Youngjae Jang, Professor at KAIST and CEO of DAIM Research, emphasized at a symposium jointly hosted by the engineering academies of Korea, China, and Japan that the essence of manufacturing Physical AI lies not in making individual robots more intelligent, but in orchestration—integrating and operating the entire factory as a single intelligent system. He explained that, unlike commercial robots, manufacturing environments require heterogeneous equipment to work together seamlessly according to production schedules, and that overall equipment effectiveness (OEE) and production throughput across the entire factory should serve as the key performance metrics.

The Physical AI unmanned factory platform introduced at the symposium integrates and manages a wide range of equipment—including overhead hoist transport (OHT) systems, autonomous mobile robots (AMRs), and humanoid robots used in semiconductor, battery, and automotive manufacturing—under a single operational intelligence. The platform achieves autonomous optimization by combining reinforcement learning, digital twins, Vision-Language-Action (VLA) models, and Software-Defined Factory (SDF) technologies. Professor Jang also stressed the need for standardized interfaces that connect robots from different manufacturers and proposed building an open manufacturing Physical AI ecosystem by combining Korea's AI orchestration capabilities with the strengths of China and Japan.