At NVIDIA GTC Taipei, NVIDIA and TSMC announced an expanded partnership to integrate artificial intelligence and accelerated computing directly into semiconductor fabrication. This collaboration aims to address the growing complexity of moving advanced chips from design to high-volume production. As semiconductor nodes shrink, tasks such as computational lithography, transistor simulation, process control, and wafer inspection now demand massive-scale simulation and real-time optimization that traditional methods cannot efficiently handle. TSMC is leveraging NVIDIA's technology across its entire design and manufacturing lifecycle to enhance turnaround time, energy efficiency, yield, and operational productivity. Jensen Huang, NVIDIA's CEO, emphasized that bringing NVIDIA AI and accelerated computing into the fab itself allows the companies to tackle some of the world's most difficult challenges using simulation and optimization. C.C. Wei, TSMC's Chairman and CEO, noted that this integration strengthens technology leadership and manufacturing excellence, ensuring the production of future products for customers. To accelerate these complex workloads, TSMC utilizes NVIDIA CUDA-X libraries and AI models. A key focus is defect inspection, where even nanometer-scale flaws can compromise chip quality. By employing the NVIDIA Metropolis platform and NVIDIA TAO Toolkit, TSMC has implemented vision AI to improve defect classification. This approach enables more accurate and faster detection while reducing the need for repeated labeling and retraining as process conditions or tool types change. Furthermore, TSMC is exploring NVIDIA Omniverse to develop FabTwin, a virtual environment designed to simulate the entire fab. This digital twin allows engineers to evaluate process tool layouts and related workflows before any physical implementation. By testing complex design scenarios in a virtual setting, TSMC can identify potential constraints early, compare configurations flexibly, and improve planning efficiency. This virtual-first strategy accelerates critical decision-making and avoids premature capital commitments. The partnership builds on nearly three decades of collaboration between the two firms. By applying AI to physics, image recognition, and optimization across the manufacturing floor, they are pushing the boundaries of what is possible in semiconductor design. The initiative is expected to significantly boost the speed and reliability of producing next-generation chips, which are essential for the future of computing. Details of the advancements and NVIDIA CEO's keynote presentation can be accessed through the NVIDIA GTC Taipei event materials.