The world has about 500,000 hours of embodied AI training data. AGIBOT just shipped a wearable kit that aims for tens of millions of hours this year — collected without ever powering on a robot.
What MEgo actually is
Three pieces of hardware plus a cloud engine. MEgo Gripper is a hand-held capture device with force and tactile sensors, shaped like a humanoid hand. MEgo View is a head-mounted vision and motion-capture rig. An operator wears both, walks into a real kitchen or warehouse, and does the task. Vision, joint trajectories, and contact forces stream out in sync. MEgo Engine reconstructs, auto-labels, and retargets the data to whichever robot body you’re training.
The pitch: decouple data collection from owning a $200K humanoid. Anyone with hands is a data labeler.
Why the API matters
MEgo Engine exposes interfaces for third-party AI teams. Pull multimodal data for VLA model training, drive remote teleop through the Gripper, or retarget trajectories across robot embodiments. AGIBOT spun out a new subsidiary, Maniformer, to operate this layer — chaired by Yao Maoqing, with a 10-billion-hour target by 2030.
Embodied AI is starved for data. MEgo bets that wearable rigs on humans scale faster than any robot fleet.
You Might Also Like
- Reolink Omvi x16 Reolink ai box at ces 2026 Cloud Free ai Surveillance Finally Hackable
- Openharness got an Open Source Idea let ai Agents Build it for you for Free
- Voicebox Turned my Macbook Into a Voice Cloning Studio for Free
- Tidy Just Turned Imessage Into my Personal Automation hub
- Ladybird Browser ai Assisted Rust Migration an Entire Javascript Engine Ported in two Weeks and Nobody Broke a Sweat
Top AI Product 
Leave a comment