NVIDIA Unveils GR00T 2 and Isaac Lab 3.0 for Robot Manipulation Training

NVIDIA introduces open robotics tools at CoRL targeting precision manipulation and industrial deployment.

IT Solution Journal News Desk

Editorial illustration: A stylized editorial illustration of a segmented blue-grey robotic hand with glowing amber neural-network lines running along its fingers, gently holding a single illuminated egg at center, with a warm golden lattice network radiating upward from above.

AI-generated editorial illustration.

At the Conference on Robot Learning in Austin, NVIDIA announced GR00T 2, an open world action model built on NVIDIA Cosmos, trained on 100,000 hours of human egocentric data. The company says the model completed 40 of 50 evaluated tasks without task-specific demonstrations and reached 95% average success across 50 tasks with five demonstrations each. Isaac Lab 3.0 adds tighter integration with the Newton physics engine, enabling deformable-object simulation and, per NVIDIA claims, up to 2.5x training throughput improvement. Foxconn, Siemens and Nexuni are among partners evaluating these tools for factory and industrial applications.

Source: NVIDIA News. Original announcement: November 09, 2026.