The robots still can’t reliably do the job – but the valuations have already arrived. As of June 2026, eight Chinese embodied-AI companies have crossed a RMB 20 billion (about $2.8 billion) valuation, earning them the nickname the “RMB 20 billion club.”
Three bets, not one
The eight – Unitree, AgiBot, Galbot, Star Era, Qianxun, Zhiyuan, Zhixiang and Lingxin Qiaoshou – split into three camps. Unitree and AgiBot have the only real shipment volumes and run a “sell hardware” loop. Galbot, Star Era, Qianxun, Zhiyuan and Zhixiang bet on the long-term value of the “brain” – model capability and data moats. Lingxin Qiaoshou bets on the part, not the whole: it builds only dexterous hands.
Shipping vs. dreaming
By Omdia’s count, AgiBot shipped 5,168 humanoid robots in 2025 (about 39% global share); Unitree disclosed 5,500 dual-foot units. Unitree is already profitable – RMB 1.699 billion revenue, RMB 590 million net profit – and cleared its STAR Market IPO at roughly RMB 42 billion. AgiBot launched its Hong Kong listing in July. The “brain” camp ships far less – some still in the hundreds – but raises aggressively: Galbot pulled about RMB 7 billion across three years; Zhixiang closed nearly RMB 5 billion in one round; Zhiyuan drew Meituan, Xiaomi, Alibaba and ByteDance as lead backers across separate rounds.
Where the robots actually work
Commercial scenarios fall into three tiers. Research and education is mature but capped – Unitree’s humanoid revenue was 73.6% academic in early 2025. Industrial lines are the biggest prize but barely proven: Qianxun’s “Xiao Mo” runs on CATL battery lines at three times a human; Galbot’s S1 works CATL 7×24 for three-plus months. Consumer retail and home use – 58 Daojia’s cleaning bots, Galbot’s unmanned kiosks, JD’s in-store coffee robots – are the most imaginative but least mature.
Why these eight
Half the founders are academics (Peking, Tsinghua, Stanford); others are ex-big-tech scientists or hardware engineers. Behind them sit state capital (Galbot’s National AI Industry Fund), industrial giants (Meituan, Tencent, BYD, JD) and market VCs. In an industry with no proven model yet, financing capacity is a core competence – and the shareholder mix sets the market’s patience.
RMB 20 billion measures expectations, not results. Nobody – including the eight – knows which route wins. The number is an entry ticket; time will do the accounting.
Source (Chinese original): Sohu IT
Translated and adapted from Sohu IT (it.sohu.com).