Anthropic’s annualized revenue run rate climbed to about $65 billion by the end of July, reflecting rapid growth in enterprise demand for its Claude models as the AI developer prepares for a potential public offering.
CNBC reported the figure after Anthropic circulated a confidential update to investors over the weekend, citing three people familiar with the information. Bloomberg first reported the new run rate. Anthropic declined to comment publicly on the financial details.
The reported figure represents roughly sevenfold growth from a year earlier and a sharp increase from May, when Anthropic said its annualized run rate had exceeded $47 billion. The company generated approximately $10 billion in revenue during all of 2025, CNBC reported.
An annualized revenue run rate is not the same as revenue recognized over a full year. It generally extrapolates recent sales over a 12-month period, making it a useful measure of current momentum but not a guarantee of future results. For a fast-growing AI provider, the metric can shift quickly as customers expand or reduce model usage, negotiate enterprise contracts, or move workloads among competing services.
Anthropic also gave investors a preliminary second-quarter revenue figure of $11.5 billion, representing a 14-fold increase from the corresponding period a year earlier, CNBC reported, citing a person familiar with the update. The company has not publicly released audited financial statements confirming the number.
The growth has been driven in part by adoption of Claude within businesses, where generative AI models are increasingly used for software development, document analysis, customer support and internal knowledge tools. Anthropic sells access through consumer and workplace subscriptions, its own application programming interface, and cloud platforms that allow companies to integrate Claude into existing systems.
Usage-based model revenue can expand rapidly because customers pay for the volume of text and other data processed. That model also carries substantial costs: developers must secure advanced chips, data-center capacity and electricity to train and operate increasingly capable systems. Strong top-line growth therefore does not by itself establish profitability, and prospective investors are likely to scrutinize Anthropic’s compute spending, margins and long-term infrastructure commitments alongside its revenue.
The investor update arrives as Anthropic takes preliminary steps toward an initial public offering. CNBC reported that the company confidentially submitted a prospectus to the US Securities and Exchange Commission in June and has held early discussions with possible investors. No formal timetable for a listing has been announced.
A public-market debut would expose more of Anthropic’s finances, including the relationship between rising model usage and the cost of serving that demand. It would also give investors a clearer basis for assessing the company’s reported $965 billion valuation. Rival OpenAI recently reached an annualized revenue run rate of about $40 billion, CNBC reported, illustrating the scale of commercial spending now flowing to leading model providers.
Anthropic’s recent expansion has continued despite disruptions involving the US government. In June, the company temporarily restricted access to two advanced models, Claude Fable 5 and Mythos 5, to comply with an export-control directive that invoked national-security authorities. Access was restored roughly two weeks later following negotiations. The episode followed a separate dispute with the Pentagon over conditions governing military uses of Anthropic’s technology.
The company’s next disclosures will be closely watched for evidence that its July pace can be sustained. Until Anthropic publishes formal financial statements, the reported run-rate and quarterly figures remain confidential investor metrics rather than publicly verified results.