 |
| Source: NVIDIA. Omniverse at work. |
Industrial software and service providers Ansys, Databricks, Dematic, Omron, SAP, Schneider Electric with ETAP, Siemens and more are integrating the NVIDIA Omniverse platform into their solutions to accelerate industrial digitalisation with physical AI.
New NVIDIA Omniverse blueprints connected to NVIDIA Cosmos world foundation models are also available to enable robot-ready facilities and large-scale synthetic data generation for physical AI development.
“Omniverse is an operating system that connects the world’s physical data to the realm of physical AI,” said Rev Lebaredian, VP of Omniverse and simulation technology at NVIDIA.
“With Omniverse, global industrial software, data and professional services leaders are uniting industrial ecosystems and building new applications that will advance the next generation of AI for industries at unprecedented speed.”
Mega, an Omniverse blueprint for testing multi-robot fleets at scale in industrial digital twins, is now available in preview on build.nvidia.com. Also available is the NVIDIA AI blueprint for video search and summarisation, powered by the NVIDIA Metropolis platform, for building AI agents that monitor activity across entire facilities.
In automotive manufacturing, Schaeffler and Accenture are starting to adopt Mega to test and simulate fleets of Agility Robotics Digit for material-handling automation. Hyundai Motor Group is using the blueprint to simulate Boston Dynamics Atlas robots on its assembly lines, and Mercedes-Benz is using it to simulate Apptronik’s Apollo humanoid robots to optimise vehicle assembly operations.
In electronics manufacturing, Pegatron is using Mega to develop physical AI-based NVIDIA Metropolis video analytics agents to improve factory operations and worker safety. Foxconn is using the blueprint to simulate industrial manipulators, humanoids and mobile robots in its manufacturing facilities for the NVIDIA Blackwell platform.
“Foxconn is constantly exploring ways to transform our operations as we continue our journey toward building the factories of the future,” said Brand Cheng, CEO of Fii, a core subsidiary of Foxconn.
“Using NVIDIA Omniverse and Mega, we’re testing and training humanoids to operate in our leading factories as we advance to the next wave of physical AI.”
KION Group, Dematic and Accenture are integrating Mega to advance next-generation AI-powered automation for warehouses and supply chain solutions. Idealworks is integrating Mega into its fleet management software to simulate, test and optimise robotic fleets, while SAP customers and partners can tap on Omniverse to develop their own virtual environments for warehouse management scenarios.
A new Omniverse blueprint for AI factory digital twins lets data centre engineers design and simulate AI factory layouts, plus cooling and electrical systems to maximise utilisation and efficiency. The Cadence Reality Digital Twin Platform and Schneider Electric with ETAP feature simulation software integrated into the blueprint, while Vertiv and Schneider Electric are also providing Omniverse SimReady 3D models of their power and cooling units to accelerate the development of AI factory digital twins.
The NVIDIA Isaac GR00T blueprint for synthetic manipulation motion generation is also now available for robotics developers, enabling large-scale synthetic data generation from Omniverse and the NVIDIA Cosmos world model development platform. The blueprint helps humanoid developers reduce data collection time from hours to minutes, fast-tracking robot development.
Ansys, Cadence, Hexagon, Omron, Rockwell Automation and Siemens are integrating Omniverse data interoperability and visualisation technologies into their industrial software, simulation and automation solutions to accelerate product development and optimise manufacturing processes.
For physical AI, Intrinsic, an Alphabet company, is enabling Omniverse workflows and NVIDIA robotics foundation models to transition from digital twins to hardware deployments using Flowstate. Databricks is integrating NVIDIA Omniverse with the Databricks Data Intelligence Platform, which will enable large-scale synthetic data generation for physical AI.
Unilever has also adopted Omniverse and physically-accurate digital twins to streamline and optimise marketing content creation for its products.
To simplify development, deployment and scale-out of OpenUSD-based applications, NVIDIA Omniverse is now available as virtual desktop images on EC2 G6e instances with NVIDIA L40S GPUs in AWS Marketplace. The Microsoft Azure Marketplace now features preconfigured Omniverse instances and Omniverse Kit App Streaming on NVIDIA A10 GPUs, allowing developers to easily develop and stream their custom Omniverse applications.
