China’s AI policy under Xi Jinping embraces both openness in sharing AI advancements and strict controls to manage associated risks, leaving the future of open-weight models uncertain.
- China supports open AI models while insisting on risk management.
- Open-weight releases bolster China’s technology presence worldwide.
- Government policy stresses developing security alongside innovation.
What happened
Since January 2025, Chinese companies such as DeepSeek, Z.ai, and Moonshot have launched open-weight AI models that allow widespread, customizable use. These open models offer competition to closed systems from firms like OpenAI and Anthropic by being lower cost and self-hostable, broadening China’s AI technology reach globally.
At the July 2026 World AI Conference, Chinese President Xi Jinping emphasized both the promotion of openness and cooperation as well as the importance of security and control in AI development. His comments reflect long-standing policies prioritizing both innovation and risk awareness, leaving open questions about how China will regulate open-weight AI moving forward.
Why it matters
China’s willingness to share advanced AI models openly challenges conventional international norms where major AI firms restrict access to protect safety or commercial advantage. By distributing powerful models openly, Chinese AI gains global integration and fosters goodwill toward China’s technology sector, potentially increasing commercial influence in cloud services and AI infrastructure worldwide.
However, unrestricted open-weight AI poses risks including potential misuse for cyber attacks or other harmful applications. China's dual focus on development and security underscores an unresolved tension, as authorities must weigh the benefits of openness against the imperative to maintain control and avoid loss of oversight over AI capabilities.
What to watch next
Future Chinese government actions regarding the release of large open-weight models will be a key indicator of how Beijing balances innovation with national security concerns. Legal measures or enforcement steps could tighten distribution if risks are deemed too high, especially as AI models grow more powerful and complex.
Meanwhile, monitoring developments from prominent Chinese AI firms and government statements will shed light on whether China continues fostering open collaboration or shifts toward a more closed, controlled AI ecosystem. The global AI safety community and international policymakers will also closely watch how China manages this critical balance.