Anthropic has, for the first time, surpassed OpenAI in paying business customers, according to the May 2026 release of Ramp's AI Index published this week. The shift caps a year of rapid enterprise gains for the Claude maker and lands as Anthropic is reportedly raising new capital at a valuation approaching $950 billion.
The Numbers
In April 2026, 34.4% of businesses tracked by Ramp were paying for Anthropic services, up 3.8 percentage points month-over-month. OpenAI's share fell 2.9 points to 32.3%. Overall AI adoption across the Ramp dataset edged up to 50.6%. Twelve months earlier, just 9% of those businesses were paying Anthropic — meaning the company has roughly quadrupled its paying base in a year, while OpenAI's business adoption grew by only 0.3 percentage points over the same period.
Ramp lead economist Ara Kharazian framed the result as a sign of unusually low vendor stickiness in the AI software market. "The truth is we have never seen a software industry as dynamic, where newcomers can disrupt market leaders in a matter of months, and where the pace of development overrides the typical forces of vendor stickiness," he said.
Claude Code Is Doing the Heavy Lifting
Most coverage of the index — including reports from TechCrunch, Axios, and VentureBeat — pins the swing on a single product: Claude Code, Anthropic's agentic coding tool, which the company describes as the fastest-growing product in its history. Anthropic has also been pushing downmarket, recently launching Claude for Small Business with prebuilt integrations including QuickBooks, Canva, DocuSign, HubSpot, and PayPal, alongside a planned ten-city promotional tour beginning in Chicago.
The Catch
The milestone arrives with caveats. VentureBeat and other outlets flagged that Anthropic's lead is fragile: heavy Claude Code usage produces eye-watering bills, with reports of individual engineers running monthly API spend between $500 and $2,000 and Uber's CTO acknowledging the ride-hailing company burned through its 2026 AI budget in roughly four months — much of it on Claude Code and Cursor.
That creates an awkward incentive structure. Cheaper models from Google, OpenAI, and a rising tier of Chinese open-weight contenders could erode share if buyers decide that frontier-tier coding agents are overkill for routine tasks.
Why It Matters
For most of the generative AI era, OpenAI's lead among paying enterprises was treated as a near-given. Ramp's data does not capture the full market — it skews toward the kinds of midmarket and tech-forward firms that use its corporate spend platform — but it is one of the few near-real-time signals of where business dollars are actually moving. The fact that Anthropic has crossed the line, even narrowly, validates a thesis that enterprises will pick whichever lab is best for the workload they care about most, and right now that workload is software engineering.
