CASE STUDY - PREMIER FORMULA 1 TEAM
NEXUS predicted the behavior of an F1 car before it hit the track
NEXUS predicted the behavior of an F1 car before it hit the track
10.3%
Front aero balance
Track time
Track time is the scarcest resource
An F1 team can't endlessly test its way to the right answer. Track time is scarce, every setup change interacts with the car, circuit and conditions, and the cost of learning in the physical world is extraordinarily high. So what if more of that learning could happen before the car ever left the garage?
Predicting the physical behavior of a machine
In a proof of concept with a Formula 1 team, NEXUS was tasked with predicting front aerodynamic balance before the car turned a wheel, using structured data describing the car and the environment around it.
Track
Track characteristics
Weather
Weather forecasts
Car setup
Car-related features
This wasn’t predicting a fan, a player or a commercial outcome. It was predicting the physical behavior of a machine.
10.3%
improvement in predictive performance
NEXUS produced a materially more accurate prediction of front aerodynamic balance from the car's setup and circuit conditions before physical testing.
Prediction doesn't stop at the digital world.
That 10.3% changes what an engineer can know before consuming one of the team's scarcest resources: track time. Better predictions allow physical testing to become more targeted and engineering decisions to start from a more accurate understanding of how the car is likely to behave.
And that's the bigger idea: prediction AI doesn't have to stop at the digital world. It can help us understand what is about to happen in the physical one.
