Foxconn & Nvidia Bring AI Agents to Taiwan’s Top Hospitals

Nvidia and Foxconn announced at GTC Taipei today that they are deploying coordinated AI agents and clinical robots across Taiwan's leading medical centers, moving the government-led Healthy Taiwan initiative from pilot projects into operating clinical infrastructure. The program has committed US$1.5 billion in regional investment and reaches a hospital network handling more than 14 million patient encounters a year.

Taiwan has one of the world's fastest-aging populations and a deepening shortage of clinical staff. The Healthy Taiwan program treats AI as the labor multiplier that closes the gap. Active deployments span Chang Gung Memorial, National Taiwan University Hospital, Taipei Veterans General, Taichung Veterans General, MacKay Memorial, and Kaohsiung Medical University Chung-Ho Memorial. Foxconn serves as ecosystem integrator, tying together government programs, hospital operators, device makers, and software vendors.

Foxconn's CoDoctor AI platform coordinates a roster of specialized agents, software systems trained to reason within a single clinical domain. New additions include an ECG AI Agent that triages cardiac patients from screening data, a Corovia agent that reconstructs the heart and coronary arteries in 3D in roughly one minute against a two-hour clinical workflow, and an Endovia agent supporting real-time lesion detection during colonoscopies. The platform runs on Nvidia Nemotron, an open-weights foundation model line, with a new CoDoClaw orchestration layer built on Nvidia NemoClaw.

Foxconn's Nurabot nursing collaborative robot completed field validation at Taichung Veterans General Hospital last year and is now rolling out to Taipei Veterans General and Tungs' Taichung MetroHarbor Hospital. By handling transport and logistics, Nurabot frees roughly two to three hours per nurse per day for direct patient care. A new Scrub Bot, an AI-enhanced scrub nurse robot, operates in live surgical suites and responds to the lead surgeon's voice commands. Foxconn builds Nvidia Omniverse digital twins of each hospital so robots are tested and trained virtually before deployment. The simulation-first approach has cut deployment time by 40 percent and produced 98 percent navigation accuracy.

Taiwan has 85 FDA- or TFDA-cleared medical AI solutions, and every major medical center is engaged in AI research or clinical trials. The Healthy Taiwan model is a sovereign, regulated AI-native health system running on Nvidia infrastructure and integrated by a Taiwan electronics firm. It is no longer a demonstration. It is the operating stack a half-dozen of the island's largest hospitals depend on, and a reference architecture for any country facing the same demographic math.