NVIDIA (NASDAQ: NVDA) announced that Taiwan Semiconductor Manufacturing Co. is implementing NVIDIA's AI and accelerated computing technologies across its semiconductor design and manufacturing operations.

The collaboration involves TSMC using NVIDIA CUDA-X libraries and AI models to enhance various manufacturing processes. TSMC is applying NVIDIA cuLitho for computational lithography, achieving 20-50% improvement in cost effectiveness or cycle time compared with CPU-based systems. The company is also using NVIDIA cuEST for electronic structure simulation, providing 50x faster chemistry simulations for semiconductor material design.

For advanced process control, TSMC is implementing the NVIDIA cuML machine learning library to accelerate large-scale analytics, enabling the processing of hundreds of thousands of process parameters across thousands of steps. The company reports significant reduction in process variation through this application.

TSMC is using NVIDIA H200 GPUs for fab operations optimization, with GPU-accelerated scheduling computation leading to improvements in fab productivity by managing complex constraints and streamlining production paths.

The companies are also collaborating on defect inspection technology. TSMC is using the NVIDIA Metropolis platform and NVIDIA TAO Toolkit to improve advanced defect classification through vision AI, enhancing detection of nanometer-scale defects while reducing the need for repeated labeling and retraining.

TSMC is exploring NVIDIA Omniverse libraries to develop FabTwin, a virtual fab environment for evaluating process tool layouts and simulation workflows. This system allows digital testing of design scenarios before physical implementation.

"TSMC is bringing NVIDIA AI and accelerated computing into the fab itself, tackling some of the world's most complex design and manufacturing challenges," said Jensen Huang, NVIDIA's founder and CEO.

The announcement was made at NVIDIA GTC Taipei, according to the press release statement.