Anthropic has suspended global access to its most advanced AI systems, including Claude Mythos 5, Mythos Preview, and Fable 5. The move follows a June 1, 2026 directive from US Commerce Secretary Howard Lutnick to CEO Dario Amodei. The order mandates strict governmental approval for any transfer of these models to non-US persons.
The restrictions apply broadly, affecting foreign nationals even when physically located within the United States. Anthropic confirmed full compliance on June 13, 2026.
The Commerce Department cites national security concerns, specifically the models' advanced cybersecurity capabilities. Officials fear these tools could be weaponized by adversaries if not tightly controlled. Previously, access was limited to US-based tech giants like Microsoft, Google, and NVIDIA through Project Glasswing, focusing on critical infrastructure defense.
This decision marks a significant shift in regulatory posture. The US government now treats frontier AI models akin to advanced semiconductors or military technology. For investors, this signals a shrinking addressable market for top-tier AI, creating friction for international enterprises and multinational corporations. The precedent suggests future export licensing will become standard for high-capability artificial intelligence.