For a long time the story was simple: ChatGPT is synonymous with artificial intelligence, everyone else runs behind it. But new data tells a different story about those who actually pay. The number of paying users and the revenue of Claude, the chatbot by Anthropic, have grown by about 75 percent since January this year.

The figures come from several sources that track real spending, not marketing. One analytics firm processes billions of anonymised card transactions from around 28 million American consumers. According to it, Claude's revenue from ordinary users is slowly closing in on that of ChatGPT - though OpenAI still has far more total subscribers.

Even more interesting is data from an online-learning platform with around 20 million users: „Claude" became the most-searched term on the site, overtaking even the term „AI" itself. Demand for Claude courses jumped 18-fold over the last 30 days. Among independent users, those who decide for themselves what they'll use, interest in Claude beats ChatGPT by a ratio of three to one.

What does this mean? That the battle for artificial intelligence isn't decided by who entered the market first, but by whom users will trust enough to take out their wallets. Anthropic is building a reputation as the „more serious" player - in March its growth spike came after the company refused to let its models be used for surveillance and autonomous weapons. In June, meanwhile, the US government banned some of its models for users outside the US, a move that tied the technology to geopolitics.

Both companies are preparing to go public. When revenue from ordinary users starts spilling from one to the other, investors read that as a signal of who owns the future. Anthropic declined to comment - which in this industry usually means the numbers speak for themselves.