Wang Xingxing’s first public speech after going public: The real breakthrough for robots will occur when two “80% moments” occur | WRC 2026
2026-09-07 10:55Products & Apps🔥 42.2 heat score
1sources
1days unfolding
42.2heat score
4mentions
SummaryAI generated
On September 7, 2026, Wang Xingxing, founder of Yuzhi Technology, stated at the WRC 2026 conference that the true breakthrough in the embodied intelligence industry depends on achieving two critical ‘80% moments’. He pointed out that the main bottleneck facing the industry currently lies in the insufficient alignment between AI models and the real physical world, which severely affects the generalization ability of robots due to minor assembly errors. To address this, Wang Xingxing advocated using basic large models to build a closed-loop system for research and development, simulation, and testing, enabling physical AI robots to achieve ‘self-evolution’ instead of traditional manual fine-tuning. Additionally, the company has shifted from pre-training to real-time generation of motion strategies, continuously iterating on technologies for wheeled composite quadrupedal robots and humanoid robots. Such products have already been applied in automobile factories, such as manned aircraft.
Wang Xingxing, founder of Yushu Technology, stated at WRC 2026 that the critical point for the outbreak of embodied intelligence industry lies in two “80% moments”: approximately 80% of tasks can be completed through voice or text commands in 80% of unfamiliar scenarios. He believes that the current bottleneck is the insufficient alignment between AI model inputs and outputs and the physical world of real robots, resulting in significant装配 errors in the last few centimeters, which severely affects generalization ability. Wang Xingxing advocates using basic large models to build a closed loop for research and development, simulation, and testing, achieving “self-evolution” of physical AI robots and replacing manual fine-tuning. The company has already deployed applications in automobile factories and products such as manned aircraft, and is committed to shifting action generation from pre-training to real-time generation, while continuously iterating technologies for wheeled composite quadruped robots and humanoid robots.