Foxconn and YUAN put physical AI on hospital rounds.

Foxconn (TWSE: 2317) unveiled its latest physical AI healthcare vision at COMPUTEX 2026 and NVIDIA GTC Taipei 2026 on June 1, showing how AI agents, collaborative robotics, digital twins, and multimodal medical AI are converging on real hospital operations. At a featured session at the Taipei International Convention Center, Foxconn B Group President Barry Chiang and Digital Health User Experience Design Director Alice Lin presented the path from Nurabot deployment to smart-hospital workflow transformation — built on NVIDIA's full-stack architecture spanning energy, infrastructure, platform, models, and applications, with technologies including NVIDIA Omniverse, Isaac, Holoscan, MONAI, Nemotron, and NemoClaw.

Nurabot, now in the wards

Nurabot, Foxconn's nursing collaborative robot, completed field validation last year and is expanding into hospitals and long-term care settings, including Taipei Veterans General Hospital, Tungs' Taichung MetroHarbor Hospital, and overseas nursing education institutions. Real-world deployment data shows it executing 75–80 tasks a day — medication delivery, specimen transport, and other clinical assistance — and reducing nursing workload by roughly 30 per cent.

A scrub nurse robot, co-developed with YUAN

Foxconn also showcased its scrub nurse collaborative robot, jointly developed with Kawasaki Heavy Industries, Taichung Veterans General Hospital, and YUAN High-Tech. The system builds on the NVIDIA Isaac for Healthcare Agent-Ready Rheo Blueprint and the Isaac GR00T Vision-Language-Action architecture, giving it multimodal perception, surgical scene understanding, and intelligent task reasoning. Since its GTC debut earlier this year the project has advanced through simulation training and surgical workflow validation, with clinical validation planned next — human-robot collaboration moving towards the operating room itself.

A closed loop for chemotherapy

A third showcase, developed with Taipei Veterans General Hospital, Kawasaki Heavy Industries, Yuyama Manufacturing, and FARobot, integrates chemotherapy compounding and delivery into one intelligent workflow. Powered by NVIDIA Omniverse, the platform pairs a virtual hospital with the physical one: ChemoRo performs high-precision drug compounding, the FARobot SMR30 transports medications autonomously through the hospital, and Nurabot handles last-mile delivery to nursing stations and patients — a closed loop from compounding to administration that improves efficiency, traceability, and medication safety in a high-risk environment.

CoDoClaw: the orchestration layer

Foxconn also introduced CoDoClaw, a clinical intelligent agent system built on NVIDIA NemoClaw, an open-source blueprint for deploying autonomous agents. CoDoClaw advances CoDoctor AI from standalone tools into a multi-agent orchestration platform, coordinating agents across breast cancer screening, ECG analysis, fundus imaging, and coronary artery analysis through a unified clinical interface — from lesion detection through scheduling, report generation, and follow-up. Foxconn envisions it as a foundational layer for the AI-native hospital.

"Healthcare's next transformation will not come from AI models alone, but from AI systems capable of perceiving, reasoning and acting in real-world clinical environments alongside healthcare professionals," said Chiang. Taiwan — with its world-class healthcare and insurance systems, semiconductor leadership, and highly digitised clinical environments — is positioned as the proving ground. Through the Taiwan Digital Health Alliance (HiMEDt), Foxconn will continue working with NVIDIA and ecosystem partners to carry physical AI into hospitals, long-term care, and home healthcare.