OpenAI Is Catching Up to Anthropic Among U.S. Businesses Based on New Spending Data
Ramp's spending data reveals OpenAI is gaining on Anthropic among U.S. businesses, with GPT-5.6 Sol driving developer adoption while Anthropic's Fable model struggles post-regulatory backlash.
By Memujo Editorial
OpenAI is making a comeback among U.S. businesses, according to new spending data from Ramp, the corporate credit card and expense management platform.
The numbers, published by Ramp economist Ara Kharazian on August 20, 2026, show that OpenAI has started closing a gap it had opened against Anthropic among Ramp's paying business customers, with more than 70,000 American companies that collectively spend billions via the platform's bill pay and corporate card products.
What the data shows
The picture has shifted several times over the past few months.
In early 2026, OpenAI was the clear leader among Ramp's business users. But by May, Anthropic had flipped the lead: 41% of AI spend went to Anthropic versus 39% to OpenAI. Anthropic held that advantage through July, reaching nearly 44% compared to OpenAI's 40%.
Then Q3 came along and OpenAI started growing faster.
As of the latest data through early August, OpenAI's momentum has picked up enough to narrow the gap significantly. Ramp declined to provide exact dollar figures or updated market share percentages, sharing only directionally. But the trend is unmistakable: Anthropic has not locked in a permanent lead, and enterprise AI spending is more volatile than the headlines from either camp would suggest.
GPT-5.6 Sol drives developer adoption
Kharazian attributed the shift to OpenAI's GPT-5.6 Sol release, calling it "really good, increasingly the choice for developers." The model, part of OpenAI's Sol/Terra product family launched in mid-2026, appears to have won over engineering teams that had been migrating to Anthropic's tools.
The contrast with Anthropic's situation is notable. Kharazian pointed to Fable 5, Anthropic's higher-end model tier, saying it "disappointed both in adoption and real-world application given price + data retention requirements imposed by regulators."
Anthropic had sparked backlash earlier this year when it informed Fable users that their data would be retained for 30 days, a policy driven by regulatory requirements but unpopular with enterprise customers who had assumed their inputs would be treated as confidential. The combination of higher pricing and reduced data privacy was enough to push some teams back toward OpenAI's offerings.
The broader picture: enterprise AI spending is growing and volatile
Ramp's data also reveals two trends that go beyond the OpenAI vs. Anthropic rivalry.
First, the market overall is expanding. The share of Ramp's customer companies that spend on AI products has climbed steadily: from just over 50% in March to nearly 56% by July. More businesses are experimenting with AI tools, regardless of which lab they choose.
Second, that expansion masks significant churn. The fact that Anthropic briefly overtook OpenAI and then lost the lead back suggests that enterprise customers are not locked into a single provider. They switch based on model quality, pricing, and policy changes, a reality that should concern both companies as they approach potential IPOs and investors evaluate their growth trajectories.
What this means
The OpenAI-Anthropic rivalry is often framed as a battle for permanent dominance. But Ramp's data, while limited to its own customer base of tech-skewing businesses, shows something more nuanced: the enterprise AI market is still being won and lost on each new model release, and customer loyalty is thin.
For GPT-5.6 Sol, that volatility has been positive. OpenAI's ability to re-engage businesses it had lost underscores that the frontier model race is far from over. Anthropic still leads on coding benchmarks and has carved out a strong position in developer tools. But as OpenAI's latest releases gain traction, the gap between the two labs may be closing faster than analysts expected.
Whether this trend holds through the rest of Q3 remains to be seen. But one thing is clear: enterprise AI spending is not a zero-sum game, and the competition between the two labs is driving both innovation and churn at a pace that benefits customers, for now.
- #openai
- #anthropic
- #enterprise
- #business
- #gpt-5