Flexcompute Launches AutoInsight to Explore Thousands of Automotive Design Variants in Real-Time

Physics-informed AI platform transforms existing simulation and test data into real-time aerodynamic predictions, enabling instant design decisions and dramatically faster design exploration.

BOSTON, March 17, 2026 /PRNewswire/ -- Flexcompute, the physics company, announced the launch of AutoInsight, a physics-informed AI platform that transforms existing computational fluid dynamics (CFD) data into real-time aerodynamic intelligence for automotive design. AutoInsight enables engineering teams to instantly evaluate aerodynamic trade-offs as vehicle shapes evolve, bringing aerodynamic insight directly into the design process where critical decisions are made.

Traditional automotive development cycles lack the speed and capacity to explore the full range of possible designs, leaving thousands of potential configurations unexplored. At the same time, organizations already possess enormous volumes of CFD data generated across past vehicle programs. Yet much of this data remains underutilized once a program is complete. AutoInsight unlocks the aerodynamic insight contained in this data, allowing teams to extract value from historical simulation and test results rather than starting each development cycle from scratch.

Using this foundation, engineers can interactively modify vehicle geometry and immediately see the aerodynamic consequences. Designers and aerodynamicists can evaluate performance trade-offs in real time, allowing decisions to be made instantly rather than waiting for new simulation runs. This dramatically expands the design space teams can explore while accelerating iteration across the entire development cycle.

AutoInsight is built as an open ecosystem capable of ingesting simulation and experimental data from existing engineering workflows, including datasets generated from widely used CFD tools and wind tunnel measurements. Organizations can immediately leverage the data they already possess without restructuring established pipelines or toolchains.

Unlike traditional surrogate modeling approaches that require massive training datasets, AutoInsight can be trained with as few as 10–20 simulation samples while maintaining high predictive accuracy. As new simulation or experimental results are generated, the model continuously improves while preserving knowledge from previous vehicle programs. The platform also enables a closed-loop workflow with high-fidelity simulation, where candidate designs can be evaluated using Flexcompute's Flow360 solver to generate new training samples that feed directly back into AutoInsight, ensuring predictions remain accurate as vehicle architectures and design spaces evolve.

"For decades, automotive aerodynamic development has been constrained by legacy workflows that separate design from simulation," said Vera Yang, President and Co-Founder of Flexcompute. "AutoInsight represents a fundamental shift. By bringing physics-informed AI directly into the design process, we are transforming how teams explore designs, evaluate trade-offs, and ultimately how vehicles are developed."

With AutoInsight, Flexcompute continues to advance engineering workflows built on the foundation of accurate physics and intelligent automation. By transforming historical simulation data into real-time aerodynamic intelligence, the company is enabling automotive teams to explore larger design spaces, make faster decisions, and accelerate the development of high-performance vehicles.

Media Contact

Steve Defibaugh: [email protected]

SOURCE FlexCompute Inc.