IBM (NYSE: IBM) and Dallara Group announced a collaboration to develop AI foundation models for high-performance vehicle design and explore quantum computing applications in automotive engineering.

The partnership focuses on creating physics-based AI models trained on Dallara's aerodynamic simulation data. Early testing on a conceptual Le Mans Prototype 2-like race car showed the AI model completed aerodynamic evaluations in approximately 10 seconds compared to several hours required by traditional computational fluid dynamics methods.

The AI approach identified optimal rear diffuser configurations with similar accuracy to conventional CFD analysis. When applied to hundreds of geometry configurations, this technology could reduce simulation time from days to minutes, according to the companies.

Dallara, which designs vehicles for racing series including IndyCar, Formula 2, and Formula 3, provided proprietary aerodynamic data for training the AI models. The company has supplied high-performance vehicles for motorsports for over 50 years and validates simulation results against real-world vehicle performance.

The companies are also investigating quantum computing integration in design workflows to enhance simulation capabilities for complex aerodynamic problems. IBM and Dallara plan to expand the AI models to cover different driving conditions and maneuvers.

"Some of the hardest engineering challenges come down to accurately simulating the physical world," said Alessandro Curioni, IBM Fellow and VP of Algorithms and Applications at IBM Research.

Andrea Pontremoli, Dallara CEO, stated the collaboration represents the company's commitment to advancing engineering boundaries through continuous learning.

The collaboration builds on IBM's Gauge-Invariant Spectral Transformers model. Initial results were detailed in a preprint study published at arXiv on April 20, 2026, with research presented at the International Conference on Learning Representations in Rio de Janeiro on April 26, 2026.