News |
By Asma Adhimi
NVIDIA and TSMC are expanding their long-running partnership by bringing AI and accelerated computing directly into semiconductor manufacturing, targeting faster chip production, improved yields and more efficient fab operations.
Announced at NVIDIA GTC Taipei, the collaboration covers a broad range of semiconductor workflows, from computational lithography and transistor simulation to defect inspection and factory scheduling. The move highlights how AI is becoming a core technology inside advanced fabs as chip manufacturing grows increasingly complex at smaller process nodes.
For eeNews Europe readers, the development is another sign that AI infrastructure is no longer limited to data centres and cloud applications. Semiconductor manufacturers are now using AI to optimise physical production processes, potentially reducing costs and speeding up the rollout of advanced chips across automotive, industrial and edge AI markets.
TSMC is deploying several NVIDIA CUDA-X libraries and GPU-accelerated tools to improve simulation and process optimisation inside its fabs.
One of the key technologies is NVIDIA cuLitho, a GPU-accelerated computational lithography library used for chip mask design. According to NVIDIA, the technology provides a 20–50% improvement in cost effectiveness or cycle time compared with CPU-based lithography approaches while maintaining the same cost of ownership.
TSMC is also using NVIDIA cuEST for electronic structure simulations, enabling what NVIDIA says are chemistry simulations that run 50 times faster on average for semiconductor material design.
In process control, TSMC is applying the NVIDIA cuML machine learning library to analyse hundreds of thousands of manufacturing parameters across thousands of process steps. The goal is to reduce process variation and improve production consistency.
CUDA-powered computation on NVIDIA H200 GPUs also accelerates fab scheduling and productivity, helping TSMC manage increasingly complicated production constraints in advanced manufacturing facilities.
“NVIDIA and TSMC have worked together for nearly three decades 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 itself, tackling some of the world’s most complex design and manufacturing challenges with simulation, optimization and AI to improve speed, efficiency and yield for the next generation of chips.”
TSMC is additionally using the NVIDIA Metropolis platform and NVIDIA TAO Toolkit to improve automated defect inspection.
The system uses vision AI to identify nanometre-scale defects during manufacturing, helping improve inspection accuracy while reducing the amount of repeated data labelling and retraining normally required as production conditions evolve.
“TSMC and NVIDIA have built a long-standing partnership rooted in advancing the technologies that make the next generation of computing possible,” said C.C. Wei, chairman and CEO of TSMC. “By using NVIDIA accelerated computing and AI across fab operations optimization, lithography, process control and inspection, TSMC is strengthening our technology leadership and manufacturing excellence to support our customers’ future products and success.”
TSMC is also exploring NVIDIA Omniverse technologies to build what it calls FabTwin, a virtual environment designed to simulate fab layouts and manufacturing workflows before physical deployment.
The idea is to digitally test different factory configurations, identify bottlenecks early and improve planning efficiency before committing to expensive physical infrastructure.
As semiconductor fabs become larger and more automated, digital twins and AI-based optimisation are expected to play a growing role in reducing operational complexity and improving manufacturing flexibility.
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。