Anthropic IPO filing leaves key questions about growth and spending

In short: Anthropic has filed paperwork for a possible IPO, but the documents leave major gaps about how its business is doing right now and what it has promised to spend.

What's going on

Anthropic, the company behind the Claude AI models, published draft IPO paperwork this week. An IPO is when a private company sells shares to the public for the first time, like opening up ownership to anyone who wants to buy in.

The Financial Times notes that three important numbers are missing from the filing. One is an up to date revenue “run rate” for October, which is basically an estimate of how much money the company would make in a year if a recent month repeated. Investors often want that number because it hints at whether growth is speeding up or slowing down right before the share sale.

Outside analysis cited by the FT suggests Anthropic’s month to month revenue growth swung sharply, from 88 percent in February to 4 percent in September. The FT links this to a burst of interest earlier this year in Claude Code, a tool aimed at helping people write software.

A second missing number is how much Anthropic has committed to spend on future computing capacity. The filing mentions more than $650 billion in deals, while some estimates put the total closer to $1 trillion, stretching out to 2045. You can think of this like signing a very long, very expensive lease for the “engine room” that powers AI.

A third question is whether recent price cuts are helping demand grow, or whether they signal a price war. Anthropic said one newer model can complete tasks 40 percent more cheaply than the previous one.

What to watch

Watch for updated numbers from Anthropic as the IPO process moves forward, especially October revenue and clearer details on long term compute spending. Those figures could shape how investors judge not only Anthropic, but also whether today’s AI boom is turning into steady, repeatable business.

Source: Financial Times