Toyota announced on September 18 that it plans to deploy 400,000 in-house-developed robots across 60 factories worldwide starting in 2028, committing an estimated ¥1 trillion (approximately $6.4 billion) annually to the project, a budget covering the full group. The effort aims to determine whether an AI system trained on human demonstrations can inherit the embodied skills of 18,000 master craftspeople before those workers retire. The robot at the center is ELEY — Embodied Learning Robot for Enhanced Yield — a 50-kilogram (110-pound) wheeled platform with two-fingered grippers rather than a full anthropomorphic hand, powered by battery or direct cord supply; it does not walk. Toyota is primarily deploying in-house developed hardware running on in-house developed AI, with the Toyota Research Institute's diffusion-policy Large Behavior Model framework at its core. Toyota's engineers at the Frontier Research Center chose wheels after a decade of lessons with ELEY's predecessor, the Human Support Robot (HSR), whose arm would damage itself or malfunction when contact produced unexpected external force. ELEY uses a quasi-direct drive (QDD) actuator architecture pairing a high-torque motor with a low ratio of 10:1, improving backdrivability so the joint yields to external force rather than resisting it.