China plans to impose export curbs on its most advanced AI models, marking a shift toward protecting AI as a strategic national security asset, according to a Reuters report. The Ministry of Commerce held private briefings with tech giants including Alibaba, ByteDance, and Z.ai to evaluate restrictions on overseas access. The proposed three-tier regime targets both proprietary systems and open-weight models like Alibaba’s Qwen series, ByteDance’s Doubao, and Z.ai’s GLM-5.2. Tier 1 covers basic models with simple government filing; Tier 2 requires mandatory state security review for high-performance models; Tier 3 bans frontier systems from foreign access. Officials are also evaluating updates to national security laws that could elevate unauthorized transfer of AI model weights to a national security crime. Europe is caught in the crosshairs, as it largely lacks a competitive domestic frontier layer, with the notable exception of France’s Mistral AI.