Reward AI unveils OM-1 robot foundation model trained solely on human data, zero-shot to any robot
yifengzhu_ut · x · 2026-09-15
Reward AI introduced OM-1, its first robot foundation model that zero-shot generalizes to any embodiment: tabletop arms, industrial arms, and humanoids.
The key shift is the data source: OM-1 is trained entirely on human manipulation data captured via sensor-equipped gloves doing ordinary tasks — no teleoperation, no robot-collected data — reaching near human-level dexterity and efficiency, with multi-robot collaboration in scope. A demo shows a robot replicating a bartender's fruit-cutting from human video alone. Since the data entry point is the human hand, training data can be accumulated today for robots that don't yet exist, breaking the traditional embodiment-bound data limitation.
More from Embodied
- NVIDIA's MotionBricks robotics motion-generation model trends on Hugging Face — nvidia · 2026-10-31
- Agility Robotics to raise $465M ahead of Nasdaq debut as AGLT on Nov 5, Foxconn leads with $201M — xmercury_one · 2026-10-31
- Tesla to open Optimus humanoid robot reservations in early December, per source — Scobleizer · 2026-10-31
- New MolmoSpaces-CommonSense benchmark: no robot policy tops 30% on manipulation tasks — RanjayKrishna · 2026-10-31
- Neuralink Unveils Next-Gen R1 Surgical Robot That Implants Brain Chips — elonmusk · 2026-10-31
- Google demos Sprout, a semi-autonomous rover that hunts sunlight to care for office plants — AI_Andrew · 2026-10-31