Anthropic faces a regulatory clash just as it prepares for a major IPO, with US export controls on one of its models threatening the company's market timing and reputation.

What happened: The US government imposed export controls on Anthropic's Fable 5 model on Friday. Anthropic staff traveled to Washington this week to address officials' concerns that the model's safety guardrails could be circumvented, following the company's earlier dismissal of White House worries about the issue.

Why it matters: Anthropic is preparing for a near-$1 trillion（約160兆円） IPO at a moment when one of its key products is blocked from market. The timing highlights a deeper tension: the company has built its recruitment appeal and safety reputation by pushing back against the White House—including on Pentagon requests to use AI for autonomous weapons—but IPOs require investor confidence in government relations. Anthropic now must balance its "trust us, we're the good guys" messaging with the ability to work constructively with US officials.

What to watch: The company will need to resolve this regulatory conflict to move forward with its public offering. The dispute hinges on Anthropic's claim that Fable is "not uniquely good" at finding security holes—an argument that, if accepted, might lift the export controls but risks undermining the safety-first positioning that has defined the company.