The Trump administration issued an order on Friday, June 13, 2026, that requires Anthropic to prevent anyone outside the United States from using its two main AI products: Mythos (used by businesses) and Fable 5 (used by everyday consumers). The order invokes export control laws — the same rules the government uses to restrict what technology can be sold overseas.

The directive cited concerns that Chinese nationals or organizations could gain access to Mythos. Both products are now blocked to non-US users, which suggests the administration views even the consumer-facing product as a potential security risk, not just the cutting-edge research version.

A Trump adviser named David Sacks said the restrictions are not connected to previous policy concerns, though he did not explain what specific threat caused the June 13 action.

The broader context matters here. Sacks's public statement does not align easily with what the White House has said about why this is necessary. The government has not yet laid out a clear explanation for what actual risk it is trying to prevent. This gap is important because what Anthropic has to do next — and whether this order holds up if challenged in court — depends on which specific law the government is using to impose it.

How Export Controls Work

The US Commerce Department has the authority to restrict access to certain technologies that could help other countries, particularly those viewed as adversaries. These rules have become stricter over time, especially for AI systems. The new rules applied to Fable 5 — a consumer product sold to millions of regular people — represent a significant expansion of how the government has historically used these controls.

For this order to work, Anthropic will need to verify that each user is actually a US citizen, not just that they live in the US. That is much harder to do than simply blocking users based on which country their internet connection comes from. The company will need to do this at the moment someone signs up, across millions of potential users worldwide. It is a serious technical and business challenge.

What Happens Now

Anthropic is in an unusual spot. It has taken money from US government-backed AI programs while also trying to build products for customers around the world. This order puts that tension on full display.

Other AI companies — OpenAI, Google, Meta — will be watching closely. If the government extends this same approach to other labs, people outside the US who use American-made AI tools could suddenly lose access with little warning.

The order also creates a immediate problem for Anthropic's business customers outside the US. Mythos is likely already built into the workflows of companies in Europe, Asia, and other regions. A sudden shutdown, rather than a gradual wind-down, would disrupt those operations.

There is a pattern worth noting. Over the past few years, the US government has restricted the sale of advanced computer chips to China, even though those restrictions have hurt American companies and their international partners. The same logic is now being applied to AI software itself: Washington appears willing to absorb economic costs in allied countries to slow China's AI development. This is the next step in that strategy, and it happened faster than many industry observers anticipated.

The key questions now are whether this order will survive a legal challenge, and whether Anthropic can actually verify citizenship at the scale needed without losing too many users in other countries. The answers will determine whether this turns out to be a major turning point or a decision that eventually gets rolled back.