Here's something that should make OpenAI's product team a little uncomfortable: Anthropic's Claude is gaining serious ground with the one customer segment that actually matters for long-term survival—people who voluntarily pull out a credit card every month to pay for an AI subscription.

That's not a vibe. That's transaction data. Credit card analytics firm Indagari, which tracks anonymized spending across roughly 28 million U.S. consumers, shows Claude's paying consumer base and associated revenue climbing steadily month over month through the first half of 2026—up approximately 75% since January alone. To be clear, this isn't measuring free-tier curiosity or API trial accounts. These are people choosing to spend real money on Claude.

Why This Is More Interesting Than It Sounds

The prevailing narrative about Anthropic has always been that it's an enterprise and developer play. Claude Code for the startup crowd, API access for the builders, and—oh yeah—ChatGPT owns the consumer market. That framing made sense for a while. OpenAI had the brand recognition, the Microsoft distribution deal, and a head start measured in years.

But transaction data doesn't care about narratives. And what Indagari's numbers suggest is that Anthropic has quietly developed a paying consumer base that's broader and stickier than the "just developers" story implies. Subscriptions, API tokens, the whole consumer spend picture—it's trending upward and hasn't plateaued.

Worth noting: Indagari's dataset is large enough to be statistically meaningful, but it doesn't tell us Anthropic's absolute revenue or total subscriber count. What it does reveal are relative trends and directional momentum. And the direction is unambiguous.

The Ethics Bump That Kept Bumping

Back in March, Anthropic reportedly saw a notable spike in consumer interest after the company publicly refused to permit its models to be deployed for mass surveillance of Americans or for autonomous weapons systems. That kind of principled stance tends to generate a short-lived media cycle and a brief traffic bump—the AI equivalent of a brand getting a viral moment.

What's more interesting is what happened afterward: the growth didn't stall. The paying consumer numbers kept climbing after that March spike, suggesting Anthropic converted some of that goodwill into actual retained subscribers rather than just curious window shoppers. Whether that's because of the ethics positioning, product quality, or some combination, the retention signal is real.

DataCamp's Search Data Tells a Similar Story

Transaction data is one lens. Search behavior on a learning platform is another, and the picture there is equally striking. DataCamp—an online AI skills education platform with around 20 million users—reports that "Claude" has become the single most searched term on its site in 2026, outranking even the generic term "AI." Let that sink in for a second. People aren't just searching for AI concepts anymore; they're searching specifically for how to use Claude.

The corporate training side still skews heavily toward ChatGPT—businesses doing bulk employee training tend to default to whatever tool their IT department already approved. But among self-directed learners, people who choose their own educational path on their own time, demand for Claude courses is reportedly outpacing ChatGPT by a three-to-one margin. That's not a rounding error. That's a meaningful preference signal from exactly the kind of engaged, motivated users who become long-term paying customers.

What's Actually Driving This?

A few honest hypotheses, none of which are mutually exclusive:

• Model quality at the task level. Claude's models have earned a genuine reputation for long-context handling, nuanced writing, and coding assistance that feels less robotic than some competitors. For users doing substantive work—not just quick queries—that difference is noticeable.
• The Claude Code halo effect. Developer enthusiasm around Claude Code may be spilling into personal subscriptions. Builders who use Claude professionally are paying for it personally too.
• The ethics positioning as product differentiation. Niche? Maybe. But the segment of consumers who explicitly care about AI governance is non-trivial and apparently willing to pay for it.
• ChatGPT fatigue. OpenAI has had a turbulent stretch of product announcements, pricing changes, and public controversies. Some consumers may simply be exploring alternatives—and sticking with what they find.

The Limits of Good News

Let's not get carried away. ChatGPT still commands an enormous lead in raw consumer scale. OpenAI's distribution advantages—default integrations, Microsoft's enterprise reach, brand ubiquity—aren't going away. Anthropic's gains are real, but they're gains from a smaller base, and the gap remains substantial.

There's also the question of whether consumer subscriptions translate meaningfully to Anthropic's bottom line compared to its enterprise API revenue. The economics of consumer AI are notoriously difficult—high compute costs, price-sensitive users, and churn that can spike whenever a competitor releases a flashy new model. Sustained growth is the metric that matters, not a single quarter's trajectory.

Still, the conventional wisdom that Anthropic is purely a B2B developer tool while OpenAI owns the consumer imagination? The data is starting to complicate that story in ways worth watching.