The US Commerce Department ordered Anthropic to block foreign access to its most advanced AI models on June 13, and within hours the company pulled the plug globally. Not just for foreign nationals. For everyone, including American users.

Canadian Prime Minister Mark Carney seized on the moment to make a point he’s been building toward since taking office in March 2025: relying on a handful of powerful American AI providers is a strategic vulnerability, not just an inconvenience.

What happened and why it matters

The export control directive specifically targeted two Anthropic models, Fable 5 and Mythos 5. The Trump administration’s justification centered on national security concerns tied to what Anthropic itself characterized as a “narrow, non-universal jailbreak” connected to undiscovered software vulnerabilities.

Fable 5 had been released more broadly earlier in June, while Mythos 5 already had limited availability due to cybersecurity considerations. The directive forced Anthropic to disable both models across the board, a decision the company complied with despite publicly expressing disagreement with the order’s lack of clarity.

The company has previously taken legal action against restrictions imposed by the Trump administration on its technology.

Carney’s play for AI sovereignty

Carney’s warnings feed into Canada’s broader strategy of diversifying its technological partnerships away from American firms. The country has been actively pursuing stronger ties with the European Union, positioning itself as a bridge between North American innovation and European regulatory frameworks.

The bigger picture for tech investors

The export ban represents a significant escalation in governmental control over advanced AI technologies. The directive signals that any AI company operating at the frontier could face similar restrictions with minimal warning. A company can be generating strong revenue, growing its user base, and expanding internationally, and then lose access to its entire foreign customer base overnight because of a security vulnerability it didn’t even create.

For companies that have built their products and services on top of these AI models, the risk calculus just changed dramatically. Any business that depends on a single AI provider’s API now has a geopolitical concentration risk that most haven’t accounted for.