An examination of billions of anonymized credit card transactions from about 28 million U.S. consumers, conducted for TechCrunch by Indagari, a consumer transaction analysis company, shows Claude gaining paid subscribers in record numbers.

Now, as with all big-data analysis, caveats exist. While this data is substantive, it doesn’t include every consumer. That means that Indagari can’t calculate Anthropic’s total current or new user numbers. It also doesn’t include Claude’s enterprise business (which is its bread and butter) or its free-tier users (those not paying Anthropic at all). Estimates for total Claude consumer users are all over the map (we’ve seen figures ranging from 18 million to 30 million) but Anthropic has not disclosed this data. A spokesperson did tell TechCrunch, however, that Claude paid subscriptions have more than doubled this year.

What’s notable is that consumers pulled out their wallets in record numbers for Claude between January and February. Also interesting, previous users returned to Claude in record numbers in February as well, Indagari told TechCrunch.

Claude total users six months Sept-Feb.Image Credits:TechCrunch

Indagari tells us that the majority of new subscribers are at its lowest tier, “Pro” users ($20 per month, compared with $100 or $200 per month).

Data through early March confirm that subscriber growth is continuing. (Data is available with a two-week delay.)

Claude weekly new consumer subscribers vs ChatGPT Image Credits:TechCrunch

To recap why consumers may have become so much more aware of Claude since January: Anthropic released several Super Bowl commercials that mocked ChatGPT’s decision to show ads to its users — and promised Claude would never do the same. The spots were funny and effective (and also got under the skin of OpenAI CEO Sam Altman).

But the bigger hullabaloo began in late January when multiple media sites, including the Wall Street Journal and Axios, began reporting on a deepening feud between Anthropic and the DoD. At its core, the dispute was about what the military could and couldn’t do with Anthropic’s AI.

Anthropic refused to allow the DoD to use its AI models for lethal autonomous operations (AI potentially killing people) or mass surveillance of American citizens. That beef grew increasingly public, with Anthropic’s CEO Dario Amodei issuing a making firm public statement on February 26 amid the DoD’s threats to hurt Anthropic’s business by labeling the company a supply risk. Which the DoD did. Lawsuits are now flying, although a federal judge this week temporarily blocked the department’s designation.

New user growth climbed sharply during this period. The increase is especially pronounced between those late January media reports and Amodei’s statement on February 26.

Claude new consumer users, six months, 09-02Image Credits:TechCrunch

Beyond the drama, Claude Code and Claude Cowork — developer and productivity tools released in January — have been drivers of subscriptions. The Computer Use feature, released this week, has also sparked a surge, Anthropic tells TechCrunch. That feature allows Claude to navigate a computer independently — clicking, scrolling, and taking actions on its own. It works with Dispatch, which lets users assign tasks from their phones. These features are not available to free-tier users.

Still, for all of Anthropic’s growth among U.S. consumers willing to pay for AI, Claude remains a long way behind ChatGPT.

While OpenAI’s uninstalls spiked immediately after it announced a deal with the DoD — a move that stood in contrast to Anthropic’s safety stand — Indagari’s data shows that OpenAI is still gaining new paid subscribers at a rapid rate and remains the biggest consumer AI platform of them all.