Anthropic has a month.
AI Labs ended May surpassing OpenAI in market share of enterprise spending for the first time, Lampe revealed. The company raised $65 billion at a $965 billion valuation (also higher than OpenAI) at the end of May, and was on track to file a confidential IPO in June on the strength of its first-ever profitable quarter.
And on Friday, the Trump administration reignited the war on the model maker by sending a letter barring non-Americans, including Anthropic employees, from accessing the company’s most advanced model, the limited-release Mythos 5 and a stricter version called the Fable 5 that was released to the public three days earlier.
This effectively forced Anthropic to withdraw its latest omnipotent model from the market entirely.
The White House invoked vague export control directives in issuing the ban, but the exact cause remains unclear. The buzz was that hackers were able to easily bypass Fable 5’s guardrails intended to prevent access to Mythos’ features. This model is so good at finding security flaws in software code that Anthropic itself promoted it as dangerous and restricted its release.
The new drama is based on the fact that Anthropic famously wouldn’t allow the government to use its models for mass surveillance of American citizens or fully autonomous weapons. As a result, in March the Trump administration declared the company exposed to supply chain risk.
That didn’t stop Anthropic from selling to businesses. Lamp’s data shows just the opposite. Ironically, this latest feud with the Trump administration, which also seems to confirm the buzz surrounding Mythos’ mythic powers, may help Anthropic rather than hurt it, according to Ara Karadzian, chief economist at Lamp. Harajian is the person who compiled AI data on corporate spending.
“If anything, it will probably give them a boost,” Karazian told TechCrunch. “Anthropic’s best month ever as far as enterprise adoption was concerned was the month that the Department of Defense classified Anthropic as a supply chain risk. There’s a lot of aura surrounding the model, especially with the name Too Dangerous for Use.”
Ramp’s data is not detailed enough to confirm how much of a financial hit the company would take by pulling Mythos and Fable 5 from the market.
Still, data from the more than 70,000 businesses that use its platform shows that customers use Anthropic’s Opus model frequently and its use in business is increasing.
For example, Ramp reported that Anthropic’s share of AI subscriptions paid by enterprises increased by 2.5 percentage points to 41% in May. By comparison, OpenAI accounted for 39.5% of customer AI subscriptions, almost flat from the previous month. (OpenAI still has a significant lead over Anthropic in overall consumer usage, according to new data from Sensor Tower.)
Beyond subscriptions, the majority of what companies spend money on is API calls to models that cover the use of tokens for activities such as coding. Anthropic’s Claude Code has a reputation as a powerful AI coding tool.
Ramp can’t always tell from spending data which model most companies are using. Once model details are visible in about a third of transactions, companies will primarily spend on different flavors of Claude Opus, especially later versions. Opus predates Mythos and is still published today.
In fact, in late May Anthropic released a new version, Opus 4.8.
Mythos was released to a limited number of users as of April and hasn’t been on the market for that long. And Fable 5 was shut down a few days later.
While it’s impossible to predict how this latest drama with the White House will affect Anthropic’s ability to go public as expected (public market investors tend to be wary of companies embroiled in disputes with the government), the numbers show that the models available to Anthropic are more popular with companies than ever.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
