Ramp data: 30.6% of US businesses paid for Anthropic's tools in March, up from 24.4% in February; OpenAI's US business adoption remained nearly flat MoM at ~35%
Divergence reflects company's recent rapid growth owing to strong interest in its Claude Code products
Context & Ripple Effects
Ramp’s earlier data showed Anthropic taking roughly 73% of spending by companies buying AI tools for the first time, after a January split that had been even with OpenAI. The March adoption change extends that new-buyer spending lead into a broader installed-business metric.
The comparison matters because Ramp had previously recorded OpenAI subscriptions across 32.4% of its US business-card users in April 2025. Anthropic’s rise toward OpenAI’s current level signals a more competitive enterprise procurement market rather than a single-provider default.
First-order effects
- Anthropic gains materially more paid US business accounts in Ramp’s sample, while OpenAI retains the larger base but shows little month-to-month movement.
- Interest in Claude Code appears to translate product demand into paid company adoption, strengthening Anthropic’s enterprise distribution position.
Second-order effects
- OpenAI faces greater pressure to defend business accounts through product differentiation, packaging, and sales execution as Anthropic closes the adoption gap.
- Corporate AI buyers are more likely to evaluate providers by task-specific utility—especially coding—rather than standardize solely on the incumbent subscription.
Third-order effects
- If this pattern persists, enterprise generative-AI spending could segment by workflow and model provider, making multi-vendor procurement more durable than a winner-take-most subscription market.
- The signal is directional rather than comprehensive: Ramp measures its card-spending customer base, but its data points to a procurement phase in which adoption momentum can shift quickly between frontier-model vendors.
The trend: Enterprise AI buying is moving from early broad experimentation toward competitive, workflow-led selection of managed model providers.