Anthropic Revenue Hits $16.8B in Q3 Ahead of IPO

Anthropic brought in $16.8 billion in revenue in the third quarter and was profitable on an adjusted basis for the first time. That is according to a Financial Times report published on October 14. The figures were shared at a private meeting between Anthropic executives and investors who want to buy shares once the company goes public.

The timing matters. Anthropic is pushing ahead with its initial public offering and is widely expected to list in November. It is reportedly aiming for a valuation of up to $2 trillion, which would put it among the most valuable companies in the world.

The numbers, and what is missing

Earlier figures from the Wall Street Journal put Anthropic's first-quarter revenue at $4.8 billion. Later reports said the company more than doubled that to $11.6 billion in the second quarter. The new third-quarter figure is a sequential increase of about 45%.

The headline number has some gaps, though. The Financial Times' anonymous source did not say whether $16.8 billion is gross or net revenue. The source also did not explain how Anthropic defines "profitability" in this case. Investors want more than a top-line figure. A key question is how much Anthropic spends to grow this fast.

Because Anthropic is still private, it has no obligation to publish that kind of detail. So any data point that does come out gets a lot of attention. Many investors treat Anthropic's revenue as a signal for the wider AI market, which has added trillions of dollars to the value of technology stocks.

Why one number can move markets

Last week showed how sensitive markets are to these figures. The Financial Times said that previously reported figures for OpenAI were wrong. It lowered its estimate of OpenAI's annualized revenue run rate from about $68 billion to $50 billion and blamed a simple miscalculation.

The correction hit tech stocks hard. The Nasdaq fell 1.25%, and Oracle, CoreWeave, Intel and Nvidia each fell by more than that. All of these firms are tied to the big AI labs, and many have signed multiyear contracts with them worth billions of dollars. When the outlook for Anthropic or OpenAI changes, their suppliers and partners feel it too.

From challenger to front-runner

OpenAI was long seen as the leader in AI, but Anthropic has overtaken it this year. The company has released a steady run of frontier models. Its Claude Code tool, which writes software from natural-language instructions, has been widely adopted by businesses.

Anthropic is now described as one of the fastest-growing companies in history. In May it closed a $65 billion funding round that valued it at $965 billion. Funding on that scale is normal in AI, where developers have to pay for very expensive chips to train and run their models.

Safety and cost pressures

The growth comes at a difficult moment for the industry. In September, Anthropic researcher Jacob Coxon resigned. He said that "the people building AI earnestly believe that it could kill us all by the end of the decade."

His comments followed several high-profile cybersecurity incidents involving frontier models. Anthropic CEO Dario Amodei then published an essay calling on companies to "slow the pace at which we improve the capabilities of AI models," so the industry has time to put proper safety measures in place. OpenAI CEO Sam Altman, Elon Musk and others supported the call. U.S. President Donald Trump rejected it and pushed instead for companies to regulate themselves.

Cost is the other pressure. Many companies are moving to open-source and open-weight models that cost much less to use, and Anthropic and OpenAI have had to respond. Last week Anthropic released Claude Haiku 5.5, its most advanced small language model. Anthropic says it costs about 75% less than Haiku 4.5, which puts it level with OpenAI's small model, GPT-6 Luna.

Our Take

The third-quarter figure suggests Anthropic's business is still growing fast. The growth rate is slowing, though: revenue more than doubled from Q1 to Q2 but rose about 45% from Q2 to Q3. That is still unusual at this scale, but a $2 trillion valuation probably assumes the curve will not flatten much further.

The bigger issue is how little is confirmed. A leaked number with no definition of revenue or profit is a thin basis for a listing at this size. The OpenAI correction showed how quickly sentiment can shift when a figure turns out to be wrong. The IPO filing should eventually answer the open questions about spending, margins and what "adjusted" means.

It is also worth watching whether lower-priced models like Haiku 5.5 reduce revenue per customer faster than they bring in new usage. And watch whether any enterprise pullback shows up in the numbers. For companies building on Claude, a public Anthropic would mean more transparency and also closer scrutiny from shareholders.