Webp 2az79zvei5lgu7fitq98t73drrf0

NVIDIA announces advanced generative AI models for diverse industry applications

ORGANIZATIONS IN THIS STORY

Jensen Huang Founder, President and CEO at NVIDIA | Official website

NVIDIA today announced significant advancements to Universal Scene Description (OpenUSD) at SIGGRAPH, aimed at expanding the adoption of the 3D data interchange framework across robotics, industrial design, and engineering. These enhancements are expected to accelerate developers' capabilities in building highly accurate virtual worlds for AI evolution.

The new developments include OpenUSD-based generative AI and NVIDIA-accelerated frameworks built on the NVIDIA Omniverse™ platform. These tools will enable various industries to develop applications for visualizing industrial projects and simulating environments for future AI and robotic technologies.

Among the new offerings are NVIDIA NIM™ microservices designed for AI models that can generate OpenUSD language responses, Python code, apply materials to 3D objects, and understand 3D space and physics. Additionally, new USD connectors facilitate integration with robotics and industrial simulation data formats, allowing users to stream extensive datasets with NVIDIA RTX™ ray tracing to Apple Vision Pro.

Rev Lebaredian, vice president of Omniverse and simulation technology at NVIDIA, stated: "The generative AI boom for heavy industries is here. Until recently, digital worlds have been primarily used by creative industries; now, with the enhancements and accessibility NVIDIA NIM microservices are bringing to OpenUSD, industries of all kinds can build physically based virtual worlds and digital twins to drive innovation while preparing for the next wave of AI: robotics."

NVIDIA's first generative AI models for OpenUSD development will be available as NIM microservices. These models allow developers to incorporate generative AI copilots into USD workflows across various sectors such as manufacturing, automotive, and robotics.

The preview microservices include:

- USD Code NIM microservice: Answers general OpenUSD questions and generates Python code based on text prompts.

- USD Search NIM microservice: Allows natural language or image input searches through large libraries of OpenUSD data.

- USD Validate NIM microservice: Checks file compatibility against OpenUSD versions and generates rendered images using NVIDIA Omniverse Cloud APIs.

Upcoming microservices include:

- USD Layout NIM microservice: Assembles scenes from text prompts based on spatial intelligence.

- USD SmartMaterial NIM microservice: Predicts and applies realistic materials to CAD objects.

- fVDB Mesh Generation NIM microservice: Generates meshes from point-cloud data.

- fVDB Physics Super-Res NIM microservice: Performs super-resolution on frames for high-resolution physics simulations.

- fVDB NeRF-XL NIM microservice: Generates large-scale neural radiance fields using Omniverse Cloud APIs.

Foxconn is already utilizing these tools to create a digital twin of a factory under development. Zhe Shi, chief digital officer at Foxconn, remarked that "digital twins will help us accelerate the next wave of industrial manufacturing and autonomous machines."

WPP has also adopted USD Search and Code NIM services in its content creation pipeline. Stephan Pretorius, CTO at WPP said that "using NVIDIA NIM microservices with NVIDIA Omniverse has made it possible for us to launch innovative new production tools with companies like The Coca-Cola Company at unprecedented speed."

New USD connectors enhance interoperability across more industries by integrating robotics data formats with streaming capabilities to Apple Vision Pro. Siemens plans to integrate OpenUSD pipelines within its Simcenter portfolio to support decision-making processes among stakeholders.

Steve May from Pixar noted that "OpenUSD is revolutionizing the way we create and interact with 3D content," highlighting expectations for accelerated growth due to these new services by NVIDIA.

NVIDIA cofounded AOUSD last year alongside Pixar, Adobe, Apple, and Autodesk. Progress includes a new OpenUSD release specification along with additional members joining AOUSD.

Currently available in preview are the USD Search, Code, Validate NIM services via the NVIDIA API catalog. The URDF connector is accessible through NVIDIA Isaac Sim™. Developers can start integrating generative AI into their workflows using new Omniverse developer tools presented during SIGGRAPH in Denver until August 1st.

ORGANIZATIONS IN THIS STORY