Two humanoid dual-arm robots have been introduced by CJ Logistics at an Olive Young distribution facility in Yongin, South Korea, representing the first deployment of humanoid robots within active logistics operations in the country. Automated systems have taken over cushioning material placement on packaging lines, transitioning from pilot programmes last year into full commercial deployment.
Rigid automated systems in fulfilment facilities have long struggled to accommodate the broad range of goods and order types that warehouses process on any given day, and this development targets that operational gap. These robotic systems can be integrated across numerous operational stages without modifications to existing facilities, functioning within the same floor plans and equipment layouts originally built for human personnel.
At present, the pair of robots handle one operation — placing padding into shipping containers — though their remit is expected to broaden over time. The company intends to gradually introduce more advanced workflows — such as item retrieval, categorisation, and quality checks — drawing on performance data from the current rollout to refine the AI systems that guide its robotic fleet.
At the heart of this initiative sits the Robot Foundation Model (RFM), an AI system functioning as the robot's central processing intelligence. The system analyses incoming visual and sensor feeds to interpret its surroundings and select suitable responses on the fly. Real-time operational data from the warehouse floor is being fed back into RFM training to sharpen the model's accuracy and ability to handle unforeseen scenarios.
The broader goal is a multi-purpose humanoid robot for logistics operations — a single unit capable of performing several tasks in succession rather than being stationed at one fixed point. Unlike traditional automation — where individual machines are designed to handle specific, dedicated functions — this approach operates differently.
The company has brought together a group of third-party collaborators to support the technology. Robotis is supplying the hardware platform, Aidin Robotics is handling the robotic hand components, and Realworld AI is leading RFM development efforts. The partnership extends across every layer of the technology, from hardware infrastructure through to the artificial intelligence systems that manage operational behaviour.
Kim Jung-hee, head of CJ Logistics' TES Logistics Technology Research Institute, outlined the organisation's approach, stating that field-gathered logistics data would be combined with artificial intelligence to develop humanoids capable of independent decision-making across varied tasks, with the longer-term aim of establishing a fulfilment environment where human workers and robotic systems operate alongside one another.
The rollout indicates that humanoid robots are moving beyond controlled testing environments into live operational settings where output volumes and system reliability are measurable priorities. Whether humanoid robots can bridge the performance gap between rigid automated systems and the adaptable workforce that fulfilment centres continue to rely on remains a central question for operations leaders. Metrics gathered by CJ Logistics on task precision, transition speed, and error recovery will inform whether the approach can be replicated at scale — and whether rival companies choose to adopt it.
