IBM and Dallara build physics-based AI models for race car aerodynamics

IBM and Dallara Group are building physics-based AI models to speed up aerodynamic design of high-performance vehicles and are exploring quantum computing for later use. In an early test on a conceptual LMP2-style car, the AI model reached the same rear-diffuser optimum as conventional simulation in about 10 seconds, versus several hours. The companies estimate it could cut days of simulation to minutes across large design sets.

• AI model matched the conventional simulation's optimal rear-diffuser design
• Evaluation took about 10 seconds versus several hours
• Companies estimate days of simulation could fall to minutes