Nearly 40 Key Autonomous Driving Talents Cross Over to Embodied AI, Driven by Tech Reuse and Capital

Since 2023, approximately 40 core executives and technical experts from China’s intelligent driving sector have shifted to the embodied intelligence field, joining more than 20 startups—over 70% of which were founded within the past two years. This talent migration is primarily driven by three factors: First, embodied intelligence and autonomous driving share a highly aligned technological closed loop in “perception-decision-action,” allowing methodologies such as multimodal fusion, end-to-end large models, world models, and data-driven feedback loops to be readily transferable. Second, intelligent driving technology paths are converging, and market structures are solidifying, with leading companies like Huawei, Li Auto, NIO, XPeng, and Horizon Robotics dominating the landscape—thereby limiting career advancement opportunities for others. Third, massive capital inflows into embodied intelligence are significantly boosting compensation packages and enhancing the feasibility of entrepreneurship. Some embodied intelligence companies are offering exceptionally high salaries to attract top talent. For instance, UBTECH advertised a chief scientist role in embodied intelligence with a base annual salary of RMB 15 million, potentially reaching up to RMB 124 million. Similarly, Yu Hao, founder of Dreame Technology, previously sought a chief scientist with an annual offer of RMB 200 million. In terms of talent outflow, Li Auto and Horizon Robotics have emerged as key sources. Li Auto has produced entrepreneurs including Shen Yanan, Lang Xianpeng, Wang Kai, Xia Zhongpu, and Zhao Zhelun. Horizon Robotics has contributed talents such as Zhang Yufeng, Yu Yinan, Sun Junkai, Pan Yangjiayi, and Fan Qingyuan. Li Auto emphasizes its AI identity and robotics orientation, while Horizon Robotics has long positioned itself as “the Intel of the robotics era,” focusing on cultivating hybrid talents with both technical depth and business acumen. Although professionals from intelligent driving bring valuable engineering implementation experience and systems-level thinking, embodied intelligence poses greater challenges in physical interaction, joint control, coordination between “large and small brains” (high-level cognition and low-level motor control), and adaptation to unstructured environments. Humanoid robot development is far more complex than automotive engineering, and clear, essential use cases have yet to emerge, leaving commercialization pathways uncertain. Moreover, the automotive industry prioritizes functional safety and long development cycles, whereas the robotics field favors rapid iteration—a cultural mismatch that adds managerial complexity. Industry insiders believe the current embodied intelligence sector faces issues such as overheated investment, talent overvaluation, and ambiguous application scenarios, and will likely undergo significant consolidation. To succeed in this cross-domain transition, entrants must shift their mindset from viewing robots as mere “tools” to recognizing them as “general-purpose partners,” and rebuild their technology, products, and business models accordingly.