Nvidia unveils GR00T 2 model that learns robot tasks from human videos

Nvidia's GR00T 2 AI model transfers human hand movements from videos to robotic arms, teaching robots new tasks without teleoperation, using training data from about 40 hours of video.

Read the full story at The Decoder ↗

Timeline · 1 report

2026-11-10 18:41 · The Decoder
Nvidia's model GR00T 2 watches humans on video, then imagines itself as the robot

More stories

NVIDIA unveils GR00T 2 and Isaac Lab 3.0 for robot learning at CoRL 2026 — Unite.AI
Nvidia CEO Jensen Huang Admits Affair With Woman Identified as Chinese Spy — Axios AI+
Anthropic targets roughly $1.3-1.4 trillion valuation in landmark IPO — CNBC Technology
Nvidia board probes CEO Jensen Huang as chip stock turbulence deepens — South China Morning Post Tech
Google adds TPU 8tx and 8i Max to eighth-generation AI chip lineup — Google: The Keyword
AI training startup Mercor raises $750 million at $20 billion valuation — Economic Times Tech
Hot inflation data revives December rate-hike expectations, weighing on stocks — Reuters AI
Local AI inference gains ground as compact desktop hardware and new Mac tooling emerge — DEV Community — Machine Learning

Get the daily brief of stories like this at 6:30 every morning →