NVIDIA (NASDAQ:NVDA) revealed that Taiwan Semiconductor Manufacturing Co. (NYSE:TSM) is deploying a range of its artificial intelligence and accelerated computing technologies throughout semiconductor development and manufacturing processes, deepening the partnership between the two companies.
The initiative spans multiple areas of chip production, from lithography and materials research to factory optimization and defect detection, as semiconductor manufacturers increasingly adopt AI-driven tools to improve efficiency and performance.
AI-Powered Lithography Delivers Efficiency Gains
One of the key technologies being implemented is NVIDIA’s cuLitho platform, which TSMC is using for computational lithography applications.
According to the companies, the solution has generated improvements of between 20% and 50% in either cost efficiency or processing cycle times compared with traditional CPU-based approaches.
The technology is designed to accelerate one of the most computationally intensive stages of semiconductor manufacturing, helping optimize chip patterning and production workflows.
Faster Materials Research Through Accelerated Simulation
TSMC is also leveraging NVIDIA’s cuEST software for electronic structure simulation, enabling significantly faster analysis of semiconductor materials.
The companies stated that the platform can deliver chemistry simulations up to 50 times faster than conventional methods, supporting the design and development of advanced semiconductor materials.
By shortening simulation times, engineers can evaluate a broader range of material candidates and accelerate research and development cycles.
Machine Learning Enhances Process Control
For manufacturing process optimization, TSMC has incorporated NVIDIA’s cuML machine learning library into its advanced process control systems.
The platform enables the analysis of hundreds of thousands of manufacturing parameters across thousands of production stages, allowing engineers to identify inefficiencies and reduce process variation more effectively.
According to TSMC, the technology has contributed to meaningful improvements in process consistency and operational performance.
GPU Computing Improves Fab Productivity
The semiconductor manufacturer is also deploying NVIDIA H200 GPUs to support production scheduling and factory management.
By using GPU-accelerated computing for scheduling calculations, TSMC has been able to better manage complex manufacturing constraints and optimize production flows within its fabrication facilities.
The companies said these enhancements have resulted in measurable productivity improvements across fab operations.
AI Vision Systems Strengthen Defect Detection
Another area of collaboration focuses on quality control and inspection.
TSMC is utilizing NVIDIA’s Metropolis platform alongside the NVIDIA TAO Toolkit to develop advanced vision AI systems capable of identifying semiconductor defects at nanometer scale.
The technology improves defect classification accuracy while reducing the amount of manual labeling and model retraining required, helping streamline inspection processes and improve manufacturing yields.
Digital Twin Technology Supports Virtual Factory Design
TSMC is additionally evaluating NVIDIA Omniverse libraries as part of its FabTwin initiative, a virtual manufacturing environment designed to simulate and optimize fabrication facilities.
The digital platform allows engineers to test equipment layouts, production scenarios and workflow configurations in a virtual setting before implementing changes in physical facilities.
This approach can help reduce deployment risks, improve planning efficiency and accelerate factory optimization efforts.
NVIDIA Highlights Growing Role of AI in Manufacturing
Commenting on the partnership, NVIDIA founder and Chief Executive Officer Jensen Huang emphasized the increasing role of artificial intelligence within advanced semiconductor production.
“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 during NVIDIA’s GTC Taipei event, where the company showcased a range of technologies aimed at expanding the use of AI across industrial, enterprise and manufacturing applications.