The race between OpenAI and Anthropic in the enterprise customer segment is showing signs of reversing, according to data released by Ramp, a corporate credit card and spend management company.
The data covers more than 70,000 U.S. businesses paying through Ramp’s products. As of July, Anthropic held nearly 44% market share among this customer group, compared to nearly 40% for OpenAI. But the notable takeaway is not the current rankings, but the speed: according to Ramp’s chief economist, OpenAI has been growing significantly faster than Anthropic since the beginning of Q3.
The reason pointed out is OpenAI’s GPT-5.6 Sol model, which has been favored by developers. Conversely, Anthropic’s Fable 5 lineup—a higher-tier offering—has been described as disappointing in both adoption and pricing, particularly after Anthropic warned that Fable users are required to allow 30-day data retention. Note that these observations come from an outside observer, not official data from either company.
Even so, this needs to be kept in perspective: Ramp’s data skews toward small and medium-sized tech companies, excluding large enterprise clients using American Express spend management tools. This is a market indicator, not the full picture.
There is another noteworthy signal: the share of businesses paying for AI in Ramp’s sample has grown steadily, reaching nearly 56% in July. The market is expanding, and both companies are benefiting.
So what: The gap between OpenAI and Anthropic in the enterprise space is not as “sticky” as it appears—customers are readily switching back and forth with each new model release. With both companies approaching an IPO, the true “loyalty” of enterprise AI spending is a crucial question for investors.