BMW elevates its AI humanoid robot strategy to include logistics

When people hear the term artificial intelligence , they usually think of chatbots or data analysis. But at BMW’s Spartanburg plant in the US, AI is now getting hands, legs, and eyes. Under the term physical AI , the automaker is integrating the new humanoid AI robt Figure 03 into its production logistics. The move comes as no surprise: For almost a year, BMW had the predecessor (Figure 02) welding body parts for more than 30,000 vehicles. The conclusion of this practical test: The machines can precisely perform monotonous, heavy tasks. Now the technology is leaving the testing phase and moving to where things get highly complex: logistics. The task: Transform chaos into order While its predecessor simply lifted sheets of metal, the further enhanced Figure 03 has to solve cognitive and tactile tasks. In logistics, it picks unsorted components from large boxes and sorts them into carts in the exact required order. Automated transport systems then take over, carrying them to the assembly line. To achieve this, the manufacturer has upgraded Figure AI. The new robot has: Cameras and tactile sensors directly in the palms of the hands for greater sensitivity Audio functions for true speech-to-speech communication in the factory hall Wireless charging for continuous, autonomous operation Softer components to increase safety for human colleagues At first glance, a humanoid robot might seem like a project solely for the production manager. That’s a misconception. This use case is relevant for everyone, and is highly relevant for CIOs. Figure 03 is ultimately nothing other than a highly complex, mobile edge client that has to process large amounts of data (video, audio, sensor data) locally and in real-time. These advanced robots, equipped with new capabilities, are taking on new tasks. BMW AG BMW is demonstrating in Spartanburg that such a robot works, but only in a fully digitized ecosystem like an automotive plant. This means that the IT department is the enabler for the production environment of the future. Virtual twins: Even before the first robot touches a box, BMW simulates Hall 52 and all movement sequences in a 3D “Virtual Factory.” IT provides the planning basis. AI Quality Control (AIQX): Error detection is performed using cameras and microphones along the production line. The algorithms perform visual and audible checks and send the feedback directly to the smart devices of human colleagues. Infrastructure scaling: When robots communicate via voice, charge wirelessly, and interact with autonomous transporters, the WLAN, 5G, and network backbone in the factory must have low latency and be fail-safe. On the one hand, the humanoid robot relieves BMW factory workers of physically demanding work; on the other hand, it forces the IT department to merge traditional IT infrastructure and factory technology (OT). “Physical AI” has thus arrived in everyday industrial practice.