When Washington closes a door, Asia opens its own. After the Trump administration barred non-Americans from access to the most powerful Anthropic models - Mythos and its stricter version Fable 5 - Asian startups did not wait long to fill the gap. For them, the American ban is not an obstacle, but an invitation.

The ban took effect about two weeks ago and directly hits a company that in May reached annual revenue of 47 billion dollars. The American government's idea is clear - the most advanced AI should remain a domestic asset. The effect, however, is the opposite of the one intended: instead of slowing the world down, the ban pushes it toward alternatives that do not depend on permission from Washington.

In Tokyo, the startup Sakana AI is launching the model Fugu, named after the Japanese poisonous fish. Behind it stand former Google engineers and a manager from Stability AI, and the message in the advert is direct: „top capabilities without the risk of export controls." In China, meanwhile, the cybersecurity firm 360 has presented two models for vulnerability detection and cyber defence, and its founder openly calls American policy „one-way transparency."

The difference in tone is telling. The Japanese say this is temporary and that „American models remain important for Asia"; the Chinese present their tools as a national strategic asset. Both reactions lead to the same conclusion - tech walls rarely stop knowledge, they only move it elsewhere. And for Europe, which sits between these blocs with no serious player of its own, this is one more reminder of how easy it is to end up a spectator in a game others have shaped for you.