Anthropic just edged past OpenAI in paid business adoption for the first time, according to TechCrunch AI citing the latest AI Index from fintech firm Ramp. The monthly index, built from expense data across more than 50,000 Ramp customers, shows 34.4% of participating businesses now pay for Anthropic services, compared with 32.3% paying for OpenAI. That flip ends a long stretch of OpenAI sitting comfortably on top.
This is a meaningful shift, even with the usual caveats about sample bias. Ramp’s customer base skews toward finance, tech, and professional services, so it isn’t a perfect mirror of the global market. But the trend isn’t isolated. OpenRouter’s developer leaderboard, which pulls from a different slice of users, last placed OpenAI ahead of Anthropic back in December 2025. Two independent vantage points telling the same story makes it harder to dismiss as noise.
The 12-month climb
The pace of change is what stands out. Per Ramp’s figures cited by TechCrunch AI, Anthropic’s adoption among surveyed businesses went from 9% in May 2025 to 34.4% this month. That’s a 26 percentage point jump in a year. Over the same window, OpenAI’s share slipped by 1%, while overall business AI adoption grew by 9%. Anthropic isn’t just riding the tide. It’s pulling share directly.
Why this happened
Ramp economist Ara Kharazian credits a focused playbook. “Anthropic has already been in the lead amongst the high adoption groups like finance, tech, professional services,” he told TechCrunch AI. “It’s across the other firms where OpenAI still has a lead, but that has been shrinking over the past couple of months.”
His read on the strategy:
- Start with a deeply technical customer base (developers, engineers, coding workflows)
- Nail execution for those users before broadening
- Expand outward with horizontal tools like Cowork
“What Anthropic did worked really well,” Kharazian said. The Claude coding momentum, enterprise deals, and Claude Code adoption all line up with that pattern.
What it means for the industry
A few things worth flagging:
- Procurement is moving. Buyers who picked OpenAI by default 18 months ago are now running side-by-side evaluations and switching where Claude performs better, especially on long-context reasoning and code.
- The two-horse race is real. Together Anthropic and OpenAI account for the vast majority of paid AI spend in this sample. Google, Meta, and the rest are still fighting for the rest of the pie.
- Vertical depth beats breadth, at least early. Anthropic’s win came from going narrow first. That’s a useful signal for any AI company trying to figure out a wedge.
Kharazian himself is cautious about whether the lead holds. Enterprise AI is still volatile, model launches reshuffle the deck monthly, and OpenAI has plenty of ammunition. But for now, the scoreboard reads differently than it has in years.
For practitioners, the practical takeaway: if your stack is OpenAI-only, it’s probably time to benchmark Claude on your actual workloads. The market is voting with budgets, and the budgets are talking. Full breakdown and Kharazian’s blog post at the original source.