Anthropic's most powerful AI models shut down by US government order

Anthropic has announced it will completely disable Fable 5 and Mythos 5, its most advanced AI models, for all global users. Why? A direct export control order from the US government citing "national security concerns." Basically, Washington decided these models are too dangerous to be accessible to anyone who isn't a US citizen.

Here's the kicker: the government didn't even explain itself properly. Anthropic got the order without specifics, just vague accusations that the company's safety systems "can be bypassed" and users might exploit these models to identify software vulnerabilities. Translation: the government thinks they're too good at doing things that could be weaponized.

This is where it gets interesting — Anthropic has zero choice here. This isn't a polite request or a suggestion the company can debate. It's an executive order with the force of law. The company said it would disable access "abruptly" — not gradually, not with warning, just: it's gone. For anyone outside the US, Fable 5 and Mythos 5 become unreachable.

The irony is brutal: Anthropic is an American company that poured billions into R&D, yet now it can't even sell its own products to international customers who might have already built these models into their systems. European startups, Japanese corporations, Canadian developers — all locked out. It's like forcing Coca-Cola to destroy every bottle shipped overseas because the government says the formula is too dangerous.

Worth noting: this isn't happening in a vacuum. The US is getting increasingly aggressive about controlling AI tech exports, trying to maintain competitive advantage over geopolitical rivals (read: China). But the cost is steep — not just for Anthropic, but for the global tech industry that now knows its products can be nuked by decree overnight.

For Anthropic, it means losing access to international markets, enterprise clients outside the US, global partnerships. For the rest of the world, it means American tech leadership keeps consolidating, not through technical merit, but simply because they can.