Figuring out humanoid logistics at Spartanburg

BMW has taken the next step in testing humanoid robots for manufacturing and logistics deployment with the introduction of the Figure 03 robot, which is being tested for use on complex parts-sequencing tasks at the carmaker’s Spartanburg plant in the US.   

The robot is made by Figure AI, a California-based robotics and AI software company founded in 2022. Figure AI is making the robots at a highly integrated manufacturing facility called BotQ in San Jose, with an annual production capacity of 12,000 units. This year, Figure AI will be using its own robots on the production line at BotQ to build the next generation of the Figure. 

At Spartanburg, Figure 03 is picking unsorted components from large containers delivered to the Spartanburg plant and sorting then onto a sequencing trolley. It is the first deployment of a humanoid robot in a logistics workflow for BMW at any of its global locations. Figure AI says the robot can move fluidly between precise picking and placement of individual parts, and forceful manipulation requiring whole-body coordination.  

BMW says the test focuses on validating Figure 03 in a real logistics sequencing application. “In this use case, delivered components arrive unsorted in larger containers, and the robot picks them and sorts them into a sequencing trolley for just-in-sequence delivery to assembly,” explains a spokesperson for BMW. 

Birthplace of humanoids 

Figure 03 follows its predecessor Figure 02 into the Spartanburg plant. The Figure 02 was used in the body shop to insert sheet metal parts into specific fixtures, which were then welded into the vehicle chassis. That task requires speed and accuracy, and can be physically demanding. The Figure 02 supported the production of more than 30,000 BMW X3 vehicles for just under a year and in the process gathered important data on efficient handling and execution. 

“Plant Spartanburg is the birthplace of humanoid robotics in BMW Manufacturing’s operational day-to-day activities,” said Ulrich Wieland, vice-president of production control and logistics, BMW Manufacturing in an earlier statement. “Having already successfully completed a pilot with Figure 02 in our body shop, we are now looking forward to deploying Figure 03 for a sequencing use case in logistics.”  

According to BMW, the Figure 02 deployment provided important experience under real production conditions and demonstrated that humanoid robots can safely perform precise, repeatable work steps. “The learnings from the pilot support the further development of task execution, accuracy, repeatability, safety and integration into production workflows,” says BMW’s spokesperson.  

Brett Adcock, founder and CEO of Figure AI, states that the 11-month deployment of Figure 02 proved that humanoids are no longer lab experiments but can be a valuable asset in establishing a flexible, reliable manufacturing workforce. “We are excited to continue our work in Spartanburg as Figure tackles the complexity of the assembly and logistics hall,” he says. 

Robot development runs in parallel at the Spartanburg plant and at Figure AI, allowing findings from the real production environment to inform the further development of the robot. Figure 03 uses the Helix 02 vision-language-action (VLA) model, which enables the robot to understand its environment and carry out tasks. 

Feeling with vision 

Figure 03 introduces several new features that expand on the robot’s applications. They include soft components designed for enhanced safety, wireless charging designed for higher availability and audio functions for speech-to-speech communication. The wireless charging is designed to increase availability with two robots working in tandem. Currently the Figure 03 can work for 4-5 hours in a row and while charging another robot can take over. 

The robot also features improved hands with tactile sensors and palm cameras designed to increase precision and dexterity. Figure AI says dynamic whole-body control is powered by its Helix 02 VLA model.  

The Figure 03’s onboard AI acts as a digital brain, translating what the robot sees and hears into safe, continuous physical movements. The AI coordinates Figure 03’s hands, arms, torso and feet, allowing the robot to manipulate parts while stepping and repositioning its body to pull a heavy cart on caster wheels. Helix enables Figure 03 to learn and improve over time as it acquires new skills. 

According to Figure, sequencing requires reactive motion based on accurate live perception by the robot. The containers and bins the parts are delivered in do not arrive in exactly the same position and may have shifted in transit. That requires Figure 03 to gauge the scene and make corrections as it goes. That is made more challenging with the simultaneous need for manipulation and ground movement. In its logistics tasks Figure 03 needs to grasp parts with both hands while adjusting the placement of its feet, maintaining reach at the same time as balance. It needs to precisely place each part into the correct slot. According to Figure AI, the Helix 02 uses high-frequency visual-motor control, continuously adapting the robots movement to special variation and correcting small errors as the task unfolds. 

Automation that is safe 

Once Figure 03 has placed the parts on the trolley, an automated tugger train or a smart transport robot (STR) picks up the trolley of parts from a designated collection location and moves them to lineside where they are given to assembly employees just-in-sequence.  

BMW says the use of humanoid robots is part of its broader strategy to expand the automation of assembly plant tasks with Physical AI, specifically those jobs that are monotonous, ergonomically demanding or have a critical safety aspect.  

“Figure 03 is being integrated into real production logistics processes at Plant Spartanburg,” says BMW’s spokesperson, adding that high occupational safety standards apply at the workstations, particularly regarding the employees. “These standards are regularly evaluated, and occupational safety experts are involved at an early stage regarding aspects such as clearances and spacing.”  

Virtual Spartanburg 

The testing of the humanoid robots and the use of Physical AI is closely linked to the digital transformation of production at the Spartanburg plant. The carmaker has expanded and updated Hall 52 at the plant where the BMW X3 and its variants are made, and where the electric BMW iX5 will be assembled from the end of 2026. BMW refers to the digital implementation for production operations as iFactory. Virtual 3D simulations are used to help optimise processes and avoid errors from the outset – a digital twin of operations. 

BMW says that its Virtual Factory helps manage complexity and refines manual processes from planning through to the production line. That includes simulating worker movements with the key objective of improving employee ergonomics. 

“BMW Virtual Factory links planning and production through virtual 3D simulation,” says BMW’s spokesperson. “At Plant Spartanburg, Hall 52 is already visualised and simulated in 3D, allowing process planners to analyse and optimise assembly processes before and during implementation. This helps refine manual processes, improve ergonomics and identify opportunities for automation at an early stage.” 

BMW Virtual Factory is used to refine manual processes, improve ergonomics and identify automation potential. While a direct link between the simulation of employee movements and the programming or training of Figure humanoids is not part of the current scope there is potential for that to happen. 

Looking forward, BMW says the logistics sequencing use case is scalable in principle, as sequencing is a common process in automotive production logistics. “The focus at this stage is on validating the application and gathering operational experience rather than communicating target quantities for future robot deployment,” according to BMW. 

AI is also being used in quality assurance at Spartanburg. BMW is using a platform called Artificial Intelligence Quality Next (AIQX) which uses camera systems and sensors during line operations and provides line employees with immediate feedback via smart devices. The technology is supporting visual and acoustic quality inspections and helping identify deviations directly in the process. BMW has established AIQX as a standard and is assessing options to make the system available to suppliers as well.