Anthropic Spent Its Weekend Fighting the Trump Admin Over Mythos 5

While the rest of the country was busy celebrating the USA’s first World Cup win and the Knicks finally taking the championship, Anthropic was having a very different kind of weekend. At 5:21 PM on Friday, the company got hit with a US export control directive that basically told it to suspend access to its newly released Mythos 5 and Fable 5 models. Not just for foreign customers—the order applied to “any foreign national” inside or outside the US, including foreign national Anthropic employees.

That’s a problem. Because Anthropic, like most AI companies, has a global workforce. The only way to comply with that order was to completely kill access to products the company had spent the past week hyping. So they did what any reasonable company would do: they disabled the models and sent people to Washington, DC to try to change President Donald Trump’s mind.

This is higher stakes than your typical regulatory spat. The directive essentially says that if you’re not a US citizen, you can’t use these models—even if you’re the person who helped build them. That’s a massive operational headache, and it raises some uncomfortable questions about how the government views AI talent and intellectual property.

I’ve seen this approach before, but not at this scale. Export controls on AI models have been floated for years, usually targeting adversarial nations. But applying them to a company’s own foreign-born employees? That’s new territory. It suggests the administration is taking a much harder line on AI sovereignty than I expected.

Anthropic’s response was swift and dramatic. They didn’t just push back with legal filings—they physically went to DC to lobby the President. That tells you how serious they think this is. The company has been positioning Mythos 5 as a leap forward in reasoning and safety alignment, and suddenly they can’t even show it to half their own team.

The irony isn’t lost on me. The same week the US celebrates global sports victories, it’s locking down AI models in a way that feels more like Cold War tech controls than modern innovation policy. And the companies caught in the middle have to decide whether to fight or fold.

I’ll be watching how this plays out. If the directive stands, it sets a precedent that could reshape how AI companies hire, where they build, and who gets to use the most advanced models. Not exactly the kind of victory lap anyone expected this weekend.