According to Bloomberg, confirmed by CNBC, by the end of July 2026, Anthropic's annualized revenue run rate reached $65 billion, about 7 times higher than a year earlier, and the company is preparing a public offering with a target valuation of $965 billion.
What happened
Over two months, the revenue run rate accelerated significantly: in May, the run rate was $47 billion, and by the end of July — $65 billion. For the second quarter of 2026, the company reported preliminary revenue of $11.5 billion, 14 times higher than the same period a year earlier, while Anthropic's revenue for all of 2025 was about $10 billion. CNBC reported that the confidential IPO prospectus was filed with the SEC in June, and meetings with investors are already underway.
Context
Anthropic is one of the leading AI labs, and its co-founder and CEO is Dario Amodei. According to the latest data, the company has overtaken OpenAI in annualized revenue run rate, which recently reached $40 billion, although the article itself acknowledges that the companies may calculate this metric slightly differently. The question of the sustainability of growth is not accidental: as recently as June, Anthropic experienced outages of the Claude Fable 5 and Mythos 5 models, which, against the backdrop of such rapid acceleration, made the financial results particularly vulnerable to criticism.
Why this matters for the industry
The main signal for the industry: the market is beginning to value leading AI labs by entrepreneurial revenue, not by model benchmarks, and the upcoming Anthropic offering will set a new benchmark for the capitalization of the entire class of AI companies. Even more important is the process itself: the public version of the S-1 will for the first time disclose the detailed financial structure of a top AI company — capex, inference costs, revenue structure. For the community, this will be the first open dataset for verifying the economics of scaling and inference unit economics, which was previously unavailable in open sources.
Why this matters for users
For readers and teams building products on LLMs, the news confirms that enterprise LLMs have already become a market worth tens of billions of dollars, and large language models have become an infrastructure category, not a pilot. Practical takeaway: dependence on a single AI provider remains a first-order risk, so it is worth planning capacity, rate limits, and an outage plan for the Anthropic API now — after the listing, the pressure of the quarterly reporting cycle on prices and limits usually intensifies. It is also worth tracking the offering prospectus as a future source of open data on the market structure.
What is still unknown / limitations
The $65 billion figure is an annualized estimate, not audited revenue, and the quarterly $11.5 billion is marked as preliminary: in the original text, the quarter is called the "last completed one," which conflicts with the wording "second quarter of 2026." The conclusion about overtaking OpenAI ($65 billion vs. $40 billion) is methodologically impure: the metrics may be calculated differently by the companies, so this is more of a narrative than a measurement. The mention in reports of positive adjusted operating profit is also weak as evidence: the metric is adjusted, unofficially disclosed, and does not disclose the cost of inference per token. The final terms of the offering, including the $965 billion valuation, have not yet been confirmed: the prospectus is in confidential mode, and the offering itself is yet to take place.
Sources
Author
Look at AI, editorial team