Trump admin weighs limits on Chinese open-source AI

The Trump administration is actively considering measures to restrict U.S. companies' access to cutting-edge Chinese open-source artificial intelligence models, reigniting a policy debate over national security, market competition, and technological innovation. These deliberations follow the recent release of Kimi K3, a powerful open-source model from Chinese lab Moonshot AI, which was unveiled on July 17, 2026, aligning with the World Artificial Intelligence Conference in Shanghai. U.S. officials are concerned about the rapid adoption of these models by American firms, drawn by significantly lower costs and competitive performance.

The administration's approach, as described by sources familiar with the discussions, appears to favor indirect pressure over an outright ban. The likely strategy involves invoking federal procurement rules, renewing threats to add Chinese AI labs to the Commerce Department's Entity List, and launching public pressure campaigns aimed at U.S. companies using the models. Officials also plan to highlight potential cybersecurity vulnerabilities, including possible backdoors, within Chinese AI systems.

This renewed push follows several earlier, stalled...