Anthropic has had a remarkable few weeks. The AI company ended May by surpassing OpenAI in business market share for the first time, raised $65 billion at a $961 billion valuation, and quietly filed IPO paperwork on the back of what was reportedly its first profitable quarter. Then the Trump administration stepped in.
On Friday, the White House sent a letter demanding Anthropic ban non-Americans, including its own employees, from accessing its most advanced models: the limited-release Mythos 5 and Fable 5, a version released to the public just three days earlier. The result was that Anthropic had to pull both models from the market entirely. The administration cited an obscure export control directive, though the exact reasoning is still unclear. Word spread that hackers had found ways around Fable 5’s safety guardrails, which were meant to block access to Mythos-level capabilities. Mythos is so effective at finding security flaws in code that Anthropic itself described it as dangerous and kept its release tightly controlled.
This is not the first clash. In March, the Trump administration declared Anthropic a supply-chain risk after the company refused to let the government use its models for mass surveillance of Americans or fully autonomous weapons. That decision didn’t hurt Anthropic’s sales. According to TechCrunch, the data suggests this latest conflict might not hurt them either.
Ara Kharazian, lead economist at Ramp, the financial platform that compiled the business spending data, put it bluntly: “If anything, it’ll probably boost them. Anthropic’s best month on record, as far as business adoption, was the month that the Department of Defense labeled them a supply-chain risk. There’s a lot of aura that comes with your model specifically being named too dangerous to use.”
Ramp’s data comes from more than 70,000 businesses using its platform. The numbers paint a clear picture of where corporate AI spending is going right now. In May, Anthropic’s share of AI subscriptions paid by businesses rose 2.5 percentage points to 41%, overtaking OpenAI, which held 39.5% and was essentially flat month over month. OpenAI still leads by a wide margin in overall consumer usage, according to Sensor Tower, but in the business segment, Anthropic has pulled ahead.
Subscriptions are only part of the story. The bigger slice of corporate AI spending goes to API calls, which cover token usage for tasks like coding. Anthropic’s Claude Code has built a strong reputation as a serious tool for software development. When Ramp can identify which specific models businesses are using, which is possible in about one-third of transactions, the answer is mostly various versions of Claude Opus, especially the newer releases. Opus is the model that came before Mythos and is still widely available. Anthropic released a new version, Opus 4.8, in late May.
That context matters here. Mythos had only been available to a limited group since April. Fable 5 was live for just three days before getting pulled. So despite the drama around both models, they had barely reached the market before the shutdown. The bulk of business spending was already flowing toward Opus, and that hasn’t changed.
Ramp’s data isn’t detailed enough to show exactly how much revenue Anthropic loses by pulling Mythos and Fable 5. But the broader trend is clear: the models that businesses are actually paying for are more popular than ever. The controversy may be adding to the mystique rather than undermining trust.
The IPO question is harder to answer. Public market investors tend to get nervous around companies in active disputes with the government, and Anthropic is now in its second major clash with the White House in three months. Whether that cools appetite for a public offering remains to be seen. What the data does show is that, at least among the businesses already using Anthropic’s tools, the government pressure hasn’t moved the needle in the direction Washington might have hoped.
