Toyota Weighs 400,000 Robots and $6.4 Billion a Year to Automate Its Factories

Toyota Motor estimates that fully automating its factories, group companies, and major suppliers could require around 400,000 robots and roughly 1 trillion yen, about $6.4 billion, in annual spending starting from 2028, according to Reuters reporting on investor discussions earlier this month. The company has not committed to the full program or said how many years spending would continue at that level.

The 400,000 figure includes both replacements for existing machines and new installations, spanning industrial robots, automated logistics, and future human-robot collaboration on factory floors.

Toyota's automation push builds on years of groundwork. At the Kamigo Plant, a piston assembly line that started robot operations in January 2025 has gone from three operators to fully automated. The plant first experimented with robots on sub-lines back in 2008, but overseas expansion exposed a skills gap: some foreign facilities lacked enough trained maintenance workers, forcing certain processes back to manual production while Japanese plant know-how was transferred abroad. Workers compensating with hand-built jigs and tools that doubled or tripled efficiency in manual processes later informed the design of newer automated systems, including the Kamigo piston line.

Current systems are learning to handle ambiguity. The KumiPro parts-picking robot, already running at Toyota Motor East Japan, uses cameras to identify loosely positioned components rather than requiring precise pre-placement, with force-feedback control compensating for recognition errors when inserting parts into tight spaces.

Toyota's Frontier Research Center frames the effort as physical AI: robots that sense their surroundings, adapt their actions to conditions, and operate autonomously. Object recognition and decision-making tech from the Human Support Robot program has already migrated into KumiPro. The next step is ELEY, the Embodied Learning robot for Enhanced Yield, a dual-arm machine on an omnidirectional base designed to cope with unexpected physical forces during contact tasks. Toyota flags long-duration reliability, positional repeatability, and robot-learning data infrastructure as the remaining hurdles, and plans to train ELEY on both successful and failed attempts in production-like conditions.

Simulation plays a growing role: thousands of humanoid robot instances train in parallel using reinforcement learning, with researchers deliberately varying sensor noise, floor friction, and actuator behavior to narrow the Sim2Real gap before transferring skills to hardware. Toyota Research Institute is also partnering with Boston Dynamics on AI control for the Atlas humanoid, which completed a walking, lifting, sorting, and packing sequence using a single Large Behavior Model in 2025.

The competitive context is intensifying. Japan was the world's second-largest industrial robot market in 2024 with about 450,500 units operating, and Hyundai Motor Group plans to deploy Atlas at its Georgia Metaplant from 2028, targeting capacity for 30,000 robots annually by 2028. Toyota's estimate is broader than IFR's industrial figures, but its scale signals that physical AI is moving from research demo to capex line item.