At the 2026 Aggregated Intelligent Industry Development Conference, held on September 9 in Wuhan's Optics Valley and co-organized by the Aggregated Intelligent Industry Innovation Center and Hubei Science & Technology Investment, Liu Hongqian, the General Manager of Puhua Basic Software, delivered a keynote address. She argued that the evolution from model capabilities to system capabilities necessitates a new generation of operating systems to support the aggregated intelligent industry.

Liu emphasized that the sector's core rests on two pillars: model capability, which dictates an intelligent agent's cognitive sharpness, and system capability, which ensures its stability and reliability. These elements are mutually reinforcing, and the vital link that elevates and connects them is the operating system itself. She highlighted a critical principle: while AI can afford trial and error, safety cannot.

According to Liu, this next-generation operating system must simultaneously guarantee real-time determinism, functional safety, and compatibility with the AI development ecosystem. It needs to uniformly manage heterogeneous computing power, sensors, and actuators, while delivering real-time communication, task scheduling, and secure isolation. Drawing a parallel to the automotive industry, she described today's intelligent vehicles as wheel-based intelligent agents operating in open physical environments, constantly perceiving, deciding in real-time, and executing at high speeds. These vehicles face unpredictable road conditions and traffic participants, requiring millisecond-level control loops that meet the highest functional safety standards. Consequently, she concluded, intelligent vehicles represent one of the most complex and mature applications of embodied intelligence to date.