NVIDIA and TSMC Advance Semiconductor Design and Manufacturing with AI in the Fab

TAIPEI, Taiwan — COMPUTEX — June 1, 2026 — NVIDIA today announced that TSMC, the world’s leading semiconductor company, is leveraging NVIDIA accelerated computing and AI to advance semiconductor design and manufacturing.

As chips transition to more advanced nodes, the process from design to mass production has become one of the world's most complex computing challenges. Computational lithography, transistor simulation, process control, and wafer inspection now require AI systems capable of providing support across large-scale simulation, real-time optimization, physics, image processing, and other applications.

TSMC is accelerating this transformation by applying accelerated computing and AI throughout the lifecycle of semiconductor design and manufacturing to improve lead times, energy efficiency, yield, and operational productivity in advanced fabs.

“NVIDIA and TSMC have worked together for nearly 30 years to push the limits of computing,” said Jensen Huang, founder and CEO of NVIDIA. “TSMC is bringing NVIDIA AI and accelerated computing into the fab to tackle the world’s most complex design and manufacturing challenges with simulation, optimization, and AI to improve the speed, efficiency, and yield of next-generation chips.”

“TSMC and NVIDIA have built a long-term partnership rooted in advancing technologies that enable next-generation computing,” said C.C. Wei, chairman and CEO of TSMC. “By leveraging NVIDIA accelerated computing and AI across fab operations, lithography, process control, and inspection, TSMC is reinforcing its technological leadership and manufacturing excellence to support our customers’ future products and success.”

TSMC Accelerates Processes with NVIDIA CUDA-X Libraries and AI

TSMC is utilizing NVIDIA CUDA-X™ libraries and AI models to accelerate these workloads on NVIDIA GPUs.

- Computational Lithography: TSMC is leveraging NVIDIA cuLitho, a GPU-accelerated library for lithography, which is the exposure technology for chip mask design. This technology improves cost efficiency or cycle time by 20% to 50% compared to CPU-based computational lithography, while maintaining the same total cost of ownership.
- Transistor, Equipment, and Process Simulation: TSMC is utilizing NVIDIA cuEST, a GPU-accelerated electronic structure simulation library, to accelerate chemical simulations in semiconductor material design by an average of 50x.
- Advanced Process Control: TSMC is leveraging NVIDIA cuML machine learning libraries to accelerate large-scale analytics on NVIDIA GPUs. This allows TSMC to speed up algorithms and extract hundreds of thousands of process parameters across thousands of steps as precise inputs for machine learning models, significantly reducing process variation.
- Fab Operations Optimization: Scheduling computation powered by GPU acceleration using CUDA has significantly boosted fab productivity with NVIDIA H200 GPUs. By leveraging CUDA-powered computation on NVIDIA H200 GPUs, TSMC has strengthened its ability to manage complex constraints, thereby optimizing production routes.