Fewer users, fatter wallets is why Anthropic tops OpenAI in LLM revenue stakes
The Register
Anthropic now edges out OpenAI in LLM revenue despite having way fewer users. Turns out fewer, richer customers beat billions of casual chatters.
Based on reporting by The Register — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
There's a neat little paradox buried in Counterpoint Research's latest numbers on the AI business: the company with the fewest users is making the most money per head, and it's not particularly close. Anthropic pulled 31.4 percent of global LLM revenue in the first quarter of 2026, just ahead of OpenAI's 29 percent. Fine margin. Except Anthropic did it with about 134 million monthly users against OpenAI's roughly 900 million. Divide it out and Anthropic is averaging $16.20 per user per month, compared to OpenAI's $2.20, Microsoft's $5, and Google's $1.10.
That gap tells you something about who's actually paying for AI right now. Anthropic has quietly built what Counterpoint calls a foothold in the high-end professional market, the kind of customers who need Claude to write code or draft contracts and don't blink at the invoice. OpenAI, by contrast, has the scale of a consumer platform but the monetization of one too, spread thin across nearly a billion free and casual users. Popularity and profit, it turns out, are not the same axis.
Meta makes that split almost comically obvious. It's got roughly a billion users touching its AI features and a monetization rate of ten cents per user. Ten cents. And yet Meta just raised its 2026 capex guidance to somewhere between $125 billion and $145 billion, up from an already enormous $115 billion to $135 billion estimate a few months ago. Wall Street's patience showed its limits: even after Meta beat on both revenue and profit for the quarter, the stock dropped 7 percent in after-hours trading once the spending number landed. Engagement, it seems, buys goodwill only up to a point.
Zoom out and the spending picture gets even bigger. Alphabet, Amazon, Microsoft, and Meta combined are projected to spend $725 billion on AI infrastructure this year, versus $410 billion last year — a 77 percent jump that assumes demand and revenue will eventually catch up to the outlay. Meanwhile Baidu is reportedly pulling in $1.30 per user, beating Google and Meta's rates combined, while Tencent and Alibaba run their own hybrid playbooks inside China's market.
Add it all up and you get more than 3.8 billion monthly LLM users worldwide generating roughly $20.7 billion in quarterly revenue. Sounds massive until you remember that most of those billions are using these tools for free, and the real money is coming from a comparatively small slice of paying professionals. The AI boom has a lot of participants. It has far fewer customers.
My take — AI-written commentary, not fact-checked reporting
This is the tell that the AI hype cycle keeps skipping past: user counts make headlines, but Anthropic's numbers show the actual business is being built on a small base of people who'll pay real money for real output. Meta torching $145 billion in capex to chase engagement metrics while monetizing at ten cents a head is the kind of math that only makes sense if you assume infinite investor patience — and after that 7 percent stock drop, I don't think that assumption holds much longer. Scale without a revenue model isn't a moat, it's a bill somebody eventually has to pay.
Read more about this at: The Register