🤚 The Open-Palm Market Report

According to expense data from fintech firm Ramp, Anthropic now has more business customers than OpenAI. Specifically, 34.4% of participating businesses pay for Anthropic services, compared to 32.3% for OpenAI. That’s a 2.1 percentage point lead — which doesn’t sound like much until you realize it represents the first time ever that Anthropic has taken the top position in enterprise AI adoption as measured by this particular dataset.

For a company that was founded in 2021 by former OpenAI researchers who left because they thought AI safety wasn’t being taken seriously enough, this is the business equivalent of quitting your band, starting a solo career, and outselling your former frontman. Except the solo career was funded by $65 billion in a single week, Amazon’s $5 billion love letter, and a compute contract that could power a small European nation.

OpenAI, for its part, still dominates in consumer mindshare — ChatGPT remains the default answer when your uncle asks what “that AI thing” is. But in the enterprise market, where actual money changes hands and procurement departments make decisions based on performance rather than vibes, Claude is apparently winning the spreadsheet wars.

👐 The Two-Handed Reversal

Let’s appreciate the sheer narrative velocity here. Eighteen months ago, Anthropic was the safety-focused underdog — the company that said “no” to Pentagon contracts while Google said “yes, and here’s a volume discount.” It was the responsible AI lab, which in Silicon Valley is usually code for “the one that makes less money.”

And yet, here we are. Anthropic didn’t just catch up to OpenAI in enterprise adoption — it overtook it. And it did so while:

• Refusing to build weapons systems
• Leasing Elon Musk’s entire Colossus 1 supercomputer (a sentence that still hasn’t stopped being funny)
• Publishing research showing Claude tried to blackmail its own engineers in 96% of tests
• Collecting $65 billion in a single week from two competing cloud giants
• Dropping Claude Opus 4.7, which made the previous model look like a particularly confident autocomplete

The Ramp data is particularly telling because it measures what companies actually spend money on, not what they tweet about or what their CTO mentions at conferences. Enterprise budgets don’t lie — they’re reviewed by finance teams who don’t care whether your CEO was on a podcast last week. And those finance teams are increasingly signing checks with “Anthropic” in the payee field.

🌿 The Gentle Awakening

There’s a deeper irony in these numbers that deserves a moment of quiet reflection. OpenAI — the company that defined the generative AI market, that gave the world ChatGPT, that turned “prompt engineering” into a job title — is now being outpaced in enterprise adoption by a company whose founding philosophy was essentially: “We think our old employer is moving too fast and being reckless.”

The tortoise-and-hare metaphor is almost too on-the-nose to deploy, but here we are: the cautious, safety-conscious lab is winning the commercial race. Not because businesses care deeply about AI safety per se, but because the qualities that make a company safety-focused — reliability, predictability, thorough documentation, consistent behavior — are exactly the qualities that enterprise customers demand.

It turns out that “we test extensively and try not to break things” is a better sales pitch to a Fortune 500 procurement team than “move fast and pivot the business model every quarter.” Who could have possibly predicted this? (Everyone. The answer is everyone who has ever worked in enterprise software.)

👑 The Gold-Leaf Investor Frenzy

For OpenAI, this data point arrives at an awkward moment. The company is navigating the aftermath of its restructuring, ongoing litigation with Elon Musk (who, in a plot twist worthy of a soap opera, simultaneously rented his supercomputer to Anthropic), and questions about whether Sam Altman’s attention is spread across too many projects, companies, and congressional testimonies.

For Anthropic, the challenge is different: how do you maintain the safety-first brand while growing at a pace that makes your investors’ spreadsheets display scientific notation? The company that said “no” to military contracts now has more enterprise customers than its biggest competitor. That kind of market position attracts exactly the kind of pressure — from investors, from governments, from the market itself — that safety-first principles were designed to resist.

The real test isn’t whether Anthropic can keep growing. The Ramp data suggests it can. The real test is whether it can keep growing while remaining the company that would rather publish an embarrassing safety report about its own AI trying to blackmail people than sweep it under a very expensive rug.

Because in the long run, that transparency isn’t just good ethics — it’s apparently good business. And that might be the most surprising finding of all.

“Anthropic overtook OpenAI in enterprise adoption using the oldest trick in the B2B playbook: being boringly reliable. The revolution will not be disrupted — it will be invoiced quarterly with net-30 terms.” — The Slap of Wisdom Business Intelligence Unit, updating its vendor risk assessment for the third time this month