One of the most striking demonstrations in the field of healthcare robotics at Computex 2026 was the artificial intelligence-supported robot nurse developed to support the surgeon in the operating room. The system exhibited by Foxconn under the name Scrub Bot works with a physical artificial intelligence infrastructure based on NVIDIA Isaac for Healthcare and Isaac GR00T. Scrub Bot, the NVIDIA-supported robot nurse exhibited at Computex 2026, is not a robot that performs surgery in the classical sense.
The system was developed to support the specific duties of the operating room nurse, who acts as a scrub nurse in the operating room team and prepares surgical instruments and hands them to the surgeon during the operation. The robot can detect the surgeon’s voice commands and react according to the process in the operating room. According to the information provided by NVIDIA, Scrub Bot is designed to work in live surgical environments and can adapt to changing needs during the operation in real time.
Behind the system, Foxconn uses NVIDIA’s physical artificial intelligence infrastructure. The Agent-Ready Rheo blueprint included in NVIDIA Isaac for Healthcare simplifies the development of hospital automation workflows. Foxconn uses this infrastructure in scene reconstruction, policy training, evaluation and transition to the real environment. The basis of Scrub Bot is not just camera or command detection. The system uses multi-modal artificial intelligence approaches to understand the surgical environment, plan the correct movement, and work safely in the same space as humans.
This structure allows the robot to move not only with fixed commands but also according to the changing flow in the operating room. NVIDIA’s work on healthcare robotics is not limited to Scrub Bot. CMR Surgical joined NVIDIA’s Physical AI healthcare robotics initiative, contributing approximately 500 hours of anonymized surgical data from its Versius surgical robotics system to the Open-H dataset. Open-H is the basis for the Isaac GR00T-H vision-language-action model developed for healthcare robotics.
The Open-H dataset combines real surgical video, robot telemetry and multi-modal data. GR00T-H is positioned as an open model developed for robots to better interpret complex surgical environments and tasks. These studies are used to first train and evaluate surgical robots in a simulation environment and then move them to the real clinical environment in a controlled manner. The Scrub Bot demonstration at Computerex 2026 showed that the new era in healthcare robotics is not just about robotic arms performing surgery.
Robot assistants that hand tools to the surgeon in the operating room, follow the task flow, respond to voice commands, and work in the same environment with clinical teams are also becoming a part of this field. Foxconn’s work on health robotics is included in a broader artificial intelligence and robotics ecosystem carried out together with hospitals in Taiwan. NVIDIA also offers its accelerated computing, simulation, edge AI and robotics infrastructure to healthcare institutions and technology manufacturers in this field.
Comments
You can write your views about this story. Comments may be moderated according to site settings.