Unexpected Benefits: How Anthropic’s Clash with the Trump Administration Could Boost Its Sales
Anthropic is experiencing significant developments this month.
The AI laboratory concluded May by outpacing OpenAI in terms of market share for business expenditures, as reported by Ramp. By the end of the month, it raised $65 billion at a valuation of $965 billion, also surpassing OpenAI, and entered June by filing confidential documents for an IPO, backed by its first profitable quarter.
However, on Friday, the Trump administration intensified its scrutiny on the company by issuing a mandate that prohibits non-American individuals, including Anthropic’s employees, from accessing its cutting-edge models: the limited-release Mythos 5 and the recently launched Fable 5.
Consequently, Anthropic was compelled to withdraw its powerful model from the market entirely.
Although the White House referenced an obscure export control regulation in its ban, the specific reasons remain ambiguous. Speculation suggests that hackers could easily circumvent Fable 5’s safeguards, designed to restrict access to Mythos’ capabilities. This model has proven exceptionally adept at detecting security vulnerabilities in software, prompting Anthropic to categorize it as potentially hazardous and limit its public release.
This unfolding situation follows Anthropic’s well-known refusal to permit the government to leverage its models for widespread surveillance of American citizens and fully autonomous weaponry. As a result, the Trump administration labeled the company a supply-chain risk back in March.
Despite this, Anthropic’s sales to enterprises have thrived, according to Ramp’s findings. Interestingly, this latest conflict with the Trump administration might bolster rather than hinder Anthropic, suggests Ramp’s chief economist, Ara Kharazian, who analyzed the data on business spending in AI.
“In fact, it might enhance their appeal,” Kharazian noted. “Anthropic’s best month in terms of business adoption coincided with the Department of Defense designating them as a supply-chain risk. Being referred to as too dangerous for use certainly adds a mystique.”
While the available data doesn’t detail the financial consequences of removing Mythos and Fable 5, it does indicate that companies widely utilize Anthropic’s Opus models, with business adoption on the rise.
For example, Ramp revealed that Anthropic’s share of AI subscriptions from businesses increased by 2.5 percentage points in May, reaching 41%. In contrast, OpenAI’s share was at 39.5%, remaining relatively stable from the previous month. Nonetheless, OpenAI still leads Anthropic significantly in consumer usage, according to new data from Sensor Tower.
Moreover, a considerable portion of corporate spending focuses on API calls to the models, which encompass token usage for tasks like coding. Anthropic’s Claude Code is well-regarded as a potent AI tool for coding.
Ramp doesn’t always determine which models businesses are utilizing based solely on spending data. When model details are available—about one-third of transactions—most companies are reportedly investing in different iterations of Claude Opus, especially the more recent releases. Opus, the predecessor to Mythos, continues to be widely accessible.
Notably, Anthropic introduced a new version, Opus 4.8, in late May.
Mythos was newly introduced and had only limited availability since April, whereas Fable 5 was discontinued shortly after its launch.
Although it’s uncertain how these latest developments with the White House may affect Anthropic’s aspirations for a public offering—given that public investors often shy away from firms involved in governmental controversies—the statistics suggest that Anthropic’s current models are enjoying greater popularity among businesses than ever before.
