CJ Logistics has deployed two humanoid dual-arm robots at an Olive Young distribution center in Yongin, South Korea, marking the first time humanoid robots have entered live logistics operations in the country. The robots are now handling cushioning material insertion on packaging lines, moving from last year's field trials into actual commercial processes.
The move addresses a persistent constraint in distribution centers: fixed automation equipment struggles with the variety of products and order profiles that warehouses handle daily. Humanoid robots can be deployed across multiple processes without requiring changes to existing infrastructure, using the same workstations and layouts designed for human workers.
From packaging to general-purpose logistics
The two robots are currently focused on a single task - inserting cushioning material into boxes - but the scope is set to expand. CJ Logistics plans to phase in more complex processes including product picking, sorting, and inspection, using operational data from this initial deployment to train the AI models that control the robots.
Central to this effort is the Robot Foundation Model (RFM), the AI system that acts as the robot's brain. RFM processes visual and sensor data to perceive the environment and determine appropriate actions in real time. The company is feeding live operational data back into RFM training to improve accuracy and responsiveness to unexpected situations on the warehouse floor.
The long-term target is a general-purpose logistics humanoid: a single unit capable of performing multiple processes sequentially rather than being dedicated to one station. This contrasts with conventional automation, where different machines are built for different tasks.
Collaboration across the robotics stack
CJ Logistics has assembled a set of external partners for the technology. Robotis is providing the hardware platform, Aidin Robotics is developing the robot hands, and Realworld AI is working on RFM development. The collaboration spans the full stack from physical components to the AI layer that governs behavior.
Kim Jung-hee, head of CJ Logistics' TES Logistics Technology Research Institute, described the direction: "By combining logistics data accumulated in the field with AI technology, we will refine humanoids so they can autonomously judge and perform diverse tasks, and in the long term, build a future fulfillment environment where humans and robots work together."
Why this matters for operations
This deployment signals that humanoid robots are beginning to prove their value outside of pilots and labs, in actual business processes where throughput and uptime matter. For operations leaders, the relevant metric is not the technology itself but whether humanoids can reduce the gap between fixed automation and the flexible labor that distribution centers still depend on. The data CJ Logistics gathers on task accuracy, changeover time, and failure recovery will determine whether the model scales - and whether competitors follow.
Your membership also unlocks:
