McLaren Integrates Physics AI Surrogate Models and AI Agents to Accelerate Design Exploration

McLaren Automotive has collaborated with engineering computing platform Rescale and NVIDIA to implement a digital engineering platform combining Physics AI models and AI agents. Integrating computer-aided engineering (CAE), systems engineering, and design data, the system addresses computational bottlenecks in traditional CAE simulations. The system utilizes surrogate models trained on prior simulation and experimental data to predict physical performance at millisecond speeds. This enables engineers to screen thousands of design variations within hours. Rather than replacing traditional CAE, surrogate models serve as an initial filtering stage, with critical parameters subsequently verified through high-fidelity numerical analysis and new data fed back into the models. AI agents automate repetitive engineering tasks and pipeline integration, supported by engineering knowledge graphs and machine learning models predicting carbon fiber manufacturing performance. Rescale claims the platform can increase engineering productivity up to threefold, though public details regarding measurement criteria and operational scope remain unspecified.

Key takeaways

• McLaren deployed an engineering platform coupling Physics AI surrogate models and AI agents to accelerate design evaluation.
• Surrogate models predict physical performance at millisecond speeds to screen designs prior to rigorous CAE verification.
• AI agents automate routine workflows alongside knowledge graphs and ML models for carbon fiber component manufacturing.
• Rescale claimed an engineering productivity increase of up to 3x, though validation criteria and scope were not detailed.