In May 2026, Anthropic quietly overtook OpenAI as the most widely paid-for AI provider among US businesses. Three months later, OpenAI is fighting back.

New data from Ramp, the corporate card and spend management platform used by more than 70,000 US businesses, shows that Anthropic still leads with 43.5% of companies subscribing to or purchasing its models, while OpenAI sits at 39.7%. But the direction of travel has shifted. OpenAI is posting a quarter-over-quarter growth rate of 82 compared to Anthropic’s 76, meaning the gap is narrowing.

The data was reported by TechCrunch on August 20, 2026, and it tells a story that matters if you’re running a business and trying to decide which AI vendor to bet on.

How We Got Here

For most of 2025, OpenAI was the default choice for businesses deploying AI. ChatGPT had name recognition, GPT-4 powered most enterprise tools, and OpenAI’s API was the path of least resistance.

Then Claude happened. Anthropic’s Claude 3 and 3.5 models earned a reputation for being better at complex reasoning, following instructions precisely, and writing in a professional tone. Developers and businesses started switching, and by May 2026, Anthropic had taken the lead among Ramp’s tracked businesses.

OpenAI responded with an aggressive model release and pricing strategy. GPT-5.6 Sol, Terra, and Luna launched with significant capability improvements. Prices dropped repeatedly. In the past month alone, OpenAI cut Sol’s API pricing three times, with the most recent cut reducing output costs by 33% through November 2026. ChatGPT Work matured into a credible enterprise product. The result: OpenAI is growing again.

What the Numbers Actually Mean

The Ramp data tracks real spending decisions by real companies, which makes it more useful than survey responses or analyst projections. A 43.5% Anthropic adoption rate against 39.7% for OpenAI means these are now roughly equal in business reach. At these scales, a few percentage points of difference matters less than the trend.

What’s more significant is that enterprise AI spending is now genuinely competitive. Businesses are no longer picking one provider and sticking with it. Ramp’s analysis notes that switching costs are falling, meaning companies are reallocating budgets as new models prove their worth. The era of vendor lock-in for AI is getting shorter.

The other striking number: enterprise revenue now makes up more than 40% of OpenAI’s total revenue, and the company expects it to reach parity with consumer revenue by the end of 2026. That’s a fundamental shift in how OpenAI thinks about its business.

The Real Competition Is With Your Existing Workflows

For most business owners, the OpenAI vs. Anthropic question is the wrong one to ask.

Both providers offer capable models. Both are pricing aggressively. Both are shipping new features monthly. The choice between them matters far less than whether you’re actually deploying AI in a way that moves the needle for your business.

The Ramp data reveals something deeper than market share: 40-plus percent of US businesses are paying for AI tools, but that doesn’t tell you how many are getting real value from them. Research consistently shows that adoption is outpacing actual productivity gains. Most companies are still in pilot mode. The gap between paying for AI and scaling it is where the real opportunity sits.

What This Means for Business

If you’re currently using one provider and wondering whether to switch, the data suggests competition is driving genuine improvements at both OpenAI and Anthropic. Prices are falling. Models are getting better. The calculus of which to choose is shifting every quarter.

But there’s a more important question: are you extracting value from whichever model you’re already using? The businesses in the top quartile of AI ROI aren’t winning because they picked the right vendor. They’re winning because they built proper workflows, trained their teams, and connected AI tools to actual business processes.

That’s the same principle behind how Enterprise DNA approaches AI integration for clients. The model choice matters a lot less than the architecture around it. Choosing the right model for each task, building agents that connect to real data, and giving teams the skills to use these tools effectively is what separates companies getting 25% productivity gains from those running the same pilot for the third time.

The OpenAI vs. Anthropic competition is worth watching. The price cuts alone will save businesses real money over the next twelve months. But the more meaningful story is in how quickly your business is moving from “we’re exploring AI” to “AI is running our operations.”

For a practical conversation about where AI can drive the biggest returns for your business, speak with the Enterprise DNA team.