Toyota’s $6.4bn robotics estimate puts physical AI in focus

Toyota Motor estimates that expanding automation across its factories, group companies, and major suppliers could require around 400,000 robots and annual spending of roughly ¥1 trillion, or $6.4 billion, from 2028.

According to AI News, citing Reuters, the estimate was discussed with investors. Toyota has not confirmed that the full investment will proceed or said how many years spending at that level might continue. The 400,000 figure includes both replacement equipment and new installations, spanning industrial and logistics systems as well as humanoid and non-humanoid robots.

From line automation to adaptable robots

Toyota already uses or tests robotics for parts transport, picking, and some medical applications. At its Kamigo Plant, a piston assembly line that began operating with robots in January 2025 has moved from three operators to full automation. The result followed years of iteration: shortages of skilled maintenance workers at some overseas plants had previously led Toyota to return certain processes to manual production while transferring manufacturing knowledge.

The company’s KumiPro picking robot uses cameras to identify and handle parts that have not been precisely positioned in advance and is operating on production lines at Toyota Motor East Japan. Force-feedback control helps compensate for camera-recognition errors when components must be inserted into tight spaces.

Physical AI in production environments

Toyota uses “physical AI” to describe systems that perceive their surroundings through sensors, optimise actions for the situation, and operate autonomously. Object recognition, decision-making, and motion technologies from its Human Support Robot programme have been applied to KumiPro.

Another project, ELEY, uses two human-like arms and an omnidirectional mobile base for tasks involving physical contact. Toyota says long-duration reliability, repeatable precision, and the data infrastructure needed for robot learning still require work. It plans to test ELEY under conditions closer to production and use both successful and unsuccessful attempts as training data.

Simulation and Boston Dynamics collaboration

Toyota is also training thousands of simulated humanoid robot instances in parallel through reinforcement learning. Because floor friction, sensor readings, and actuator behaviour differ between simulation and physical hardware, researchers vary those conditions during training and use real-robot data to narrow the Sim2Real gap.

Toyota Research Institute is separately working with Boston Dynamics on AI control for the Atlas humanoid. In 2025, the organisations demonstrated Atlas walking, lifting, sorting, and packing with a single Large Behavior Model, with skills added through human demonstrations. This research partnership is separate from Hyundai Motor Group’s controlling ownership of Boston Dynamics.

Market scale and comparison limits

Figures from the International Federation of Robotics cited in the report put Japan as the world’s second-largest industrial robot market in 2024, with about 450,500 robots in operation. Japan’s automotive industry installed roughly 13,000 industrial robots that year. Toyota’s 400,000 estimate is not directly comparable, however, because it includes replacement machines and multiple robot types across Toyota, group companies, and major suppliers.

The estimate therefore indicates the potential scale of manufacturing change rather than a confirmed spending programme. Delivery will depend on cost, reliability, maintenance capacity, skills transfer, and whether robots can work safely and consistently in real factory environments.