NVIDIA: AI turns chip fabs into “AI-defined factories”

NVIDIA says AI is pushing chip fabs into an "AI-defined factory" era. With accelerated simulation, learned physics models, AI agents and digital twins, chip design, testing and production decisions are getting much faster. As demand for advanced chips grows, processes get more complex, and real-line tests are slow and costly—engineers often need days to find defect causes.

For example, COMSOL sped up simulations up to 10x with NVIDIA's cuDSS. Siemens boosted key simulations up to 17x by adding cuLitho to its Calibre software. Applied Materials' platform saw up to 35x speedup with PhysicsNeMo, and plasma behavior predictions dropped from days to milliseconds.

NVIDIA says tasks that once took days can now be done in minutes, letting engineers focus on judgment and optimization. Applied Materials is also building a full digital twin of a fab on Omniverse to simulate material flow, spot bottlenecks and test strategies—final decisions still stay with engineers.