NVIDIA Opens SimReady Central With 2,000 OpenUSD Assets
NVIDIA launches SimReady Central, a free library of 2,000 OpenUSD assets on Hugging Face for robotics simulations, AI factory twins and agent workflows.
Summary
On October 21, 2026, NVIDIA launched SimReady Central, a free, ungated Hugging Face library of more than 2,000 simulation-ready OpenUSD assets for physical AI training and AI factory digital twins. Unlike raw CAD or 3D models, assets carry profile-specific collision geometry, mass, friction, joints and limits, semantic labels, materials, sensor attributes, and electrical or thermal metadata. SimReady Foundation supplies specifications, profiles, and validation tools, while NVIDIA USD Search indexes USD properties and simulation metadata for natural-language discovery by developers and AI agents.
Contributors include Vertiv; robotics OEMs Comau, FANUC, Franka Robotics, Robotiq, Standard Bots, and Universal Robots; LG Electronics; ForgeCAD; and Palatial, whose catalog includes the NVIDIA Robotics 1K Dataset created with NVIDIA Robotics. OEMs manage their own Hugging Face datasets and can validate, package, and submit CAD-derived assets. OpenAI’s GPT-6 Astra is presented as an authoring agent that can find assets, generate OpenUSD scenes and Python code, produce reproducible Isaac Sim and PhysX pick-and-place previews, compare two rack and cooling layouts for overlaps and aisle clearance, and help prepare submissions. Developers can vary layered scenes, run Isaac Lab tests, connect power, cooling, BMS, and IT telemetry, or support projects such as Hugging Face and Pollen Robotics’ open-source Microduck. NVIDIA plans more partners and assets, with feedback requested on asset gaps, runtimes, bugs, and requirements.
Positives
- More than 2,000 free, ungated OpenUSD assets give developers a reusable base for robotics simulations and AI factory digital twins.
- SimReady profiles add physics, semantic, sensor, electrical, and thermal properties that raw CAD and conventional 3D models may lack.
- NVIDIA USD Search enables natural-language discovery across participating Hugging Face datasets using indexed simulation metadata.
- Vertiv, Comau, FANUC, LG Electronics, Palatial, ForgeCAD, and other suppliers broaden the available industrial, robotics, and household equipment.
- GPT-6 Astra can assist with asset discovery, scene assembly, Python generation, layout checks, and SimReady submission preparation.
- OEMs can manage their own Hugging Face datasets while distributing validated assets directly to simulation developers.
Risks & concerns
- Assets must match the correct SimReady profile, or their capabilities may not satisfy a developer’s simulation workflow.
- GPT-6 Astra’s pick-and-place and layout examples are described as expected workflows rather than documented production results.
- Power, cooling, and operational digital twins still require external equipment data, analysis tools, and live BMS and IT telemetry.
- Library coverage remains dependent on additional OEM contributions, with more partners and assets still on the roadmap.
- NVIDIA is requesting feedback on integration gaps, runtime support, bugs, and missing requirements, leaving some needs unresolved.