Unitree's Wang Xingxing: Robot generalization capability is nearing a breakthrough; robots doing work in the home may happen sooner than expected
Wang Xingxing, chairman of Unitree Robotics, judged at the APEC CEO Summit in Shenzhen that the robotics industry's most difficult generalization capability challenge has reached a breakthrough tipping point, and that robots entering millions of households to actually do work may happen sooner than expected. He also cautioned that a model breakthrough is only the first half; if AI models make a major breakthrough next year and Unitree scales from tens of thousands of units a year to millions or even tens of millions, manufacturing, reliability and after-sales service will all come under pressure. In the first half of 2026, global humanoid robot shipments totaled about 19,100 units, up 272% year on year, with Chinese manufacturers accounting for more than 97%; Unitree ranked second globally with 5,900 units.
Body · AI translation
Add crowdsourced testing promotional copy Add crowdsourced testing promotional copy end Add App-exclusive traffic diversion
Kuai Technology News, November 18: Recently, Wang Xingxing, Chairman of Unitree Technology, appeared at the APEC CEO Summit in Shenzhen and, during the seminar session "Building an Open and Inclusive AI Innovation Ecosystem," offered his judgment: the robotics industry's thorniest problem—generalization capability—has reached a breakthrough tipping point, and robots entering millions of households to actually do work may happen sooner than imagined.
The theme of this year's summit is "Promoting Prosperity Together, Winning the Future Together," attracting more than 2,000 political and business representatives.
Wang Xingxing described 2026 as "the most exciting," saying that the AI and robotics achievements the public sees are only above the water's surface, while below the surface technological iteration is surging, with breakthroughs happening much faster every day, and the direction for the coming years is already clear.
The difficulty lies in the fact that robots face constantly changing spaces, objects, and tasks, and changing to a new scenario requires retraining once. However, combined with this year's global technological progress, the tipping point may come relatively soon.
Wang Xingxing also cautioned that model breakthroughs are only the first half. The question he often thinks about recently is what to do if AI models achieve a major breakthrough next year: Unitree currently produces only tens of thousands of units a year, and if in the future it suddenly needs to produce millions or even tens of millions of units, manufacturing, product reliability, and after-sales systems will all come under enormous pressure, and computing power demand may far exceed most people's imagination.
Industry data is confirming this enthusiasm. Institutional statistics show that in the first half of 2026, global humanoid robot shipments were about 19,100 units, up 272% year on year, with Chinese manufacturers accounting for more than 97%, and Unitree ranking second globally with 5,900 units.
Unitree itself is also speeding up. The company listed on the STAR Market on August 19, becoming the A-share "first embodied intelligence stock," with revenue of 1.817 billion yuan in the first three quarters, up 55.71% year on year, and net profit attributable to parent company of 387 million yuan, up 270.54% year on year.
Wang Xingxing's expectations have also been moving forward all along. In August at the World Robot Conference, he still expected the ChatGPT moment for embodied intelligence to arrive in 2 to 3 years at the earliest; by October, that had shortened to the next 1 to 2 years, with confidence coming from actual test feedback that GPT-6 took over the G1 robot to copy and paint, exceeding expectations.
Wang Xingxing finally called on all parties to make the cake bigger through cooperation and to focus on the future scale of the industry rather than the current market share.
[This article ends] If reprinting is needed, please be sure to indicate the source: Kuai Technology
Editor in charge: Shuangyue article_adlist[
Report article content ]article_adlist Statement: Sina.com exclusive manuscript, unauthorized reproduction prohibited. tech_web_article_v2015_block_bottom Non-targeted 300*250 button 17/09 wenjing begin
Non-targeted 300*250 button end
AI Trends