1 hour ago
2
The US government just did something it has never done before in the commercial AI sector: it told one of the world’s most valuable companies to cut off foreign users from its best products, effective immediately.
On June 12, Anthropic disabled access to its Fable 5 and Mythos 5 models for all foreign nationals, following a directive from the US Department of Commerce. The trigger was a high-risk jailbreak vulnerability that CEO Dario Amodei flagged in a blog post, which quickly escalated into a national security conversation involving tech leaders including Amazon CEO Andy Jassy.
What happened, and why it matters beyond AI
The Commerce Department’s directive represents one of the most significant government interventions in the commercial AI sector to date. Think of it as the AI equivalent of export controls on advanced semiconductors, except this time the product being restricted isn’t hardware. It’s intelligence itself.
Anthropic, valued at somewhere between $965 billion and approximately $1 trillion, isn’t some scrappy startup that can be quietly sidelined. It’s a cornerstone of the global AI infrastructure.
European officials have already started framing AI dependency as a sovereignty issue. The logic is straightforward: if your hospitals, banks, defense systems, and government services run on AI models controlled by a foreign government that can flip the switch overnight, you don’t really have technological independence. You have a subscription.
The sovereign AI scramble
The concept of “sovereign AI” isn’t new. France, Germany, and the EU broadly have been investing in local AI capabilities for years. But those efforts have consistently lagged behind the scale and capability of American frontier models.
The vulnerability that prompted the shutdown adds another layer of complexity. Amodei’s disclosure of the jailbreak risk suggests that even the most safety-focused AI labs can produce models with serious security flaws. On June 10, two days before the government order, Amodei had underscored the urgent necessity for third-party testing of frontier AI systems due to their exponential growth.
What this means for investors
First, Anthropic itself. A valuation approaching $1 trillion is extraordinary for a company that just had its flagship products restricted by its own government.
Second, the broader AI competitive landscape is about to get more fragmented. European sovereign AI initiatives will attract fresh capital. Local AI startups in the EU, UK, and Asia that previously struggled to compete with American frontier labs now have a compelling pitch: “We can’t be turned off by Washington.”
Third, additional regulatory scrutiny for existing AI providers is essentially guaranteed. If the US government is willing to intervene this aggressively with Anthropic, a company known for its cooperative stance on safety, it signals that no AI company is immune from sudden policy shifts.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.
English (US) · 