Applied Materials and NVIDIA Partner on AI-Accelerated Chip Design and Manufacturing Simulations

Applied Materials and NVIDIA are integrating CUDA-X libraries and AI into semiconductor simulation tools, claiming speedups of up to 55x for atomic-scale modeling and 35x for manufacturing process simulations.

This article was drafted with AI assistance from multiple sources and was reviewed and approved by a human editor before publication.

Applied Materials and NVIDIA have announced a broad collaboration to integrate NVIDIA's CUDA-X libraries and artificial intelligence into Applied Materials' semiconductor simulation and manufacturing tools. The companies say the effort creates an end-to-end digital development model for chip innovation.

As part of the partnership, Applied Materials' Ginestra platform, which simulates material properties and defects at the atomic scale, now incorporates NVIDIA's cuDSS direct sparse solver. According to the companies, this integration delivers up to a tenfold speedup over CPU-only approaches while maintaining accuracy.

NVIDIA's cuEST library is being used to accelerate density functional theory (DFT) workflows. On NVIDIA B200 systems, a DFT simulation that took five days on 64 CPU cores can now be completed in about two hours on a single GPU, representing a roughly 55-fold improvement, the companies state.

The collaboration also extends to Applied Materials' ACE+ platform, which simulates chamber design, hardware configuration, and process parameters using multiphysics models. NVIDIA's PhysicsNeMo converts ACE+ simulation results into real-time digital twins of process behavior, according to the companies.

At the factory scale, Applied Materials is leveraging NVIDIA Omniverse for optimization, though details were not provided. The companies also highlight the Endura metallization system as a platform for validated process innovations.