MENU

NVIDIA Isaac Platform: Humanoid Reference and Learning Frameworks

NVIDIA Newsroom USA
Overview
NVIDIA has expanded its Isaac open robotics platform with the introduction of the Agile ONE SN reference humanoid and the Isaac Lab 3.0 learning framework. These tools, powered by Jetson Thor and GR00T, aim to accelerate the transition of general-purpose robots from simulation to real-world deployment.
In Depth

Key Findings

NVIDIA has expanded the Isaac open robotics development platform to provide an integrated pipeline for general-purpose robotics. The expansion includes the Agile ONE SN reference humanoid, the Isaac OS runtime, and Isaac Lab 3.0. These technologies are designed to reduce the time required to move robot skills from development to physical deployment.

Technical Details

The Agile ONE SN, developed by Agile Robots, is a full-stack reference humanoid preconfigured with NVIDIA Isaac GR00T and powered by the NVIDIA Jetson Thor edge AI platform. Its technical specifications include:

  • Physicality: 1.40 meters tall, weighing 40 kilograms, with 31 degrees of freedom and a 3-kilogram payload per arm.
  • Manipulation: Dual WUJI Hand 2 five-finger hands with 20 degrees of freedom each, featuring 1kHz force control and tactile sensing.
  • Compute: NVIDIA Jetson Thor delivering 2,070 FP4 teraflops and 128GB of unified memory.
  • Sensing: Head and wrist cameras, lidar sensors, and inertial measurement units.

On the software side, NVIDIA Isaac Lab 3.0 is now generally available as an open-source learning framework using multi-GPU workflows and physics engines including Newton and NVIDIA PhysX. Additionally, NVIDIA Isaac OS, built on NVIDIA Holoscan, provides a runtime framework for real-time performance. For data generation, the SimReady Central Library offers over 2,000 assets, and the Human 1M Dataset provides synchronized three-view video and 3D hand tracking.

Background & Context

The platform integrates several NVIDIA ecosystems: the GR00T foundation model for whole-body control, the NVIDIA Halos safety stack for real-world deployment, and NVIDIA Cosmos world models for synthetic data generation. Reference workflows for NVIDIA DGX Station now allow developers to refine GR00T 1.7 models using reinforcement learning and ChatGPT Astra.

Significance & Outlook

The Agile ONE SN is scheduled for availability in early 2027 and has drawn interest from ETH Zurich and Stanford University. The ecosystem is supported by a wide range of partners, including ABB Robotics, KUKA, and Universal Robots, signaling a move toward scalable, adaptable machines capable of working across human-centric environments.

Source: https://nvidianews.nvidia.com/news/nvidia-advances-isaac-platform-for-general-purpose-robotics-with-new-reference-humanoid-frameworks-simulation-pipelines-and-data

Get our weekly technology intelligence — free

Receive an infographic that lets you judge at a glance whether each field’s analysis report is worth reading.

Subscribe Free — Weekly Tech Intelligence

By subscribing, you’ll receive Troy-Technical’s weekly technology intelligence newsletter.

  • Your email and selected fields are used only to deliver the newsletter.
  • We never share your information with third parties.
  • You can unsubscribe anytime via the link in each email.

See our Privacy Policy for details.

Takes about a minute · Unsubscribe anytime

Published by Troy-Technical, an independent site run by one engineer with a career in materials development.
About the author / Contact info@troy-technical.jp
Let's share this post !

Author of this article

TOC