US Government Restricts Access to Anthropic AI Models Amid Ongoing Dispute

The Donald Trump administration deepened its high-profile dispute with Anthropic , the artificial intelligence startup that has become now become a multi-billion-dollar valued giant, this week by restricting foreign access of two of its most advances AI models. These past few months, we’ve seen US Defence Secretary Pete Hegseth repeatedly issue threatening deadlines targeting co-founder and CEO Dario Amodei’s company over not budging from its principles, barring the unrestricted use of the AI models “for all legal purposes.” On Saturday (US time), stoked the fire even further by penning a begrudging message that singled out Anthropic on social media. “Three months ago, @DeptofWar kicked @AnthropicAI out of our building–forever,” he wrote on X. “Every passing day proves why that was the right move.” Hegseth’s virtual missive surfaced shortly after Anthropic publicly revealed that the US government issued an “export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national,” including even the company’s own foreign international employees, over “national security concerns.” Complying with the order, Anthropic said that it was abruptly disabling its models Fable 5 and Mythos 5 for all customers. Both sides have tussled over the use of AI in warfare since early 2026, with the current administration ultimately blacklisting the Dario Amodei-led company and forging agreements with so many of Anthropic’s rivals: SpaceX, OpenAI, Google, Microsoft, Nvidia, Amazon Web Services, Oracle and Reflection. Three months ago, @DeptofWar kicked @AnthropicAI out of our building—forever. Every passing day proves why that was the right move. 🇺🇸 However, even, before that, US Secretary of Defence Pete Hegseth laid out an ultimatum for the AI giant in late February, urging the co-founder and CEO to relent within days or face one of two punishments – compulsion under the Defence Production Act or be expelled from the defence supply chain as a “national security risk.” Anthropic, on the other hand, chose not to budge from its principles. In a blog post dated February 26, CEO Dario Amodei wrote, “In a narrow set of cases, we believe AI can undermine, rather than defend, democratic values. Some uses are also simply outside the bounds of what today’s technology can safely and reliably do. Two such use cases have never been included in our contracts with the Department of War.” The pointers, which ultimately resulted in the company being shunned by the Trump administration, Amodei chose to speak out against were: use of their systems for mass domestic surveillance being incompatible with democratic values, and AI-powered fully autonomous weapons that could put USA’s warfighters and civilians at risk. US President Donald Trump has since also characterised the Amodei’s AI giant as a “radical left, woke company” on social media. Meanwhile, Hegseth announced on X that Anthropic was a “Supply-Chain Risk to National Security,” compelling military contractors and suppliers not to engage in commercial business with it. Since then, the General Services Administration has renounced ⁠Anthropic’s government-wide contract, with Treasury, State and the Federal Housing Finance Agency also publicly cutting ties. Days later, the Amazon-backed company escalated the dispute by suing the US government over retaliation for refusing to remove safety limits of its Claude model. Although the company had long been willing to work with the military, it did not want to do it on any terms. The company ultimately raised five legal claims against the government, arguing that the actions violated the Administrative ​Procedure Act, the First Amendment, the Fifth Amendment, the president’s statutory authority, and the APA’s prohibition on unauthorised agency sanctions. Moreover, Anthropic accused the Pentagon of launching a significant air attack on Iran using its tools hours after Hegseth declared the company a “Supply-Chain Risk to National Security.” The supply chain blacklisting of Anthropic is set to take effect later in the year, while the San Francisco AI company behind Claude is bound for its initial public offering as early as October 2026. Although an official date has yet to be announced, the company has already set the groundwork for a potentially blockbuster Nasdaq debut by filing confidential IPO paperwork with the US Securities and Exchange Commission (SEC) earlier this month. The major move came after a massive funding round in May boosted Anthropic’s valuation to approximately $965 billion. Overtaking Sam Altman’s OpenAI, Amodei’s company became the most valuable AI startup in the world.