October 27, 2026 – Agibot’s heavy-duty humanoid robot, Elf G2 Max, has entered JD Logistics’ Zhi Lang warehouse to handle inbound tasks such as pallet breaking, stacking, box moving and standard container handling. The deployment marks one of the first applications of a humanoid robot in a real warehouse production environment.
On the hardware side, G2 Max can carry 20 kg with a single arm, 40 kg with both arms, and up to 50 kg at peak load. Its repeated positioning accuracy is kept within 0.5 mm, enabling stable handling of cartons and turnover boxes. In warehouse trials, the robot achieved an operation success rate above 99 percent, breaking down 80 boxes and stacking 75 boxes per hour. With hot-swappable dual batteries and autonomous recharging, it can operate around the clock without manual shift handovers.
Agibot said it chose JD Logistics’ Zhi Lang warehouse because the facility offers large-scale, real-world supply chain scenarios for testing industrial reliability. The inbound stage sits at the beginning of warehouse operations, and its stability directly affects downstream storage, transportation and sorting, especially during peak e-commerce periods.
Zhi Lang is JD Logistics’ self-developed “goods-to-person” robotic warehouse. It uses about 12 meters of vertical space, raising storage efficiency by roughly six times compared with traditional warehouses and improving outbound picking efficiency by about three times, with hundreds of robots working simultaneously.
G2 Max is also being used on multiple production lines, including Agibot’s Qiyuan T1 assembly line and consumer goods manufacturing lines. It natively supports Agibot’s Genie Studio 2.0 embodied operation platform, which allows faster task configuration and on-site adaptation.
JD Logistics has built a “Wolf Pack” robot portfolio covering storage, picking, transportation and delivery, including Zhi Lang, Yi Ling, Du Lang, Fei Lang, Tian Lang, Di Lang, Cang Ling, Mu Lang and Baomu Lang systems, coordinated by its Chao Nao large model. The deployment shows that humanoid robots are moving from demonstrations into actual industrial workflows, although large-scale commercial replication will still require further validation.
