BMW is set to become the first carmaker to use humanoid robots in European car manufacturing, with two machines planned for production from this summer.
The two robots, made by Hexagon Robotics, are currently undergoing a test deployment at BMW’s factory in Leipzig ahead of their full production role.
Michael Nikolaides, head of process management and digitalisation at BMW, described the development as a turning point for the industry.
“This will be the future of automotive production,” said Nikolaides, pointing to the robots’ ability to integrate into existing workplace environments.
The move towards human-shaped robots is driven partly by economics, as the cost of robots has fallen while redesigning assembly lines remains expensive.
“When a robot costs 17 million, you’d re-organise your factory around the robot, but it doesn’t anymore,” said Bill Ray, distinguished VP analyst at Gartner, adding: “So now you want to fit it into your existing way of working.”
The robot, named Aeon, stands 1.65m tall, weighs 60kg, and is equipped with 21 sensors including cameras, radar, a microphone, and force and torque sensors.
Aeon has a top speed of 2.4 metres per second and can carry 15kg for short periods, or 8kg continuously during sustained operation.
At BMW, the robots were trained using teleoperation, simulation inside a digital twin of the factory, and reinforcement learning supported by software from Nvidia.
Arnaud Robert, president of robotics at Hexagon, said imitation learning represents one of the most exciting frontiers in physical AI development.
Robert said imitation learning can reduce training time from months to just days, with the robot learning by observing human movement through video or motion sensors.
“The best translation [from the human to the robot] is when the teacher and the student have the same form factor,” Robert said.
Asked whether a robot could simply watch someone packing boxes and then join in, Robert described it as “the ultimate scenario,” adding: “You’re describing probably something that’s a year or two out.”
Ray at Gartner estimated that within three to five years, robots will be capable of following simple voice instructions to carry out tasks effectively.
Aeon’s battery life is limited to three hours, so the robot has been designed to swap its own battery in around three minutes, including travel to and from its charging station.
At BMW, the robots will feed parts to manufacturing tools and carry out pick-and-place tasks for battery assembly, roles that are repetitive and physically demanding for human workers.
Nikolaides noted that humanoid robots could also help address a looming labour shortage, saying: “We know that staff will be short in a matter of years, and humanised robots help.”
He drew a comparison to the automation of car production in the 1970s, arguing that new technology historically creates rather than destroys jobs.
BMW already has experience with humanoid robots at its Spartanburg facility in the US, where the Figure O2 robot helped build 30,000 model X3 cars at a pace matching that of a human worker.
A notable finding from the US deployment was that AI-based robots handled variance far better than traditional machinery, continuing to function even when components were slightly shifted or tilted.
Unlike the Figure robot, which walks, Aeon uses wheels, which Nikolaides said makes more practical sense on a factory floor.
Ray cautioned that humanoid robots have been significantly overhyped, stating: “The primary use case for a humanoid robot today is to walk on stage and artificially inflate your share price.”
He warned that seeing a humanoid robot walk leads people to assume far greater capabilities, saying: “Your brain fills in those gaps. We’re having unrealistic expectations when people deploy these robots.”