When the user pays out of pocket: the metric that best gives away a product that works

Not all metrics carry the same weight. A free user is cheap to acquire and cheap to lose; a subscriber who re-enters their card month after month is casting a vote that is hard to fake. That is why the analysis published by TechCrunch on June 25, 2026, based on transaction data from the firm Indagari, deserves attention: it points to a rise of nearly 75% since January 2026 in Claude's paying consumers, with sustained month-on-month growth rather than a mere spike.

That nuance—the continuity—is the truly revealing figure. In March 2026 Anthropic generated a sharp surge of attention after refusing, according to the article, to let its models be used for mass surveillance and autonomous weapons. The expected outcome would have been a media high followed by a reversion to the mean. What the analysis describes is the opposite: acquisition held up after the episode, a sign that the branding decision crystallized into structural adoption, not passing curiosity.

The signals converge from other observatories. The educational platform DataCamp says that 'Claude' has become its most searched term, ahead even of the generic 'AI,' with demand for courses that reportedly triples that of ChatGPT among self-learners. When people invest their time in learning to use a tool, they are anticipating that they will use it seriously. Interest in training is usually a leading indicator of professional use.

It is worth, of course, keeping a sense of proportion, and the article itself does so well: data from Sensor Tower confirm that ChatGPT still leads comfortably in absolute terms. OpenAI's more modest percentage growth is explained by basic arithmetic—it is extremely hard to grow fast on an already gigantic base—not by weakness. Claude is closing the gap in dollars billed, but it starts far behind.

The underlying takeaway, at a moment when both OpenAI and Anthropic are approaching, according to TechCrunch, the threshold of going public, is diversification. The dominant perception placed Claude as a tool for developers and businesses via Claude Code; these data suggest a broader retail base than assumed. For a future investor, a business that grows simultaneously in the individual segment and the corporate one—even after the regulatory setback that pulled Mythos 5 and Fable 5 from the market for non-U.S. users—is simply a less fragile business. The consumer AI market is ceasing to be a monologue and starting to be a conversation.