CNAS Insights | A Closing Door on Open Weights? Implications for AI Sovereignty | CNAS

Implications for AI Sovereignty." The piece argues that foreign countries and companies can likely no longer count on the regular release of highly capable open-weight models. Open-weight models are those whose parameters are published and can be downloaded and run by outside users, and the insight frames the question in terms of AI sovereignty — the extent to which foreign governments and companies can rely on such releases for their own AI development.

The insight cites analysis from Epoch AI finding that China's leading open-weight model, GLM-5.3, lags the U.S. frontier in cyber capabilities by just six to seven months. frontier in a specific capability area. The insight's title poses the question of whether open-weight releases are becoming less frequent, and states that regular access to highly capable open-weight models can likely no longer be assumed.

The implication the piece draws concerns AI sovereignty: if highly capable open-weight models arrive less regularly, foreign countries and companies would have less to build on, and the assumption that such releases will keep coming would no longer hold. The six-to-seven-month gap between GLM-5.3 and the U.S. frontier in cyber capabilities is presented as the measure of how little separates the leading open release from the frontier. The insight is published under the CNAS Insights label.