Moqi Intelligence CTO Argues for Large Model and Embodied AI Integration at 2026 World Internet Conference

At the WIC 2026 forum, Huang Qingqiu of Geekplus stated that large models alone cannot control robots due to latency and closed-loop needs. He emphasized using on-device execution, non-intrusive data collection like Ego and UMI, and real-world validation for embodied AI progress.

At the 2026 World Internet Conference in Wuzhen, Geekplus CTO Huang Qingqiu stated that large language models will not replace embodied AI because physical real-time closed-loop interaction remains a core capability difficult for digital models to simulate. He emphasized that high-quality, synchronized data from real-world operations is essential for model iteration and that future development requires simultaneous progress in both digital intelligence and physical hardware deployment.

At the 2026 World Internet Conference in Wuzhen, Geekplus CTO Huang Qingqiu stated that large language models will not replace embodied AI because physical real-time closed-loop interaction remains a core trait difficult for digital models to simulate. He emphasized that high-quality real-world operational data and whole-body coordination during movement are essential for scaling embodied AI across industries within the next three to five years.