NVIDIA announced on Wednesday, October 14, 2026, the NVIDIA cuPotential library and that leading semiconductor companies are adopting CUDA-X libraries — including the new NVIDIA cuPotential — to accelerate the development of AI-defined fabs.
What is NVIDIA cuPotential?
NVIDIA cuPotential is a CUDA-X library that allows developers to build AI agents capable of simulating interactions between atoms and molecules. In early-access testing, NVIDIA cuPotential accelerated already GPU-optimized workflows by an additional 2 to 3 times.
In addition, NVIDIA highlighted NanoVDB, an open-source library for 3D data that helped Lam Research achieve simulations up to 30 times faster in wafer etch modeling, and NPP for GPU-accelerated inspection, adopted by KLA.
Why it matters
AI-defined fabs are semiconductor factories where AI agents and models optimize processes, solve production issues, and improve yield before changes reach the line. With demand for advanced chips surging, tools like NVIDIA cuPotential reduce simulation and inspection times at atomic and nanometer scales.
Tim Costa, NVIDIA vice president and general manager of industrial and computational engineering, said: “The semiconductor industry made the AI revolution possible. Now, AI is transforming how chips are built.”
Context
The announcement of NVIDIA cuPotential took place during SEMICON West and continues NVIDIA’s work with TSMC and other companies to bring accelerated computing and AI inside factories. Companies such as Synopsys, Cadence, Lam Research, KLA, and others are among those adopting the libraries.
NVIDIA cuPotential is in early access. The goal is to make fabs more autonomous and capable of handling the growing complexity of manufacturing processes for AI chips.
Sources
Transparency: This content was created, edited, or reviewed with the assistance of artificial intelligence. Information was cross-referenced with public posts on X and sources available on the internet. Consult the original sources to verify the full context.
By GeekikiBot
