Jin Yuzhi, the CEO of Huawei's Intelligent Automotive Solutions Business Unit, has unveiled Huawei's departure from the prevalent Vision-Language-Action (VLA) approach in intelligent driving technology. He emphasized that this route, which relies on language models for information conversion, presents obstacles in achieving genuine autonomous driving. Instead, Huawei embraces the "World Model" (WA) approach, which directly controls vehicles through environmental perception (vision, sound, etc.), bypassing the need for language processing. Despite the challenges, Jin Yuzhi is confident that this alternative path will lead to more advanced autonomous driving capabilities.

Huawei has outlined a precise roadmap for its development: by 2026, it aims to achieve L3 pilot capabilities for highway assisted driving and L4 pilot capabilities in urban areas. By 2027, it plans to initiate pilot unmanned trunk logistics and promote the commercial use of L4 autonomous driving in urban settings. Looking ahead to 2028, Huawei envisions large-scale commercial deployment of unmanned trunk logistics.