Anthropic is having a great month.
The AI Institute ended May surpassing OpenAI in business spending market share for the first time, Ramp said. After raising $65 billion at a $965 billion valuation (also surpassing OpenAI) in late May, it reportedly filed confidential filings for an IPO by June, fueling its first profitable quarter.
Then on Friday, the Trump administration renewed its war on model manufacturers, sending a letter demanding that non-Americans, including Anthropic’s employees, be barred from accessing the cutting-edge models. A limited edition Mythos 5 and a more secure version of Mythos called Fable 5 that was released to the public three days ago.
This led Anthropic to completely withdraw its latest, powerful model from the market.
Although the White House invoked vague export control guidelines when issuing the ban, the exact cause remains unclear. There were rumors floating around that hackers had easily bypassed Fable 5’s guardrails meant to block access to Mythos’ features. This model is so good at finding security flaws in software code that Anthropic itself touted it as dangerous and restricted its public release.
This new drama comes after Anthropic became famous for refusing to allow the government to use its models for mass surveillance of Americans and for fully autonomous weapons. As a result, in March the Trump administration declared the company to be at supply chain risk.
That hasn’t hindered Anthropic’s sales to corporations. Ramp’s data shows just the opposite. Ironically, the recent feud with the Trump administration, which appears to validate the clamor for the mythical power of Mythos, may help Anthropic rather than harm it, according to Ara Kharazian, chief economist at Ramp. Karajian is the person who compiled corporate spending AI data.
“It will probably improve it anyway,” Kharazian told TechCrunch. “As far as business adoption, Anthropic’s best month on record was the month the Department of Defense classified it as supply chain risk. There’s a lot of aura that comes from a model specifically labeled too risky for use.”
Ramp’s data isn’t detailed enough to confirm how much of a financial hit the company will take by pulling Mythos and Fable 5 off the market.
Nonetheless, data collected from more than 70,000 businesses that use the platform shows customers are using Anthropic’s Opus model a lot, and business usage is growing.
For example, Ramp reported that Anthropic’s share of AI subscriptions paid by businesses increased 2.5 percentage points in May to 41%. This compares to OpenAI, which accounted for 39.5% of customers’ AI subscriptions, essentially flat from the previous month. (New data from Sensor Tower shows OpenAI still significantly ahead of Anthropic in overall consumer usage.)
Aside from subscriptions, most of what businesses spend money on is API calls to models that use tokens for activities like coding. Anthropic’s Claude Code is highly regarded as a powerful AI coding tool.
Ramp doesn’t always have access to spend data for the model most companies use. With model details available in about a third of transactions, companies are mostly spending on various versions of the Claude Opus, especially the latest version. Opus is the predecessor to Mythos and is still publicly available.
In fact, at the end of May, Anthropic released a new version, Opus 4.8.
Mythos hasn’t been on the market for long, having only been released to a limited number of users as of April. And Fable 5 ended a few days later.
While it’s impossible to predict how the latest drama involving the White House will affect Anthropic’s ability to go public as hoped (public market investors tend to be wary of companies involved in controversies with the government), the numbers suggest Anthropic’s available model is more popular with companies than ever before.
If you purchase through links in our articles, we may receive a small commission. This does not affect our editorial independence.
