Photo credit: venturebeat.com
Nvidia recently announced that prominent industrial software and service providers are incorporating Omniverse into their platforms to accelerate the digitization of industries through advanced physical AI technologies.
This integration includes major players such as Ansys, Databricks, Dematic, Omron, SAP, Schneider Electric with ETAP, and Siemens, among others. This news was shared during Jensen Huang’s keynote address at the Nvidia GTC 2025 conference held in San Jose, California.
New blueprints for Nvidia Omniverse, linked to the Nvidia Cosmos models, have been introduced to facilitate the development of facilities equipped for robotics and to generate extensive synthetic data essential for advancing physical AI.
Rev Lebaredian, Nvidia’s Vice President of Omniverse simulation technology, remarked, “Omniverse serves as an operating system connecting the vast expanse of physical data with the domain of physical AI. This platform enables global leaders in industrial software, data, and professional services to collaborate, creating innovative applications that will push the advancement of AI in industries faster than ever before.”
New Blueprints Enable Robot-Ready Facilities and Large-Scale Synthetic Data Generation
The Mega blueprint, aimed at testing and managing multi-robot fleets within industrial digital twins, has now entered its preview phase on build.nvidia.com. Additionally, the Nvidia AI Blueprint designed for video search and summarization has been launched, leveraging the Nvidia Metropolis platform to develop AI agents capable of monitoring activities throughout large facilities.
Leaders in manufacturing are already employing these blueprints to enhance their operational efficiencies utilizing physical AI technologies. For instance, in the automotive sector, companies like Schaeffler and Accenture are utilizing the Mega blueprint to simulate and test fleets of Agility Robotics’ Digit for automating material handling. Meanwhile, Hyundai Motor Group is simulating Boston Dynamics’ Atlas robots on their assembly lines, and Mercedes-Benz is implementing it to optimize vehicle assembly processes using Apptronik’s Apollo humanoid robots.
In electronics, Pegatron is leveraging the Mega blueprint to create Nvidia Metropolis video analytics agents, ultimately improving factory operations and enhancing worker safety. Similarly, Foxconn is utilizing this blueprint to simulate various robotic systems in its manufacturing processes, aiming to innovate its facilities as part of its vision for the factories of the future, according to Brand Cheng, CEO of Fii, Foxconn’s core subsidiary.
For warehousing and supply chain solutions, companies like KION Group, Dematic, and Accenture are harnessing Mega to develop next-generation AI-driven automation. Idealworks is incorporating it into its fleet management software to optimize robotic systems, while SAP partners can create custom virtual environments for warehouse scenarios using Omniverse.
A new Omniverse Blueprint focused on AI factory digital twins enables data center engineers to design and simulate efficient factory layouts, cooling, and electrical systems. This initiative sees early partnerships with Cadence Reality Digital Twin Platform and Schneider Electric with ETAP, alongside the provision of Omniverse SimReady 3D models by Vertiv and Schneider Electric to enhance the development of AI factory digital twins.
The introduction of the Nvidia Isaac GR00T Blueprint for generating synthetic manipulation motions is also noteworthy, offering substantial advantages for robotics developers by allowing them to generate large-scale synthetic data from within Omniverse and Cosmos, thus significantly reducing data collection time.
Omniverse Physical AI Operating System Expands Across Industries
The digital transformation presents unique challenges for industries anchored in traditional physical processes, particularly as outdated systems generate vast amounts of data that lead to inefficiencies. Omniverse seeks to resolve these issues by serving as an operating system based on the OpenUSD framework, which facilitates the integration of physical-world data and applications.
Key companies such as Ansys, Cadence, Hexagon, Omron, Rockwell Automation, and Siemens are now incorporating Omniverse’s data interoperability and visualization capabilities into their existing industrial software and automation products. This effort aims to expedite product development cycles and improve manufacturing efficiency.
For those working within the physical AI realm, Intrinsic—an Alphabet subsidiary—is enhancing Omniverse workflows, utilizing Nvidia robotics foundation models to transition digital twins into tangible hardware applications. Furthermore, Databricks has formed a partnership to integrate Nvidia Omniverse with its Data Intelligence Platform, enhancing capabilities for large-scale synthetic data generation tailored for physical AI applications.
General Motors, the largest auto manufacturer in the U.S., has adopted Omniverse to refine its factory operations and train models for various processes including material handling and precision welding. On a different note, Unilever is leveraging Omniverse to optimize the creation of marketing content through accurate digital twins of its products.
Omniverse in Every Cloud
In a bid to streamline the creation and scaling of applications based on the OpenUSD framework, Nvidia has made Omniverse accessible as virtual desktop images on AWS EC2 G6e instances equipped with NVIDIA L40S GPUs. Additionally, the Microsoft Azure Marketplace now offers preconfigured Omniverse instances and streaming capabilities through Omniverse Kit App on Nvidia A10 GPUs, simplifying development and deployment for creators.
Looking ahead, Nvidia plans to extend these cloud-based developer tools and services to Oracle Cloud Infrastructure and Google Cloud, where they will be supported on newly released Nvidia RTX Pro 6000 Blackwell Server Edition, expected to launch later this year.
OpenUSD Unifies Robotics Workflows
During the GTC event, Nvidia unveiled the OpenUSD Asset Structure Pipeline for Robotics, developed in partnership with Disney Research and Intrinsic. This innovative structure seeks to unify various robotic workflows by applying best practices from OpenESt, establishing a standardized language for diverse data sources.
The GTC 2025 conference is set to continue until March 21, providing a platform for further discussions and demonstrations of these advancements.
Source
venturebeat.com