Why Automakers Invest in Robotics: Focus on Hands That Turn Screws, Not Brains That Write Slides

Last week I visited an OEM plant in East China. In the final assembly workshop, two six-axis robotic arms were installing door seals. A production line engineer standing nearby told me this was already the third version of the solution—the first one came from some high-profile startup. Their motion control algorithms looked elegant, but after running continuously for two hours, the end-effector jitter exceeded 0.5mm, causing the seals to be misaligned. The second version switched to another company, using imported harmonic drives for the joint modules, but spare parts had a 12-week lead time, and every stoppage cost hundreds of thousands. The final adopted solution came from a small company whose founder used to be a smart manufacturing director at a car manufacturer. They used domestic reducers rated for 8,000 hours of life. After three months on the actual production line, the failure rate was only 0.3%.