In recent times, prominent global AI enterprises and chip manufacturers, including OpenAI and Anthropic, have ramped up their investments in inference chips. This shift is largely attributed to the fact that real-world application workloads are now reshaping computing architectures in a reverse manner. The uninterrupted task chain workloads initiated by Agent technology have introduced novel demands for the design of computing power architectures. China’s AI sector boasts strengths in application deployment, chip development trajectories, and open-source ecosystems. Nevertheless, the industry still grapples with high adaptation costs and fragmentation issues across different segments of the industrial chain.

The AICC2026 Artificial Intelligence Computing Conference will center its attention on application workloads, fostering collaborative progress throughout the entire industrial chain, encompassing chips, interconnects, and storage solutions. The conference will explore topics such as computing architectures in the Agent era and advancements in cutting-edge architectures. Additionally, it will unveil the ‘2026 China AI Computing Power Development Assessment Report’ and present rankings of AI computing power cities, offering valuable insights into the evolving landscape of AI computing in China.