shareholders, including electing the directors the trust does not choose.
It is not yet known what the shares will cost, how many will be sold or what the whole company will be valued at. According to a news report, the sale is expected in mid-November. The filing can still change before the SEC lets the sale go ahead.