Why Governments Cannot Simply Flip a Switch to Shut Down Deployed AI Models

When the U.S. Department of Commerce forced Anthropic to disable its newly released Fable 5 and Mythos 5 models in June 2026, the company had no choice but to shut them down globally for every customer — even though the technical infrastructure supporting them cannot always be controlled by a single vendor. In a landmark moment for AI governance, the 18-day outage revealed a critical infrastructure challenge facing regulators worldwide: once frontier AI models are deployed across distributed cloud systems, pulling the plug is far more complex than flipping a switch.