NVIDIA expands Isaac stack for general-purpose robotics
NVIDIA expanded its Isaac robotics platform with a new runtime, an updated learning framework, datasets and simulation assets.
NVIDIA has expanded its Isaac platform with new software, datasets and reference hardware for general-purpose robotics.
What happened
The update introduces Isaac OS, an open-source runtime connecting AI models, sensors, controls and robot hardware. Isaac Lab 3.0 is now generally available with multi-GPU training and support for multiple physics engines.
NVIDIA also added a library containing more than 2,000 simulation-ready assets, a dataset covering varied human activities and several reference workflows for robot training and synthetic-data generation.
Agile ONE SN, a humanoid reference platform built around the stack, is planned for availability in early 2027.
Why it matters
Robotics developers currently combine tools from many vendors to move from simulation to deployment. A more integrated stack can reduce setup work and make experiments easier to reproduce.
The bigger picture
NVIDIA is extending its position beyond chips by providing more of the software layer used to build physical-AI systems. Broad adoption could make the Isaac stack an important development standard for robotics.