The 13th Annual Conference on Intelligent and Connected Vehicles (CICV 2026) was held in Shanghai from May 21 to 22, 2026. At the event, Sheng Kehua, CTO of DiDi Autonomous Driving, stated that the industrialization and deployment of Robotaxi services require deep integration across AI technology, vehicles, productization, and operational systems—not just breakthroughs in isolated technologies. In terms of core technologies, DiDi Autonomous Driving has independently developed a multi-link fusion system. Building upon the mainstream perception-prediction-planning architecture, it incorporates reinforcement learning to enhance handling of long-tail scenarios and continuously optimizes user experience and efficiency within safety boundaries through on-device large models. In Q4 2025, DiDi Autonomous Driving launched round-the-clock passenger-carrying tests in designated demonstration zones in Guangzhou and Beijing. Its end-to-end operations hub, “Huiju Port,” provides Robotaxi services to users in these areas, supporting the development of smart cities. Sheng Kehua noted that Robotaxi vehicles will evolve through stages—from retrofitted models and pre-installed adaptations to purpose-built new vehicle architectures. DiDi Autonomous Driving is collaborating with GAC Aion to advance vehicle upgrades, and their jointly developed next-generation Robotaxi model, R2, was delivered in January 2026 and has since begun road testing, featuring significant improvements in autonomous driving performance, cabin experience, and reliability. Sheng emphasized that autonomous driving is a long-term, capital-intensive systems engineering endeavor. DiDi Autonomous Driving will continue strengthening its systemic capabilities and steadily advance industrialization and global expansion.