Explore the Best of GTC Taipei

The next generation of AI agents doesn't just respond—it plans, reasons across modalities, and operates continuously over hours and days to complete real work. Explore how NVIDIA is building the model infrastructure, open source tools, and AI factory foundation to make long-running intelligent agents real and production ready.

AI has reached an inflection point. Through platforms like OpenClaw, always-on "claw" agents can now continuously learn, adapt, and generate value in real time—redefining how work gets done. For enterprises, this means faster drug discovery, 24/7 customer operations, and compounding productivity and innovation through decentralized agent creation.

As AI factories grow to hundreds of thousands—and soon millions—of interconnected GPUs, the network has become the defining architecture of large-scale AI. Achieving uncompromised performance requires fabrics that not only scale up and out within a single data center, but also scale across multiple sites with predictable latency, lossless throughput, and end-to-end resilience.

Healthcare is undergoing a platform shift, where AI moves beyond training models to deploying intelligent systems that can perceive, reason, and act in the physical world. As agentic AI goes mainstream and physical AI expands across robotics, science, and healthcare, intelligent machines and digital agents are transforming how medical devices, hospitals, and life sciences are built and operated.

The integration of large language models into healthcare demands uncompromising accuracy and resource efficiency. This session explores the architectural journey of building EBM-Guardian, detailing how Mackay Memorial Hospital overcomes the engineering challenges of deploying massive 100B+ parameter models using advanced quantization and dynamic task loading to deliver verifiable, zero-hallucination clinical consultations.

As healthcare systems worldwide confront aging populations, workforce shortages, and rising clinical complexity, the next frontier of transformation is embodied intelligence—AI models integrated with physical robotic systems to enable real-time, action-oriented clinical operations. 

One three-computer architecture—NVIDIA DGX, OVX, and AGX—one global AI factory. Told through four real scenarios. (1) A heatsink scratch missed by humans. (2) Thermal designs still demand weeks of prototyping. (3) Line reconfiguration once meant costly trial and error. 4) Shift reports and decisions previously took hours.

Discover how physical AI is redefining how factories are built and operated at global scale. By integrating the NVIDIA Omniverse platform with simulation, perception, and AI agent technologies—including VSS, Cosmos, Isaac Sim, AR, and NeMo-based agents—we enable factories to evolve from automated systems into intelligent, self-aware operations. Through this approach, new factories can launch faster, ramp more efficiently, and operate with greater resilience and adaptability.

See physical AI at work on a real factory floor with PEGAVERSE and PEGA AI—built with NVIDIA Omniverse, Physical AI Data Factory, and NVIDIA Metropolis. Explore how sim-ready digital assets, synthetic data generation skills, and AI agents drive control and real-world sensor feedback in a single closed loop. Discover how Pegatron used synthetic defect data for PCBA inspection to cut loss rates by 50% on day one, saving months of effort. Walk away with a practical blueprint for closing the sim-to-real gap in manufacturing operations—from individual robots and agents to entire AI factory infrastructure.