329
Audio & Video Production330
Automation & Workflow218
Software Development248
Marketing & Growth203
AI Infrastructure & MLOps152
Writing & Content Creation204
Data & Analytics128
Customer Support130
Design & Creative153
Sales & Outreach125
Photography & Imaging143
Voice & Speech132
Operations & Admin93
Education & Learning122
Ramp data suggests more US businesses are paying for Anthropic’s AI tools, while OpenAI’s share was flat in March.
In short: New payment data suggests more US businesses are paying for Anthropic’s AI tools, narrowing the gap with OpenAI.
Data from payments company Ramp says nearly one in three US businesses paid for Anthropic’s tools in March. Ramp says this was up by more than 6 percentage points from the month before.
In the same month, Ramp says OpenAI’s business adoption was flat at 35 percent. Ramp’s numbers are based on card and invoice spending it can see across about 50,000 customers and $100 billion in annual spending.
The Financial Times reports that Anthropic’s growth is being driven by demand for its Claude Code products and related add-ons, which help automate parts of office work (like a helpful assistant that can handle repetitive tasks). Ramp also said Anthropic’s tools are now more widely used than OpenAI’s across information, financial, and professional services, and are growing in other areas like construction and hospitality.
OpenAI disputed the Ramp data. OpenAI said it did not recognise the figures cited, and pointed to its own metrics, including 3 million weekly users for its Codex coding agent and more than 15 billion “tokens” processed per minute (tokens are small chunks of text that AI systems read and write, like counting words in pieces).
Anthropic said it does not comment on third-party data. The FT also noted that a recent dispute involving the US Defense Department did not appear to slow Anthropic’s business.
If this pattern continues, more companies may split their spending across multiple AI providers, instead of standardising on just one. It is also worth watching how much Ramp’s view of spending matches large enterprise contracts, which OpenAI has argued are not captured well by this kind of dataset.
Source: Financial Times