On August 20, at the 2026 World Robotics Conference, Wang Xingxing, founder and chairman of Yushu Technology, gave a keynote speech on “From Exhibits to Products: The Next Decade of the Humanoid Robot Industry”. Wang Xingxing said that over the years, the company has actually been promoting robots to actually enter life, such as implementing applications in automobile factories, deploying robots in its own factories for simple tests, etc. In addition, the AI team is also promoting robots to work at home. However, he mentioned that since the overall efficiency and general usability capabilities are still not high enough at present, Yu Shu has not carried out large-scale promotion. Wang Xingxing hopes that moving towards an intelligent ChatGPT moment will ideally take 2-3 years, or 5-10 years longer. He believes that when this point is reached, robots can successfully complete about 80% of tasks through voice or text instructions in 80% of unfamiliar scenes. Currently, the core challenge is aligning the input and output of a large AI model with a real robot.

On August 20, at the 2026 World Robotics Conference, Wang Xingxing, founder and chairman of Yushu Technology, gave a keynote speech on “From Exhibits to Products: The Next Decade of the Humanoid Robot Industry”. Wang Xingxing said that over the years, the company has actually been promoting robots to actually enter life, such as implementing applications in automobile factories, deploying robots in its own factories for simple tests, etc. In addition, the AI team is also promoting robots to work at home. However, he mentioned that since the overall efficiency and general usability capabilities are still not high enough at present, Yu Shu has not carried out large-scale promotion. Wang Xingxing hopes that moving towards an intelligent ChatGPT moment will ideally take 2-3 years, or 5-10 years longer. He believes that when this point is reached, robots can successfully complete about 80% of tasks through voice or text instructions in 80% of unfamiliar scenes. Currently, the core challenge is aligning the input and output of a large AI model with a real robot.