Toyota turns its humanoid research into a factory-first automation plan
Toyota is formalizing a USD 6.4B-a-year factory automation plan: 400,000 robots from 2028, its own plants first, humanoids included.
Toyota is formalizing a factory automation program worth about USD 6.4 billion a year, built around roughly 400,000 robots from 2028, as it turns two decades of robotics research into a commercial plan. About 250,000 of those robots are earmarked for Toyota group companies and their suppliers, and a dedicated robotics team is being set up under the chief production officer. The humanoid effort sits inside that wider program rather than as a standalone robot business: the first customer is Toyota itself.
The factory humanoid at the center of the plan is an internally developed 50-kilogram machine designed to learn tasks by watching factory workers, using an in-house foundation model. Its first jobs are fine-motor factory work such as moving parts between boxes. The same research lineage extends beyond Toyota's own hardware: Toyota Research Institute's learning-model work, which turns high-level decisions into precise physical actions and real-time motor control, already feeds into Boston Dynamics' Atlas through a research tie-up announced in 2025.
Deployment is already happening through Walden Robotics, the Toyota Research Institute spinout that emerged from stealth in July with USD 300 million in funding at a USD 1.1 billion valuation. Its general-purpose robots are already working at Toyota plants in North America — the factory-first strategy in practice, ahead of any external robot business. Toyota's humanoid lineage runs from its basketball-playing CUE robot to these factory machines, a reminder that the company has been iterating on humanoid hardware for years.
The sequencing is deliberate. Toyota's head of humanoid robotics research, Tomohiro Nomi, said earlier this year that the highest priority is deploying robots in the company's own factories, where demand is real and the environment can be controlled. A new assembly plant near Nagoya — Toyota's first new assembly plant in Japan since 2012, due to begin operating in the early 2030s — gives the program another integration site. Bernstein analyst Masahiro Akita argues Toyota's advantage may lie less in any single machine than in integrating robots, AI and production processes into one operating system.
The competitive frame is clear: Tesla is ramping early Optimus production in Fremont with external deliveries possible by the end of 2027, while Chinese manufacturers dominate current humanoid shipment volumes. Goldman Sachs analysts estimate Toyota could capture 3 to 8 percent of the humanoid market by 2035, equivalent to 190,000 to 540,000 units — an analyst projection, not a Toyota commitment. Two cautions frame the story: the 400,000 figure covers robots of all kinds, not humanoids alone, and Toyota has disclosed no humanoid production volume or price. The test is whether factory deployments compound into a repeatable operating system, not how many robots a headline counts.