General Motors cuts FEA simulation time from 15 hours to one minute using AI/ML virtualization tools

Sterling Anderson, Chief Product Officer at General Motors, described GM's shift into a 'third epoch of engineering' where AI and machine learning collapse traditionally separate engineering functions—aerodynamics, structures, design—into a single, probabilistically informed development method.

GM's finite element analysis (FEA) simulations, which historically took 15 hours per run, now complete in one minute using AI/ML-powered virtualization. Engineers can now run thousands of design variations and iterate rapidly instead of setting up overnight simulations.

The virtualization approach extends beyond early analysis to vehicle handling tests, crash performance, and HVAC system design. For example, GM can now simultaneously optimize hardware and software in a virtual environment to simulate Consumer Reports' avoidance maneuvers, and crash analysis that took 15 to 18 hours now takes less than one minute.