US Government Forces Anthropic to Disable Models for Foreign Nationals
A federal directive blocking foreign-national access to Anthropic's newest models reveals how export-control logic is moving directly into AI deployment.
3. US Government Forces Anthropic to Disable Models for Foreign Nationals
The Trump administration directed Anthropic to disable access to its most recently released models for foreign nationals, AI Business confirmed on May 20, 2026. The order followed shortly after Anthropic shipped two new high-capability models, and represents one of the most direct federal interventions into a commercial AI product's availability to date. The scope of the shutdown applies specifically to foreign nationals, not to geographic regions, making identity and citizenship status the operative access filter rather than IP geolocation alone.
This is export-control logic applied at the model layer, and it changes the competitive math in ways that extend well beyond Anthropic. OpenAI and Google DeepMind both operate frontier model APIs with significant international usage. If the federal government treats high-capability model access as a controlled export, every lab with a frontier product now faces potential compliance obligations that no standard cloud provider has had to engineer before. Anthropic, which has positioned itself as the safety-first, government-friendly alternative to OpenAI, now finds that proximity to Washington cuts both ways: preferred partner status comes with preferred compliance exposure. The directive also hands foreign competitors, particularly Chinese labs like Baidu and Zhipu AI, a ready-made narrative about US model access being structurally unreliable.
The broader pattern here is the convergence of AI capability thresholds with national security frameworks that were built for semiconductors and weapons systems, not software APIs. Watch for whether the directive triggers formal rulemaking, which would force the Commerce Department to define what capability level triggers export restrictions. That definition, whenever it comes, will set the boundary conditions for every frontier lab's international go-to-market strategy.
Source: Anthropic Forced to Disable New Models by US Government