An artificial intelligence startup utilizing advanced foundation models developed by Anthropic PBC has filed a lawsuit against the United States government, challenging an unprecedented export control directive that forces tech companies to bar foreign nationals from accessing top tier AI software.
The legal challenge, first reported by Bloomberg, marks an escalating confrontation between commercial AI developers and federal regulators over the boundaries of national security, software code, and digital border control.
The legal battle stems from an emergency export control directive issued by the U.S. Commerce Department. The order commanded Anthropic to immediately suspend access to its newly released flagship models, Claude Fable 5 and Mythos 5, for all foreign nationals regardless of whether they reside inside or outside the United States.
Because Anthropic and its enterprise clients lack the infrastructure to instantly verify the citizenship of hundreds of millions of global users, the company was forced to take a nuclear option: pulling the models completely offline worldwide.
“The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance,” Anthropic stated in an official announcement detailing the disruption. “Access to all other Anthropic models will not be affected.”
While Anthropic itself has been locked in ongoing friction with the administration over federal supply chain blacklists, this new lawsuit introduces a vital third party to the fray: the commercial software startups building on top of these models, whose entire business models have been upended overnight by the government’s intervention.
According to government officials, the extraordinary intervention was driven by severe national security apprehensions. Intelligence agencies raised alarms that the “Mythos” class models possess an advanced, autonomous ability to identify hidden software vulnerabilities that could be weaponized in cyberwarfare or state sponsored digital attacks.
However, Anthropic and its ecosystem partners strongly reject the government’s justification, arguing that federal authorities have overreacted to standard industry limitations. Anthropic clarified that the U.S. government had only provided “verbal evidence of a potential narrow, non universal jailbreak,” which simply involved prompting the system to read a specific codebase and fix minor bugs.
“We disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people,” Anthropic stated, warning of industry wide consequences if the directive stands. “If this standard were applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers.”
Historically, U.S. export controls aimed at containing foreign adversaries like China have targeted physical infrastructure specifically high end semiconductor chips and lithography tools from companies like Nvidia and ASML.
By targeting the weight files and algorithmic accessibility of software, the Commerce Department is entering uncharted regulatory waters. Tech policy analysts warn that enforcing citizenship checks on software access transforms a technical safety issue into a massive identity management crisis for the entire global tech economy.
Dean Ball, a former White House official who contributed to the 2025 AI Action Plan, noted the severe practical implications of the order on social media:
“This means you should expect to have to prove your citizenship to use Anthropic models.”
The anonymous startup’s lawsuit seeks a preliminary injunction to halt the enforcement of the government’s order, claiming the broad restriction violates regulatory due process and inflicts immediate, catastrophic financial harm on businesses utilizing the API. The case is now moving through federal court, where a ruling could redefine how artificial intelligence is distributed globally.
