The ChatGPT company just got outspent by the company that said it would never chase revenue.
The Summary
- Anthropic hit 34.4% business adoption in April, overtaking OpenAI's 32.3%, according to Ramp's AI Index tracking billions in corporate AI spending from 50,000+ US businesses.
- This marks the first time Anthropic has more verified business customers than OpenAI, a complete reversal from January when OpenAI dominated enterprise adoption.
- The shift was driven by surging adoption of Claude Code, particularly in software development, research, finance, and customer support.
- The AI market remains extremely volatile, with leadership positions that seemed permanent six months ago now up for grabs.
The Signal
In January 2025, OpenAI was the clear enterprise AI leader, dominating corporate spending across every major business function. Four months later, that lead evaporated. The company that launched the ChatGPT moment just lost the enterprise race to a competitor founded by former OpenAI employees who explicitly positioned themselves as the "safe AI" alternative.
Ramp's AI Index tracks corporate card and bill-paying activity from more than 50,000 US businesses, capturing billions in monthly AI service spending. It's not comprehensive, but it's the best real-time proxy we have for where enterprises are actually putting their money, not where they tell TechCrunch they're "exploring partnerships."
"Anthropic has crossed a symbolic threshold in the AI race: businesses are now spending more on Anthropic than OpenAI for the first time."
The driver? Claude Code. While OpenAI was busy shipping ChatGPT Enterprise and fighting regulatory battles, Anthropic quietly built the tool developers actually wanted: an AI that could write, review, and refactor code without hallucinating itself into production incidents. Software development is the highest-value use case in enterprise AI. Anthropic won it.
This isn't just a product story. It's a go-to-market story. OpenAI built for consumers first, then tried to enterprise-ify the product. Anthropic built for researchers and developers from day one, the people who control enterprise AI budgets. They didn't need a sales team. They needed a product that shipped.
Key adoption drivers:
- Claude Code's reliability in production environments
- Developer trust in Anthropic's technical rigor
- Enterprises moving from experimentation to deployment, where accuracy matters more than brand recognition
The volatility here is the real signal. Six months ago, the narrative was that OpenAI's lead was insurmountable, that being first to market with GPT-3 and ChatGPT created an unbreakable moat. That moat just got breached. In Web4, adoption curves move faster than anyone's comfortable with. The AI stack hasn't ossified yet. There's no Windows, no iPhone. Just a bunch of APIs fighting for developer mindshare quarter by quarter.
Ramp's economist noted the market remains "volatile," which is economist-speak for "nobody knows who's winning next month." OpenAI still has consumer dominance, the ChatGPT brand, and Microsoft's distribution engine. But enterprises don't care about brand. They care about what doesn't break in production.
The Implication
If you're building on foundation models, this is your reminder that the ground keeps shifting. The "pick the market leader and build on top" strategy doesn't work when market leadership changes every quarter. Multimodel architectures, abstraction layers, and vendor-agnostic tooling aren't paranoia. They're the only sane approach to a market where April's winner is January's second place.
Watch where the enterprise budgets go next. OpenAI will respond. They have to. But Anthropic just proved that in the agent economy, the best product for people who ship code wins, regardless of who has the most famous chatbot.