NVIDIA Unveils AI-Powered Surgical Robotics Suite at GTC 2026

NVIDIA Brings Physical AI to the Operating Room

At its annual GTC conference in San Jose on March sixteenth, NVIDIA unveiled a sweeping set of AI tools designed to revolutionise how surgical robots are trained and deployed. The announcement signals a major push by the chip giant into healthcare robotics, with CEO Jensen Huang declaring that physical AI has arrived and that every industrial company will become a robotics company.

Digital Twins for Hospitals

The centrepiece of the announcement is Rheo, a developer blueprint within NVIDIA's Isaac for Healthcare framework. Rheo allows developers to build physically accurate digital twins of hospital environments, modelling everything from medical device interactions to human movements and logistics. The goal is to let robotic systems train extensively in simulation before ever entering a real clinical setting.

New AI Models for Surgery

NVIDIA also released GR00T-H, a vision language action model trained on roughly six hundred hours of surgical data from the Open-H-Embodiment dataset, described as the world's largest healthcare robotics dataset with seven hundred and seventy-six hours of surgical video from thirty-five collaborators. The model translates text commands for clinical tasks into robotic motion commands and has already demonstrated the ability to execute an end-to-end suture in benchmark testing.

Companion model Cosmos-H generates physically plausible synthetic surgical video from kinematic actions, dramatically cutting simulation time from two days of real-world benchtop work to approximately forty minutes.

Industry Heavyweights Sign On

Several major medical robotics firms are already adopting the platform. CMR Surgical is using Cosmos-H to train its Versius surgical system and contributing nearly five hundred hours of surgical video to the project. Johnson and Johnson MedTech is applying Isaac Sim workflows to its Monarch Platform for Urology. Medtronic is exploring NVIDIA's IGX Thor for functional safety in surgical systems, while Moon Surgical and PeritasAI are also among early adopters.

Addressing a Critical Workforce Gap

The push arrives against a backdrop of acute labour pressure, with the World Health Organisation estimating a global shortfall of eleven million health workers by twenty thirty. Huang framed this as healthcare's inflection point where AI-driven automation becomes not just viable but necessary. All models are now available as open source on GitHub and Hugging Face for developers to adapt to specific surgical scenarios.