Key Takeaways
- Toyota plans to deploy 400,000 robots across 60 plants starting in 2028, supported by roughly $6.4 billion in annual manufacturing investment.
- Human workers train Toyota’s ELEY humanoid robot using jigs, a method Toyota calls “embodied learning,” to capture “takumi” artisan expertise.
- Toyota Executive Vice President Hiroki Nakajima has stated a human-robot coexistence goal, but Toyota has not disclosed humanoid numbers or job divisions. After roughly 1,500 training iterations over two weeks, Toyota’s ELEY humanoid robot folded T-shirts with near-perfect accuracy at a mid-September demonstration at the company’s European regional headquarters. That detail, a human teaching a machine by repetition, not by code, captures what separates Toyota’s automation strategy from simulation-first approaches: the knowledge being transferred belongs to the factory floor, and it moves through physical practice.
The 400,000-Robot Rollout
Toyota is targeting deployment of 400,000 robots across its global manufacturing operations starting in 2028. The plan splits roughly 150,000 units into Toyota’s own automotive plants and 250,000 into group companies handling components and materials production. Rollout is tied to factory renovations and the introduction of new vehicle models, not a standalone automation push, but integration into planned capital cycles.
The 60 production facilities involved span Toyota’s global footprint, and the company intends them to serve as live testing grounds for the new systems. Toyota has not disclosed what share of the 400,000 total will be humanoids like ELEY; the bulk of units across both segments are expected to be conventional industrial robots.
Workers Train ELEY with Jigs
ELEY is a battery-powered, wheeled humanoid, fitted with hands for picking components and folding fabrics.
The target knowledge base is Toyota’s pool of roughly 18,000 “takumi”, veteran artisans whose manufacturing expertise sits in muscle memory, not documentation. Capturing that expertise digitally, then distributing it across a fleet, is the operational logic behind what Toyota calls “embodied learning.” Data gathered from worker training sessions will be centralised and used to program additional robots, and, Toyota says, could eventually support onboarding for new human employees as well.
Coexistence or Job Redefinition?
Hiroki Nakajima, Toyota’s Executive Vice President, has spoken about human-robot coexistence rather than replacement in relation to the programme.
The debate over where humanoid robots fit in working environments extends well beyond Toyota. What makes Toyota’s case distinct is the data loop: as trained robots generate operational data, that data feeds back into programming new units, compressing the time between deployment and full capability. The long-term workforce math on that loop is not public.
$6.4 Billion Annual Investment
Toyota plans to commit roughly $6.4 billion (JPY 1 trillion) per year from 2028 to renovate and rebuild manufacturing facilities across the Toyota Motor Group, covering both its own plants and affiliated companies. The phased timeline aligns capital spend with factory upgrade cycles and new model introductions rather than treating robotics as a separate line item.
For context on what this kind of capital commitment signals at the hardware level, manufacturing investment at scale is reshaping supply chains across the AI and robotics sectors. Toyota’s investment covers the full group, not just its passenger vehicle operations, which means the automation push reaches into the supplier tiers that feed final assembly.
At the mid-September European demonstration, the T-shirt folding result came after two weeks and roughly 1,500 training cycles, a data point Toyota has chosen to make public as evidence of ELEY’s learning rate. How that rate scales from a controlled demo to 60 plants running concurrent production schedules remains the open engineering question.
Originally published at https://autonainews.com/toyota-plans-400000-robot-rollout-with-humans-training-eley/
Top comments (0)