BMW Deploys Figure 03 Humanoid in Spartanburg

Figure 03 just added hands and palm cameras to BMW’s logistics line. BMW Group is doubling down on its humanoid program after a successful in-plant run with Figure 02 that proved the robots can handle high speed and accuracy dependent tasks in production.

Plant Spartanburg is the birthplace of humanoid robotics in BMW Manufacturing’s operational day to day activities, Ulrich Wieland, vice president of production control and logistics, said as the automaker confirmed the next step in its rollout. After Figure 02 completed a body shop pilot, producing more than 30,000 BMW X3s and performing welding related sheet-metal insertion, BMW announced it will deploy Figure 03 for a sequencing use case in logistics. The company’s 11-month deployment of Figure 02 was pitched as proof that humanoids are no longer lab experiments but a viable, flexible backbone for a modern manufacturing workforce, according to Figure AI founder and CEO Brett Adcock.

Figure 03 broadens the robot’s repertoire beyond the prior version. The newer robot introduces soft components designed for enhanced safety, and new perception and manipulation features, including tactile-sensor hands, palm cameras, wireless charging and speech-to-speech audio that lets the system communicate with humans and other devices. These enhancements are not cosmetic, they map to practical needs in busy production environments where robots must handle parts with care, navigate around people, and stay-powered through longer shifts without frequent human intervention.

The findings BMW cites from its first manufacturing deployment form the basis for Figure 03’s next steps. In Spartanburg, the body shop tasks that tested the earlier robot (heavy sheet-metal handling and high-speed precision in welding workflows) proved that a humanoid can contribute meaningfully to a traditional automotive line rather than sit on the margins as a curiosity. The leap to logistics sequencing signals BMW’s intent to push humanoids into tasks that sit between the line and the warehouse, where flexible sequencing, part routing, and line-side support can reduce downtime and improve throughput.

From a practitioner’s lens, the BMW case offers tangible signals about what it takes to move humanoids from pilot to production-like settings. First, perception and manipulation are now core enablers, not afterthought add-ons. Palm cameras and tactile sensing address the long-standing bottleneck of gripping and handling parts safely at speed, but they also introduce new maintenance and calibration demands. Second, power strategy matters. Wireless charging is cited as a feature, but operators will want to know how often a robot requires recharging in a work cell and how that downtime stacks against production goals. Third, integration with existing control and logistics software remains a real constraint. The robot must speak the plant’s language to understand sequencing priorities, work orders, and safety protocols, which means robust software interfaces and clear human-robot handoffs become essential, not optional.

Looking ahead, observers should watch whether Figure 03 expands beyond sequencing in logistics to other high-mix, high-velocity tasks in the assembly hall. BMW’s Spartanburg experience suggests the path to broader deployment lies in careful tuning of perception, safety, and power, combined with a clear plan for sustaining operations over long shifts and varied tasks.