A legal technology company that sued the US government over a directive cutting off access to Anthropic’s most advanced AI models has voluntarily dropped the case. The reason: the government blinked first.

Legion LegalTech Corp filed its lawsuit on June 23 after a Bureau of Industry and Security directive, issued June 12, ordered Anthropic to deny access to its Fable 5 and Mythos 5 models for all foreign nationals. Anthropic responded by suspending global access to those models the same day the directive dropped. By July 1, after the Trump administration quietly lifted the export controls and told Anthropic that licenses were no longer necessary, the whole thing was over.

The entire saga, from directive to reversal, played out in roughly 18 days.

What happened and why it matters beyond legal tech

The BIS directive was remarkably broad. It targeted “any foreign national, whether inside or outside the United States,” which meant even foreign employees of US companies working on American soil could theoretically lose access. Legion LegalTech, which had been running its AI-powered legal platform on a commercial license for Fable 5, claimed the cutoff caused “immediate, irreparable, and existential” harm to its operations.

Legion is reportedly the first known company to file a lawsuit challenging these types of BIS directives regarding AI access. The lawsuit sought to vacate the BIS directive entirely. But around June 30, the Trump administration reversed course, notifying Anthropic that the licensing requirements had been dropped. Access was restored on July 1, and Legion subsequently withdrew its legal challenge.

The AI export control tug-of-war

Anthropic, founded by former OpenAI researchers and backed by billions in investment from Amazon and Google, has previously navigated conflicts between defense agencies and its own policies around model access. The BIS directive reflected a broader pattern of US governmental actions concerning AI export policies, particularly around frontier models that could have dual-use applications in military or intelligence contexts.

What this means for investors

The legal tech sector’s dependence on third-party AI models is now a documented vulnerability. Legion’s description of the harm as “existential” wasn’t hyperbole for a courtroom. Companies that build their entire value proposition on top of a model they don’t control are one government directive away from crisis.

Regarding Anthropic specifically: the company complied immediately with the BIS directive, suspending global access on the same day. Any enterprise customer evaluating Anthropic’s models now has a concrete example of access being cut off with zero warning.