Anthropic's annualized revenue run rate reached $65 billion in July, up from $47 billion in May and $9 billion at the end of 2025, the company told investors, according to TechCrunch. Anthropic expects to finish 2026 between $100 billion and $120 billion at the current growth rate, TechCrunch reported, citing Bloomberg and the Financial Times.
The growth has not translated evenly into use of Anthropic's newest models. Writing on his blog, developer Simon Willison cited the Ramp AI Index, which tracks token spending across tens of thousands of companies, showing Opus 4.8 still drew 28% of model spending in July while Opus 5, released July 24, had captured only 3.5% in its first week on the market. Anthropic's higher-end Fable 5 model took 8% of spending, which Willison attributed to its cost relative to cheaper, good-enough alternatives.
Anthropic also told investors it has 6,000 customers spending more than $100,000 a year on its models, according to Willison's post citing the Financial Times. OpenAI's own annualized revenue jumped 35% in the most recent quarter to more than $40 billion, helped by the July launch of GPT-5.6, the same FT reporting found.
The pattern is not unique to Anthropic. Across the market, newer and more capable models are drawing praise on release and then losing budget share within weeks to systems that cost less and are judged close enough on the tasks that matter.
For anyone budgeting AI spend, the takeaway is concrete: model leadership on a benchmark is not the same as model leadership in a company's actual bill, and the gap between the two is where cheaper competitors are making their money.