China Promotes Comprehensive IPv6 Support for Generative AI Large Models

China has launched a one-year special initiative to promote comprehensive IPv6 support in generative AI large model applications and APIs. The initiative, joint-launched by the Cyberspace Administration of China and five leading AI firms, aims to upgrade network infrastructure to handle surging AI token traffic. As AI applications grow, the explosion of token traffic demands robust network capacity, which IPv6 can support through its massive address space, efficient routing, and built-in security. This transition will accelerate the integration of AI with next-generation internet protocols, laying a foundation for native intelligent networks. The initiative aims to increase the share of IPv6 traffic in LLM API calls, alongside a pilot in Xiong'an New Area targeting full IPv6 single-stack operations in new districts by 2030. Additionally, China's broader national plan targets 950 million active IPv6 users and a 42% network traffic share by 2030.