Omniverse Libraries Expand AI Agent Toolkit for Physical AI Workflows

TLDR

• NVIDIA expands Agent Toolkit with Omniverse libraries for physical AI
• NVDA adds simulation-ready Omniverse tools for AI development
• NVIDIA launches Blender blueprint and new Omniverse libraries
• NVDA broadens AI toolkit with physics and sensor simulation tools
• NVIDIA grows physical AI platform with open-source Omniverse tools

NVIDIA Corporation expanded its Agent Toolkit by adding Omniverse libraries for physical AI development. NVDA stock traded at $205.17, up $2.36, or 1.16%, during the latest session. The update adds simulation-ready workflows and open-source tools for software developers building AI-powered applications.

NVIDIA Corporation, NVDA

NVIDIA extends Agent Toolkit with Omniverse libraries

The expanded toolkit now includes Omniverse libraries that prepare 3D assets for simulation environments. The software components support AI agents with physics, validation, and sensor simulation capabilities.Developers can integrate physical AI functions into existing applications more efficiently.

The release introduces ovrtx for RTX sensor simulation and ovphysx for GPU-accelerated physics. NVIDIA also added CAD-to-SimReady skills that convert CAD files into OpenUSD simulation assets. All three components are available through GitHub under an open-source distribution model.

NVIDIA also released a Blender integration blueprint alongside a SimReady sample workflow. These additions help developers extend familiar content creation tools with simulation features. The workflows support deployment across both cloud systems and local hardware platforms.

Physical AI systems require digital testing before operating in real environments. Robots, factories, and autonomous machines depend on simulation during development and validation. Simulation-ready assets remain a critical part of modern AI workflows.

Partners adopt Omniverse libraries across development platforms

Several software companies have integrated the new Omniverse libraries into their products. SideFX, PTC, ForgeCAD, Lightwheel, Moonlake AI, and Palatial are among the early adopters. Their implementations extend physical AI workflows into established engineering and design software.

SideFX is incorporating OpenUSD workflows together with ovrtx and ovphysx inside Houdini. The integration supports procedural content generation and simulation preparation within the application. Technical artists can also validate digital environments before simulation begins.

PTC integrated OpenUSD and ovrtx into its Onshape CAD and product data management platform. The connection links engineering design, collaboration, and simulation within one workflow. This approach simplifies product validation during development.

NVIDIA also demonstrated a SimReady Blender workflow during SIGGRAPH. The demonstration combines Omniverse libraries with NemoClaw to support simulation-ready asset preparation. Developers can add sensor simulation, physics, and validation inside Blender.

Open-source release strengthens NVIDIA’s software ecosystem

NVIDIA made the Omniverse libraries openly available through GitHub for software developers. Open availability encourages broader adoption across simulation, robotics, and industrial AI projects. Developers can integrate these components without changing existing production workflows.

The CAD-to-SimReady tools convert engineering models into OpenUSD assets for simulation environments. Ovrtx generates virtual camera, lidar, and radar outputs from digital scenes. Ovphysx applies GPU-accelerated physics to simulate realistic object behavior.

Several startups have already adopted these capabilities in commercial software products. Palatial automates SimReady asset creation from CAD files using NVIDIA’s conversion tools. Lightwheel applies OpenUSD-powered content agents to generate simulation-ready assets from text prompts.

ForgeCAD and Moonlake AI are also integrating Omniverse libraries into their software platforms. Their implementations support AI-assisted asset preparation and scene validation workflows. These additions expand the ecosystem around NVIDIA’s physical AI software strategy.