The warehouse robotics firm’s latest win recognises a zero-shot learning system deployed at a Schneider Electric facility in Shanghai
Beijing-headquartered Geekplus (HKEX: 2590.HK) has won the 2026 RBR50 Innovation Award for its Robot Arm Picking Station, an intelligent system that automates item picking, which is widely regarded as the last major bottleneck in warehouse automation.
The award, presented annually by The Robot Report at the Robotics Summit & Expo, recognises the 50 most innovative companies in global robotics. It is the fifth time Geekplus has appeared on the list, placing it alongside ABB, Amazon, Boston Dynamics and Nvidia among repeat honourees.
Tackling the picking bottleneck
Autonomous mobile robots have reshaped how goods move through warehouses over the past decade, but the moment an item must be identified, grasped and placed into an order container has stubbornly resisted automation.
Rapidly changing SKU catalogues, wide variation in product size and shape, and the cost of training conventional vision models for each new item have kept human hands at the centre of the process.
The Robot Arm Picking Station targets that bottleneck by pairing Geekplus’s proprietary embodied intelligence foundation model, Geek+ Brain, with zero-shot learning. This is a technique that allows the robotic arm to handle unfamiliar items without per-SKU training.
The system plugs directly into existing Geekplus AMR infrastructure, creating what the company describes as an end-to-end unmanned picking workflow.
The RBR50 judging panel noted the significance of the approach. In its citation, the panel said the system strengthens an already robust goods-to-person portfolio by automating what it called the most difficult step in warehouse workflows, and moves the company closer to fully autonomous, end-to-end warehouse operations.
The technology was validated at a Schneider Electric warehouse in Shanghai, building on an existing partnership between the two companies. According to Geekplus, the system doubled manual picking throughput, achieved accuracy of 99.99 percent or higher, reaching 100 percent in testing, and was production-ready within 48 hours of deployment.
The company says the station’s zero-shot learning capability eliminates the retraining cycle typically required when new SKUs are introduced, while an on-premise architecture ensures operational data remains on-site. The system was integrated into Schneider Electric’s existing warehouse network and production processes through Geekplus’s All-in-One software platform.
Broader ambitions
For Geekplus, the picking station represents a strategic evolution. The company has built its reputation on AMR-based goods-to-person systems for storage and transport, but automating the pick itself has been a missing link in achieving fully unmanned warehouse operations.
The Schneider Electric deployment is intended to serve as a replicable model for industrial warehousing applications. Geekplus has indicated it plans to expand the solution across additional industries and use cases, though the company has not yet disclosed specific deployment timelines or target sectors.
The award comes during an active period for Geekplus. Recent announcements include a partnership with OMLOG to automate luxury fashion logistics in Hong Kong, and a deal with Latin American systems integrator Mindugar to accelerate warehouse automation adoption across the region.
