For the first time in the brief but ferociously contested history of commercial AI, Anthropic has more paying business customers than OpenAI — and the data comes from one of the most credible sources imaginable: real corporate card transactions.
Ramp, the corporate spend management platform used by more than 50,000 businesses, publishes a monthly AI Index compiled from actual payment data. Its May 2026 edition, released on May 13, confirmed what the trend lines had been signaling for months: Anthropic's business adoption rate climbed 3.8 percentage points in April to reach 34.4%, while OpenAI's fell 2.9 points to 32.3%. The gap — narrow as it is — represents a genuine inflection point. OpenAI had held the top position since the index launched a year ago.
"Our results mark a stunning reversal in the competitive market dynamics for AI model providers," wrote Ara Kharazian, Lead Economist at Ramp, in the index report. The dataset, drawn from corporate card and invoice-based payments, is broadly regarded as a real-money proxy for enterprise intent — not survey responses, not developer registrations, but actual procurement.
A Year of Quadrupling
The speed of Anthropic's ascent is difficult to overstate. In May 2025, just 9% of businesses on Ramp were paying for Anthropic products. By April 2026, that figure had reached 34.4% — an increase of roughly 280% in twelve months. OpenAI's share, by contrast, grew by only 0.3% over the same period.
The April 2026 Ramp AI Index had already signaled the crossover was imminent. At that point, Anthropic stood at 30.6% to OpenAI's 35.2% — a gap of 4.6 percentage points, down from 11 points just two months earlier in February. Kharazian had publicly predicted Anthropic would overtake OpenAI within two months. It happened on schedule.
The data also reflects a broader market expansion: overall AI adoption among Ramp's business customers crossed 50% for the first time in March 2026, reaching 50.4%, and edged up another 0.2 points to 50.6% in April.
How Anthropic Did It
Ramp's earlier reporting points to a deliberate sequencing strategy. Anthropic concentrated first on the highest-adoption, most technically sophisticated segments — finance, software, and professional services — before expanding outward. Among venture capital-backed companies, which show an 80% AI adoption rate overall, Anthropic already leads OpenAI 66% to 59%. Among VC-backed firms in the information sector, the gap is even wider: 63% to 54%.
"What Anthropic did worked really well," Kharazian told TechCrunch. "Which was — start with a very technical customer base, focus on their needs, really succeed in execution and then start broadening out through tools like Cowork."
The strategy mirrors how enterprise software giants have historically conquered markets: win the sophisticated early adopters who have the highest standards and the most influence, then let their adoption patterns cascade downward into the broader market. Ramp's own longitudinal data supports this mechanism. As Kharazian noted in the April index: "The pattern in our data is consistent: what early adopters do today, the broader market does a few months later."
OpenRouter's rankings provide independent corroboration. On that leaderboard — which samples a different pool of users via API traffic — OpenAI had not ranked above Anthropic since December 2025.
Three Clouds on the Horizon
Kharazian is explicit that the data should not be read as a declaration of durable dominance. He outlined three structural headwinds that could erode Anthropic's lead.
First, a cost-alignment problem. Anthropic earns more revenue when customers consume more tokens, creating an incentive to push users toward its most expensive frontier models even when cheaper alternatives would suffice. Ramp's own data shows AI token spend per firm is up 13x since January 2025. The Information reported that Uber's CTO announced the company had already blown through its entire 2026 AI budget — a high-profile cautionary tale that is prompting finance teams to scrutinize their AI spend. A new deal between Anthropic and SpaceX to expand compute capacity may resolve supply-side constraints in the short term, but the cost pressure on customers remains.
Second, recent service reliability issues. Claude experienced a wave of outages, rate limits, and user complaints about degraded output quality in April. Anthropic reset usage limits for all users in response and acknowledged the problems publicly. The company's new SpaceX compute agreement is expected to address the underlying capacity crunch.
Third, a product roadmap concern. Ramp Econ Lab analysis found that Anthropic's latest model update would triple token costs for any prompt containing an image — a move that compounds rather than alleviates the cost problem at a moment when enterprise customers are already watching their AI budgets closely.
Meanwhile, the fastest-growing vendors on Ramp's platform in May 2026 included four AI inference platforms — Fireworks AI, fal AI, DeepInfra, and Vast.ai — each pitching access to a broader, cheaper mix of models than what Anthropic or OpenAI offer directly. The market is beginning to fragment. Enterprises are routing different workloads to different models based on price and capability, rather than committing wholesale to a single provider.
What It Means
The Ramp data carries real weight precisely because it is built from payments, not preferences. A company counted in the Anthropic column wrote a check. That said, market share in AI is more volatile than in virtually any other software category. Kharazian himself has been careful to note this dynamic from the start: "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."
For Anthropic, the achievement validates a strategy built on earning the trust of demanding enterprise buyers through consistent model quality and thoughtful API design. The company now faces the harder challenge: holding that position against an OpenAI that retains a substantial installed base, a dominant consumer brand, and a product pipeline that includes Codex — a coding tool that Ramp's data suggests is gaining developer traction at competitive price points.
The throne has changed hands. Whether it stays that way is a question the June index will begin to answer.
---
Sources: Ramp AI Index (May 2026), Ramp AI Index (April 2026), Ramp Top SaaS Vendors (May 2026), TechCrunch
"What Anthropic did worked really well — start with a very technical customer base, focus on their needs, really succeed in execution and then start broadening out through tools like Cowork."— Ara Kharazian, Lead Economist, Ramp