Anthropic's Claude claws its way towards the top of the AI market - theregister.com
Anthropic is eating OpenAI's lunch in the business market — and the numbers explain exactly how it happened.
According to Ramp's AI Index for March 2026, Anthropic added 4.9 percentage points of business subscription share in February, its largest monthly gain since the fintech firm started tracking AI adoption. OpenAI lost 1.5 percentage points in the same period — the steepest single-month decline for any AI model company in Ramp's dataset. The two companies now sit at 24.4% and 34.4% of business subscriptions respectively, with Anthropic having crossed the threshold of one in four businesses on Ramp paying for Claude. A year ago, that figure was one in 25.
The trend is even sharper among first-time AI buyers. Businesses selecting AI services for the first time now choose Anthropic roughly 70% of the time when given a choice between it and OpenAI, according to Ramp economist Ara Kharazarian. "It's a complete reversal of the trends we observed in 2025, when OpenAI adoption accelerated faster than any other model company," Kharazarian wrote.
What changed? The Pentagon helped.
Anthropic's public break with the Defense Department — refusing to remove model guardrails to make Claude more amenable to military applications, then publishing a pointed rebuttal when Washington designated it a supply chain risk to national security — appears to have burnished its brand among a specific class of customer: people who pay $20 or $200 per month for AI access and notice such things. Katy Perry announced she switched to Claude in the wake of the dispute. Senator Brian Schatz, a prominent tech policy voice, said the same. Ramp's data shows those moments correlating with a surge in Claude installs and ChatGPT uninstalls, per app tracking firm Sensor Tower.
OpenAI's decision to work with the Pentagon and CEO Sam Altman's acknowledgment that the company handled the situation poorly didn't help. Neither did the ads in ChatGPT.
There's a deeper dynamic here that Kharazarian flags as genuinely unusual in enterprise software: Anthropic is charging more for what analysts describe as roughly equivalent performance. Claude Code and OpenAI's Codex are comparable products; on certain benchmarks Codex arguably wins, and it's cheaper. And yet Anthropic can't meet demand — every plan still has usage limits and rate caps because it lacks compute to serve all the revenue it's turning away. In most enterprise markets, when two products perform similarly, the cheaper one wins. Anthropic is winning anyway.
Maybe the moat isn't benchmarks or pricing. Maybe it's culture. Maybe Anthropic has become — as Kharazarian puts it — cool.
"Is this the moat in AI?" he asks. "Not benchmarks and pricing but whether your model is the one that culturally aware people want to use? There might be a world in which choosing between OpenAI and Anthropic becomes less like an enterprise procurement decision and more like the green bubble/blue bubble distinction in iMessage: a signal of identity as much as a choice of technology."
That sounds absurd for an enterprise software category. But Ramp's data suggests something like it is already happening.
Anthropic's CFO Krishna Rao put a concrete number on the commercial trajectory in a court filing from March 8, 2026: the company has won more than $5 billion in revenue since entering the commercial market. The company's annual revenue run rate stands at $14 billion, according to Anthropic's own announcement accompanying its $30 billion Series G raise at a $380 billion post-money valuation — a figure that was itself the subject of the BMG lawsuit, which alleged that valuation was built partly on copyrighted content.
OpenAI, meanwhile, is reportedly revising its strategy to focus on selling AI to businesses and software developers — precisely the markets where Anthropic is pulling ahead.
The question is whether Anthropic can sustain the positioning. Being the "responsible AI" company is a tricky brand to maintain when your models are reportedly being used in US military operations. And compute constraints mean Anthropic is actively leaving revenue on the table, which limits how fast it can compound its lead.
But for now, the momentum is real, and the data is unambiguous: the AI company that positioned itself as the grown-up in the room is winning the customers who want to believe it.