Qingzhou Zhihang Unveils General Physical AI Roadmap to Accelerate Autonomous Driving and Logistics Deployment

On April 12, 2026, at a high-level forum on the development of intelligent electric vehicles, Yu Qian, Co-founder, Chairman, and CEO of QCraft, delivered a keynote speech asserting that autonomous driving represents the optimal entry point toward achieving General Physical AI. He emphasized that world models and reinforcement learning are critical pathways to this goal. According to Yu, the AI capabilities of intelligent vehicles have undergone three evolutionary phases, and in 2026, the industry has entered a new era of "superhuman intelligence," shifting its core focus from building "driving systems" to creating a "brain for the physical world." Centered on General Physical AI, QCraft launched its Chengfeng 2.0 intelligent driving solution, featuring a unified architecture combining Vision-Language-Action (VLA) models with world models. The solution offers three product tiers—AIR, PRO, and MAX—to address full-scenario needs ranging from highway NOA to advanced urban NOA. It prioritizes extreme optimization on single chips and tight hardware-software co-design, reducing reliance on high-compute hardware and accelerating the adoption of advanced urban NOA in mainstream passenger vehicles. In terms of commercial deployment, QCraft’s L4 Robovan autonomous logistics solution has achieved "production-to-operation" readiness. Designed for cargo volumes between 3 and 12 cubic meters, it supports round-the-clock (7×24) operations across diverse applications including express delivery, cold-chain logistics, supermarket distribution, industrial logistics, and security patrols. The solution cuts operational costs by 50% on the same routes, with an autonomous driving system cost under RMB 10,000. Its map-free technology further lowers deployment barriers. The Robovan is already in operation in cities such as Ningbo, Jinhua, Wuhu, and Suzhou. Internationally, QCraft’s solutions are compatible with major chip platforms from NVIDIA, Qualcomm, and Horizon Robotics. The company has also obtained key international certifications—including ASPICE CL2, ISO 26262, ISO 9001, and ISO/SAE 21434—ensuring compliance with global regulatory requirements. QCraft adheres to a "dual-engine" strategy, leveraging shared technological foundations for both L2 and L4 systems to drive intelligent, green, integrated, and global development. Yu Qian stated that 2026 marks the beginning of a golden decade for driverless technology, and QCraft will continue advancing the application of General Physical AI across all scenarios, empowering China’s new energy vehicle industry to ascend to the high end of the global value chain.