NVIDIA announced a major expansion of its Isaac open robotics development platform at GTC Berlin on October 21, headlined by the Agile ONE SN, a ready-made humanoid reference platform built together with German startup Agile Robots. The package adds a new open-source runtime, a generally available learning framework, open datasets and reference workflows that, according to NVIDIA, give developers an integrated path from data to deployment for bringing general-purpose robots into the real world faster.

What Was Announced

The centerpiece is Agile ONE SN, which NVIDIA describes as a frontier humanoid reference platform for research. Built by Munich-based Agile Robots, the system ships preconfigured with NVIDIA Isaac GR00T software and runs on the NVIDIA Jetson Thor edge AI platform.

Robotics researchers and developers often struggle to integrate the separate simulation, training and deployment tools needed to take robot skills from development to the physical world, NVIDIA said. The Agile ONE SN is meant to remove that friction: the company says it cuts setup time from weeks to days.

Key specifications, per NVIDIA’s official announcement:

  • 1.40 meters (4.59 feet) tall, 40 kilograms (88 pounds), 31 degrees of freedom, 3-kilogram payload per arm
  • Dual WUJI Hand 2 five-finger hands, each with 20 degrees of freedom, 1 kHz force control and tactile sensing
  • NVIDIA Jetson Thor onboard compute delivering 2,070 FP4 teraflops and 128 GB of unified memory for onboard reasoning and real-time control
  • Multimodal sensing including head and wrist cameras, lidar sensors and IMUs, with planned integration of the NVIDIA Halos safety system

Stanford University and ETH Zurich have expressed interest in using the platform, according to NVIDIA. Agile ONE SN is scheduled to be available in early 2027.

“General-purpose robotics is advancing at extraordinary speed: the ChatGPT moment for robotics is fast approaching,” said Jensen Huang, founder and CEO of NVIDIA. “With NVIDIA Isaac, our full-stack open robotics platform, we are giving developers the models, data, simulation tools and computing systems to bring physical AI to the world’s largest industries.”

“The breakthrough for humanoids is usefulness at scale: one adaptable machine that learns many skills and works across environments built for people,” said Zhaopeng Chen, founder and CEO of Agile Robots. “Today, Agile Robots introduced Agile ONE SN, a full-stack, open reference humanoid built on NVIDIA Isaac GR00T, powered by Jetson Thor, giving developers a ready system to create, train and validate the capabilities needed to put humanoids to work.”

Agile ONE SN key specifications
Graphic: GPTOnline, using Agile Robots and NVIDIA assets

New Software: Isaac OS and Isaac Lab 3.0

Alongside the humanoid, NVIDIA announced Isaac OS, a new open-source runtime framework built on NVIDIA Holoscan that connects AI, sensors, controls and robot hardware, targeting deterministic, real-time performance from simulation to physical robots.

Isaac Lab 3.0 is now generally available: an open-source, agent-ready learning framework for training and refining robot policies from desktop to data center, using multi-GPU workflows and physics engines including Newton and NVIDIA PhysX.

NVIDIA also released three open-source reference workflows for NVIDIA DGX Station to accelerate robot data generation, training and evaluation: refining GR00T 1.7 models with reinforcement learning, training dexterous manipulation skills in Isaac Lab, and post-training NVIDIA Cosmos world models for synthetic data generation using ChatGPT Astra.

Data was the third pillar. The new SimReady Central Library brings more than 2,000 simulation-ready robotics and factory assets, and the Human 1M Dataset covers diverse human activities with synchronized three-view video and 3D hand tracking. Both are on Hugging Face, with the full Human 1M dataset open to Cosmos Coalition partners. NVIDIA also described Halos, a full-stack safety platform built on NVIDIA IGX Thor for moving autonomous vehicles and robots to safer real-world deployment at scale.

NVIDIA Isaac platform expansion overview
Graphic: GPTOnline, using NVIDIA assets

Europe’s Robotics Ecosystem Signs On

NVIDIA framed the expansion as an anchor for Europe’s robotics industry. Companies building on the open platform include ABB Robotics, KUKA Group, Sereact, Humanoid, Hexagon Robotics, Bear Robotics, Toyota Material Handling Europe, Idealworks, Analog Devices, Franka Robotics, Bellboy, Flexion, Schaeffler and NEURA Robotics, alongside Hugging Face and Pollen Robotics, according to The Circuitry.

Why It Matters

The announcement turns GTC Berlin into a robotics showcase: the Isaac expansion sits alongside the OpenShell reference architecture for safe AI agents and the DSX POD containerized AI factories that NVIDIA unveiled the same day. Together they position Isaac as NVIDIA’s full-stack path for “physical AI,” the company’s term for intelligence embodied in machines that operate in the physical world. An open reference humanoid shipping to researchers next year, plus generally available training frameworks, simulation assets and safety tooling, removes several of the integration hurdles that have kept general-purpose robots in the lab. It is also a strategic play for Europe: with Chinese manufacturers building the vast majority of today’s humanoids, NVIDIA is putting a German-made design at the center of its robot platform.

Sources