- NVIDIA has introduced significant updates to its OpenUSD framework, integrating generative AI models and NIM microservices.
- The new NIM microservices include tools for generating OpenUSD-Python code, searching through data libraries, and validating file compatibility.
- Upcoming microservices will offer features such as scene assembly from text prompts and realistic material application for CAD objects.
- Major industry players like Foxconn and WPP are already leveraging these advancements for digital twin creation and enhanced content production.
- New connectors and APIs will facilitate data integration with Siemens’ Simcenter and enable streaming to Apple Vision Pro.
- The OpenUSD Exchange SDK will allow developers to build custom data connectors, expanding the ecosystem.
Main AI News:
NVIDIA has announced significant advancements to its Universal Scene Description (OpenUSD) framework, promising to broaden its adoption across diverse industries such as robotics, industrial design, and engineering. These innovations are set to propel the development of highly accurate virtual worlds and accelerate digital twin creation for the next wave of AI applications.
At the core of NVIDIA’s update are the newly introduced OpenUSD-based generative AI models and NIM microservices, integrated into the NVIDIA Omniverse™ platform. These tools are designed to enhance capabilities for visualizing complex projects and simulating environments, further advancing the integration of AI in industrial and robotic applications.
Key Innovations and Microservices
NVIDIA’s latest offerings include a suite of NIM microservices that are poised to transform how developers interact with OpenUSD. These services provide advanced functionalities such as:
- USD Code NIM Microservice: Generates OpenUSD-Python code based on user prompts and answers general questions about OpenUSD, facilitating the visualization of 3D data in applications like Pixar’s usdview and NVIDIA Omniverse Kit.
- USD Search NIM Microservice: Allows developers to search through extensive OpenUSD and 3D data libraries using natural language or image inputs.
- USD Validate NIM Microservice: Ensures file compatibility with OpenUSD release versions and produces RTX-rendered, path-traced images through NVIDIA Omniverse Cloud APIs.
Upcoming microservices include:
- USD Layout NIM Microservice: Assembles OpenUSD 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: Creates OpenUSD-based meshes from point-cloud data.
- fVDB Physics Super-Res NIM Microservice: Enhances physics simulations with AI-driven super resolution.
- fVDB NeRF-XL NIM Microservice: Produces large-scale neural radiance fields in OpenUSD using Omniverse Cloud APIs.
Industry Applications and Partnerships
Foxconn, a global manufacturing leader, is leveraging NVIDIA’s platform to develop a digital twin of a factory under construction. Zhe Shi, Chief Digital Officer at Foxconn, noted, “Digital twins will help us accelerate the next wave of industrial manufacturing and autonomous machines,” highlighting the transformative impact of NVIDIA’s technologies.
WPP, a prominent marketing and communications firm, is an early adopter of the USD Search and USD Code NIM microservices. The company integrates these tools into its AI-powered content creation pipeline, facilitating rapid and innovative production processes for clients like The Coca-Cola Company.
Expanding OpenUSD’s Reach
NVIDIA is also enhancing OpenUSD’s versatility with new connectors and developer tools. The collaboration with Siemens aims to integrate OpenUSD with the Simcenter portfolio, advancing simulation technology and decision-making in industrial contexts. Additionally, a new connector facilitates data transfer from Unified Robotics Description Format to OpenUSD, expanding applications in robotics design and simulation.
The release of the OpenUSD Exchange SDK will enable developers to create custom OpenUSD data connectors, further enriching the ecosystem. New APIs and tools are now available for streaming large-scale OpenUSD scenes to Apple Vision Pro via the NVIDIA Graphics Delivery Network, marking another leap in interoperability and advanced authoring capabilities.
Steve May, Chief Technology Officer of Pixar, emphasized the transformative potential of these updates, stating, “OpenUSD is revolutionizing the way we create and interact with 3D content.” The latest innovations by NVIDIA are expected to accelerate USD’s adoption, facilitating greater engagement with its ecosystem across various industries.
Conclusion:
NVIDIA’s latest enhancements to OpenUSD and the introduction of NIM microservices are set to revolutionize the 3D visualization and simulation markets. By integrating generative AI and expanding data interoperability, NVIDIA is driving innovation across various industries, including manufacturing, robotics, and creative content production. These advancements will likely accelerate the development of digital twins and physically accurate virtual worlds, offering new opportunities for companies to improve operational efficiencies and explore cutting-edge AI applications.
