Practice Makes Precise: At CoRL, NVIDIA and Partners Reveal How Robots Rehearse Delicate Work in Simulation
Reported by NVIDIA and IntraMind LLC2 sources ↓
Latest in this story: NVIDIA Launches GR00T 2 and Isaac Lab 3.0 at CoRL
At CoRL, NVIDIA introduced robotics models and simulation tools for training and evaluating delicate tasks such as aligning gears, connecting cables and assembling small components.
Key points
- NVIDIA introduced GR00T 2, an open model trained on robot and human data
- NVIDIA says GR00T 2 completed 40 of 50 tasks at least once without task-specific demonstrations
- NVIDIA reports 71% average success with one demonstration per task and 95% with five
- Isaac Lab 3.0 with Newton achieved up to 2.5 times the training throughput of an earlier setup
- Nexuni, Foxconn and Siemens are testing or evaluating GR00T 2 for robotics applications
Sources
Details
NVIDIA says GR00T 2 was trained on diverse robot and human data, including 100,000 hours of its open-source human egocentric data. The company reports that the model achieved a 71% average success rate across 50 tasks with one demonstration per task, rising to 95% with five. It also says the model completed 40 of 50 evaluated tasks at least once without task-specific demonstrations.
NVIDIA introduced Isaac Lab 3.0, an open-source robot-learning framework that can use the Newton or PhysX physics solver. The company says its Isaac Lab 3.0 and Newton setup achieved up to 2.5 times the training throughput of Isaac Lab 2.3 with PhysX in a vision-based dexterous-manipulation benchmark. Newton integration is intended to simulate deformable objects such as cables and cloth, as well as complex articulated systems.
The CoRL announcements also included RoboTTT, a robot model and training recipe that NVIDIA says scales visuomotor context three orders of magnitude beyond state-of-the-art policies without increasing inference latency. NVIDIA said two of its papers, including work on RoboTTT and a pipeline for generating simulation-ready scenes, were among six finalists for the conference’s Best Paper Awards. The company named Nexuni, Foxconn and Siemens as organizations testing or evaluating GR00T 2.
Written from coverage as of
Images
- NVIDIA
- NVIDIA
Stocks in this story
Change since open: 20 Nov to 21 Nov 2026
Last refresh