For the maker and the artist, the latest friction between Anthropic and the Trump administration is a double-edged sword: while it restricts access to the most powerful new models, it simultaneously cements the brand’s reputation for safety and reliability in a market increasingly wary of unregulated AI.
Market dominance amidst political pressure
Anthropic finished May by overtaking OpenAI in the share of business spending for the first time, a shift highlighted by data from Ramp. The company had just raised $65 billion at a $965 billion valuation in May, matching OpenAI’s figure, before filing confidential paperwork for an initial public offering (IPO) in June. This move reportedly relied on the strength of its first-ever profitable quarter.
That momentum was abruptly halted on Friday when the Trump administration issued a directive demanding that non-Americans, including Anthropic staff, be banned from accessing its cutting-edge models. The order specifically targeted Mythos 5, a limited-release model, and Fable 5, a version made public three days prior.
Consequently, Anthropic was forced to remove its latest flagship model from the market entirely.
The security controversy
Although the White House cited an obscure export control directive, the precise reasoning remains murky. Industry chatter suggests that hackers managed to bypass Fable 5’s safety guardrails, which were designed to prevent access to Mythos’ capabilities. The model is so adept at identifying security flaws in software code that Anthropic itself marketed it as hazardous and restricted its public availability.
This escalation follows a well-known refusal by Anthropic to permit the government to utilise its models for mass surveillance of Americans or for fully autonomous weaponry. In March, these stances led the Trump administration to classify the company as a supply-chain risk.
Adoption rises despite the fallout
Surprisingly, this political drama has not dampened corporate demand. Ramp’s data indicates the opposite: the feud, which arguably validates the hype surrounding Mythos’ immense power, may actually benefit Anthropic. Ara Kharazian, Ramp’s lead economist and the compiler of the business-spending data, noted that Anthropic’s best month for business adoption coincided with the Department of Defense labelling them a supply-chain risk.
“If anything, it’ll probably boost them,” Kharazian told TechCrunch. “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.”
While Ramp’s data lacks the granularity to quantify the financial impact of removing Mythos and Fable 5 from the market, the broader trends are clear. Across more than 70,000 businesses using the platform, customers continue to heavily utilise Anthropic’s Opus models, with business usage showing steady growth.
In May alone, Anthropic’s share of AI subscriptions paid for by businesses increased by 2.5 percentage points to reach 41%. This stands in contrast to OpenAI, which held 39.5% of AI subscriptions among its customers, a figure that remained essentially flat from the previous month. It is worth noting that OpenAI still leads significantly in overall consumer usage, according to recent Sensor Tower data.
Beyond subscription fees, the bulk of corporate expenditure goes toward API calls for activities such as coding. Anthropic’s Claude Code maintains a strong reputation as a potent AI coding tool.
Reliance on available models
Ramp cannot always identify specific models in spending data, but when details are visible in approximately one-third of transactions, businesses are predominantly spending on various iterations of Claude Opus, particularly the newer versions. Opus is the model that preceded Mythos and remains openly available.
In late May, Anthropic released Opus 4.8. By comparison, Mythos had only been available to limited users since April, and Fable 5 was shut down after just a few days.
While the uncertainty surrounding the White House relationship may complicate Anthropic’s hoped-for IPO, given that public-market investors often shy away from companies entangled in government controversies, the numbers suggest that Anthropic’s currently available models are more popular with enterprises than ever before.
Key takeaways
- Anthropic overtook OpenAI in business spending share for the first time in May, driven by a record valuation and a profitable quarter.
- A US government ban on non-citizens accessing Mythos 5 and Fable 5 forced Anthropic to pull these models, yet adoption of existing Opus models continues to grow.
- Data from Ramp shows Anthropic captured 41% of business-paid AI subscriptions in May, outpacing OpenAI’s 39.5% in that specific metric.
- Paradoxically, the controversy surrounding the company’s safety and autonomy stances may have enhanced its “aura,” boosting business trust and usage.
