On June 12, 2026, the U.S. Department of Commerce required Anthropic to restrict access to its newest AI models for foreign nationals. The lack of effective citizenship verification tools led to a фактический global shutdown of these models, exposing the risks of using export controls to manage API access.

image

What Happened

The U.S. Department of Commerce imposed mandatory access restrictions on two of Anthropic's new models for users outside the United States. Because the company lacks mechanisms for large-scale nationality verification of every client, this decision resulted in blocking API access for a wide range of foreign users, effectively implementing a "kill switch" scenario.

Context

The incident is related to the application of Export Administration Regulations (EAR) to access via APIs. Unlike physical goods, managing access to "live" AI models at the level of government regulators occurs without clear technical standards for identity verification, turning software access into a tool for rigid geopolitical control.

Why It Matters for the Industry

For the AI industry, this precedent creates critical legal uncertainty and the risk of arbitrary shutdowns of advanced technologies. This could lead to the fragmentation of the global market into "trusted" and "untrusted" zones, as well as force companies to change their access management architecture and implement new, more complex authentication mechanisms.

Why It Matters for Users

Developers and international startups face the risk of sudden service interruptions for those critically dependent on Anthropic's API. Geopolitical decisions can instantly make the use of advanced models an unstable asset, requiring the implementation of fallback mechanisms to open-source solutions or local deployments to ensure business continuity.

What Is Not Yet Known / Limitations

At this time, it is unclear whether international standards for citizenship verification for cloud AI will be developed or how long export restrictions will apply to API access.

Sources

Author

Look at AI, Editorial Team