Humanoid robots 'the future' of car making, says BMW

BMW will deploy humanoid robots for car manufacturing in Europe for the first time this summer, a move its digitalisation chief calls 'the future of automotive production' [1]. Two robots named Aeon, built by Hexagon Robotics, are in a test deployment at BMW's Leipzig factory [1]. Standing 1.65m tall and weighing 60kg, the wheeled robots can carry 15kg for short periods and are equipped with 21 sensors [1]. Their initial tasks will be feeding parts to tools and performing pick-and-place operations for battery assembly [1]. BMW has prior experience with humanoid robots at its Spartanburg, US plant, where a Figure O2 robot helped build 30,000 X3 models [1]. The global automotive industry's pivot toward advanced automation is part of a broader technological shift, with companies like Toyota, the world's largest automaker producing about 10 million vehicles per year, also investing in new production technologies [1][4]. The cost of such robots has fallen, making it more economical to integrate them into existing human-centric workflows rather than redesigning entire assembly lines [1]. Analyst Bill Ray of Gartner notes that when a robot cost 17 million, factories were reorganized around it, but that is no longer the case [1]. Training for these robots combines teleoperation and simulation using reinforcement learning, with Hexagon's president, Arnaud Robert, stating that imitation learning can cut training time from months to days [1]. The robots have a three-hour battery life but are designed to swap their own batteries within three minutes to accommodate an eight-hour shift [1]. While addressing potential labor shortages, BMW's Michael Nikolaides argues that, like the automation of the 1970s, this technology will create new jobs rather than eliminate them [1]. Other manufacturers are following suit; Hyundai, for instance, is using Boston Dynamics' Spot and Atlas robots, while Toyota plans to deploy Agility Robotics' Digit humanoids [1]. The industry's move toward physical AI and robotics mirrors the ongoing, complex development of autonomous driving software, which faces significant hurdles in achieving full autonomy across all conditions [2].