NVIDIA Extends Physical AI to Robotics

At the recent GTC conference, NVIDIA announced partnerships with leading robotics companies to integrate its physical artificial intelligence (AI) technologies with robotics systems. The company also announced new NVIDIA Isaac simulation frameworks and NVIDIA Cosmos and NVIDIA Isaac GR00T open models. intelligent robotics training applications.

Robotics companies involved in the initiative leveraging the NVIDIA platform include ABB Robotics, AGIBOT, Agility, FANUC, Figure, Hexagon Robotics, KUKA, Skild AI, Universal Robots, World Labs and YASKAWA.

“Physical AI has arrived — every industrial company will become a robotics company,” said Jensen Huang, founder and CEO of NVIDIA. “NVIDIA’s full-stack platform — spanning computing, open models and software frameworks — is the foundation for the robotics industry, uniting a worldwide ecosystem to build the intelligent machines that will power the next generation of factories, logistics, transportation and infrastructure.”

FANUC, ABB Robotics, YASKAWA and KUKA are integrating NVIDIA Omniverse libraries and NVIDIA Isaac simulation frameworks into their virtual commissioning solutions to develop and validate complex robot applications and entire production lines through physically accurate digital twins. To power advanced intelligence on the production line, the companies are integrating NVIDIA Jetson modules into their controllers for real-time AI inference at the edge.

In addition, developers such as FieldAI and Skild AI are building generalized robot brains using NVIDIA Cosmos world models for data generation and Isaac simulation frameworks to validate policies in simulation, enabling any robot to master new tasks with minimal retraining. World Labs is using Isaac Sim to validate its generative world models, while Generalist AI is using Cosmos to explore generating synthetic data.

At the GTC event, NVIDIA announced Cosmos 3, what the company describes as the first world foundation model unifying synthetic world generation, vision reasoning and action simulation to accelerate the development of generalized robot intelligence for complex environments.

Companies including 1X, AGIBOT, Agility, Agile Robots, Boston Dynamics, Figure, Hexagon Robotics, Humanoid, Mentee and NEURA Robotics are building the next generation of humanoids using Cosmos world models, Isaac Sim and Isaac Lab to accelerate the development and validation of their robots.

During GTC, NVIDIA introduced Isaac Lab 3.0 in early access, enabling faster, large-scale robot learning on NVIDIA DGX-class infrastructure. Built on the new Newton physics engine 1.0 and the NVIDIA PhysX software development kit, it adds multiphysics simulation and improved support for complex, dexterous manipulation.

AGIBOT, Humanoid, LG Electronics, NEURA Robotics and Noble Machines are also adopting NVIDIA Isaac GR00T N models to accelerate industrial deployments of their humanoids. NVIDIA announced GR00T N1.7 is now available in early access with commercial licensing, bringing generalized robot skills including advanced dexterous control to production-ready robot deployments.

In addition, during his GTC keynote, Huang previewed GR00T N2, a next-generation robot foundation model based on DreamZero research. Built on a new world action model architecture, the model helps robots succeed at new tasks in new environments more than twice as often as leading vision language action models, the company says. 

These systems are powered by the NVIDIA Jetson Thor robotic computing platform, enabling developers to move from simulation training to real-world deployment with greater speed, intelligence and reliability.

Physical AI for Healthcare Robotics, New Collaborations

CMR Surgical is using Cosmos-H simulation to train and validate robotic intelligence for its Versius surgical system prior to clinical deployment. Johnson & Johnson MedTech is using Isaac Sim- and Cosmos-based post-training workflows to train and validate systems for the Monarch Platform for Urology. Medtronic is exploring NVIDIA IGX Thor to deliver mission-critical precision and functional safety in surgical robotic systems.

In addition, Skild AI is partnering with ABB Robotics and Universal Robots to deploy its generalized robot intelligence across different industries and tasks. By embedding a shared intelligence layer into widely deployed industrial and collaborative systems, manufacturers can extend automation into more dynamic and variable applications without building task-specific code for every workflow. Simultaneously, Skild AI is partnering with Foxconn on high-precision assembly for NVIDIA Blackwell production lines, enabling Foxconn’s AI-driven dual-arm manipulators to master the industry’s most complex manufacturing tasks.

Lightwheel is codeveloping and calibrating the Newton physics engine to enable Samsung’s assembly robots to master intricate cable handling in simulation, delivering higher precision and faster assembly lines. PTC is announcing a new robotics design-to-simulation workflow from its cloud-native Onshape CAD and product data management platform to NVIDIA Isaac Sim, creating a seamless CAD-to-OpenUSD bridge that will enable engineering teams like FANUC America Corporation and Fauna Robotics to design and validate their robotic systems within physically accurate digital twins.

WORKR is integrating its AI platform with ABB Robotics industrial robots, using NVIDIA Omniverse libraries as part of its WorkrCore to train a robotic workforce that can be deployed by small- and medium-sized manufacturers in minutes, without programming knowledge.

KION Group, the supply chain solutions company, is working with NVIDIA and Accenture to advance autonomous warehouse solutions. Using Omniverse and a physical AI-powered digital twin and systems architecture pioneered by Accenture, KION engineers can create large-scale, physics-accurate warehouse digital twins to train and test fleets of NVIDIA Jetson-based autonomous forklifts for GXO, the world’s largest pure-play contract logistics provider.

Microsoft Azure and Nebius are integrating the NVIDIA Physical AI Data Factory blueprint to enable scalable, agent-driven synthetic data generation for their developers, including FieldAI, Teradyne Robotics, Hexagon Robotics on Azure, and RoboForce on Nebius. CoreWeave is integrating NVIDIA Isaac Lab to build robot learning pipelines, while Alibaba Cloud is integrating NVIDIA’s entire physical AI stack into its Platform for AI to accelerate end-to-end robotics development.

Kamino, a GPU-accelerated physics simulator built by Disney on the NVIDIA Warp framework and integrated into Newton, enables the training of robot policies for Disney’s Olaf and BDX Droids — allowing Olaf to learn to manage its own heat and reduce impact noise, and enabling BDX Droids to navigate complex environments. In his GTC keynote, Huang was joined by Disney’s Olaf ahead of his debut at Disneyland Paris on March 29.

Sources: Press materials received from the company and additional information gleaned from the company’s website.