At the 2026 World Robot Conference, Yuan Foyu, Vice President of Baidu Group, highlighted that the embodied AI industry is shifting from the phase of “capability emergence” to “productivity realization,” with system-level synergy emerging as a critical competitive factor. Robots must achieve a seamless integration of models, data, and applications to effectively address the complexities of real-world scenarios. She emphasized that embodied AI models are evolving from language comprehension to predicting and interacting with the physical environment. By leveraging the Baige AI computing platform, Baidu has constructed a comprehensive AI infrastructure to streamline model engineering processes. At the data level, corporate competition will center on accumulating effective task duration—a metric reflecting how long AI systems can perform tasks successfully in real-world settings. Baidu Intelligent Cloud is fostering industrial collaboration by establishing a closed-loop data system, ensuring continuous improvement and refinement. In terms of application deployment, embodied AI will initially prioritize scenarios with well-defined boundaries and measurable outcomes, such as industrial material handling. Commercial success will depend on metrics like task completion rates, with Baidu Intelligent Cloud offering infrastructure support and minimizing migration costs through standardization.