These cloud-based NVIDIA Omniverse developer tools and services are expected to be available later this year on Oracle Cloud infrastructure compute bare-metal instances with NVIDIA L40S GPUs, as well as the newly-announced NVIDIA RTX PRO 6000 Blackwell Server Edition on Google Cloud.
At GTC, NVIDIA introduced the OpenUSD Asset Structure Pipeline for Robotics with Disney Research and Intrinsic. This new structure and data pipeline uses today’s best practices within OpenUSD to work toward unifying robotic workflows, providing a common language for all data sources.
To help design and optimise AI factories — purpose-built infrastructure dedicated to AI training and inference, NVIDIA unveiled the NVIDIA Omniverse blueprint for AI factory design and operations. The blueprint uses
OpenUSD libraries that enable developers to aggregate 3D data from
disparate sources such as the building itself, NVIDIA accelerated
computing systems and power or cooling units from providers such as
Schneider Electric and Vertiv.
During his GTC keynote, NVIDIA founder and CEO Jensen Huang showcased how an application on the Omniverse blueprint can be used to plan, optimise and simulate a 1 gigawatt AI factory. A single gigawatt AI factory could require tens of thousands of workers
across suppliers, architects, contractors and engineers to build, ship
and assemble nearly 5 billion components and over 210,000 miles of fibre
cable. Every day of downtime at such a factory can cost over US$100
M, NVIDIA said. By solving infrastructure challenges in advance, the
blueprint reduces both risk and time to deployment.
Connected to simulation tools such as Cadence Reality Digital Twin Platform and ETAP, engineering teams can test and optimise power, cooling and networking long before construction starts. By unifying the design and simulation of billions of components, the blueprint helps engineers address complex challenges like:
- Component integration and space optimisation.
- Cooling system performance and efficiency.
- Power distribution and reliability.
- Networking topology and logic.
Using the blueprint, engineers can also bring together different teams — power, cooling and networking, avoiding inefficiencies and potential failures:
- Collaborate in full context, iterating in parallel, sharing live simulations that reveal how changes in one domain affect another.
- Optimise energy usage through real-time simulations that enable teams to find the most efficient designs for AI workloads.
- Validate redundancy configurations before deployment.
By integrating real-time simulation across disciplines, the blueprint allows engineering teams to explore various configurations to model cost of ownership and optimise power utilisation. For example, a small tweak in cooling layout might significantly improve efficiency — a detail that could have been missed on paper. And instead of waiting hours for simulation results, teams can test and refine strategies in seconds.
Once an optimal design was finalised, Omniverse streamlined communication with suppliers and construction teams — ensuring that what gets built matches the model, down to the last detail.
NVIDIA is already working on agentic AI for the next evolution of the blueprint. Vertech is collaborating with the NVIDIA data centre engineering team on NVIDIA’s advanced AI factory control system, which integrates IT and operational technology data to enhance resiliency and operational visibility.
Phaidra, on the other hand, is working with NVIDIA to integrate reinforcement-learning AI agents into Omniverse. These agents optimise thermal stability and energy efficiency through real-time scenario simulation, creating digital twins that continuously adapt to changing hardware and environmental conditions.
Built on the Blackwell architecture that was introduced a year ago, Blackwell Ultra includes the NVIDIA GB300 NVL72 rack-scale solution and the NVIDIA HGX B300 NVL16 system. The GB300 NVL72 delivers 1.5x more AI performance than the NVIDIA GB200 NVL72, as well as increases Blackwell’s revenue opportunity by 50x for AI factories, compared with those built with NVIDIA Hopper, NVIDIA said.
Also new are the NVIDIA Spectrum-X and NVIDIA Quantum-X silicon photonics networking switches, which enable AI factories to connect millions of GPUs across sites while reducing energy consumption and operational costs.
Hashtag: #GTC2025
No comments:
Post a Comment