AI Sovereignty: A Growing Concern for Businesses Amidst Vendor Dependence

On June 12, 2026, at 5:21 p.m. Eastern time, a letter reached Anthropic, prompting its two most advanced models to shut down for every customer worldwide. The U.S. Commerce Department had invoked export-control authority after a reported jailbreak, citing that the directive barred access by 'any foreign national, whether inside or outside the United States.'

Anthropic couldn't verify nationality in real-time across various platforms, so it switched everything off, leaving its model Claude Fable 5 down for 18 days until Commerce lifted the controls on June 30.

This incident highlights the importance of AI sovereignty, defined by IBM as an organization's capacity to control its AI stack: infrastructure, data, models, and operations. The goal is not ownership but rather selective sovereignty, a set of dials that allow organizations to manage their dependencies.