Bold claim: Track limits in Formula 1 have long been one of the sport’s most heated battlegrounds, and 2026 brings a game-changing AI-driven approach to policing them. But here’s where it gets controversial: can technology truly settle debates that often hinge on perception, rivalry, and human judgment?
For years, enforcing track limits has sparked ongoing controversy at every race. Drivers push the boundaries to shave milliseconds from laps, teammates and rivals rush to report infringements, and denials run hot when accusations fly. This tug-of-war can consume significant time and energy, sometimes delaying outcomes. A vivid illustration from the 2023 Austrian Grand Prix: the FIA faced more than 1,000 suspected track-limit violations, underscoring the need for smarter tools to speed up decisions.
To address this, the FIA and Catapult collaborated on an automated detection tool that identifies when a car crosses the white lines and integrates it into RaceWatch—the race-control system that monitors track activity remotely. With hundreds of potential incidents per race, a system that can accelerate review and notify teams within seconds is invaluable. In FIA terms, this computer-vision based solution has reduced human review needs by about 95% when reaching a decision.
RaceWatch is the FIA and Catapult’s race management software that includes this AI-assisted capability. The system’s computer vision recognizes car silhouettes and compares their movements against predefined reference points captured by cameras to determine if a reference line was crossed.
Looking ahead to 2026, several practical upgrades are planned to increase transparency and speed. First, the FIA will deliver footage of track-limit infringements directly to the involved teams, reducing quibbles and speeding feedback. This enhances clarity and streamlines workflows, allowing teams to act quickly on the evidence.
The second upgrade is arguably the most transformative: a smarter detection system that rethinks data analysis. The AI-based recognition will run on high-performance GPUs, processing real-time data to verify every single lap with greater accuracy and speed.
A central component of this evolution is a centralized camera controller that lets officials set track distances from one point and distribute processing across the network. Chris Bentley, the FIA’s Single Seater Head of Information Systems Strategy, explains that the system will run computer-vision software on any networked machine, processing video in chunks and returning results, thereby handling more data than ever before.
This data surge ties into a separate, advanced positioning system developed with Catapult. By merging multiple data streams, the FIA can track a car’s position with high precision and build a real-time digital twin of track activity—combining positioning data, sector times, and ideal racing lines to interpret what’s happening on track, even when cameras don’t capture every angle.
“We’ll use geofencing on positioning data and timing delays at key positions to determine where a car went off and how its line deviated,” Bentley explains. He notes that while there’s always an optimal racing line, deviations are rarely dramatic and can be detected with richer data.
The core concept is ECAT—Every Car All Turns. This approach compares a car’s behavior against a reference model and cross-references it with micro-sector timing to identify when and where an infringement may have occurred. The system can monitor multiple cars simultaneously and flag deviations from the ideal line by analyzing longer travel distances and sector-time differences.
As Bentley describes, the aim is to enrich everything with video and data so the system diagnoses what happened rather than requiring manual hunting for infractions. The objective is to push the process from manual to semi-automatic, retaining a human review for final strike decisions and flagging, but reducing the time and effort involved.
The FIA is also working with circuits to improve camera coverage, though this isn’t uniform across venues. With ECAT and geofencing, cameras remain important but are no longer the sole foundation of analysis. RaceWatch can now indicate potential track-limit infringements using positioning data alone: abnormal deviations, virtual zones, or trajectories drifting from the ideal line trigger alerts.
“Moving to this level lets us manage cameras in one place, distribute computer-vision processing, and combine other data streams,” says Bentley. By year’s end, these capabilities will be operational, with ongoing improvements throughout 2025 and into 2026.
The topic of track limits has always carried a subjective edge. A more data-driven approach increases the fidelity of decisions, and many fans hope it will quell endless debates. Still, opinions will vary about whether the technology should or could remove controversy entirely, or whether it may introduce new debates about data interpretation and control.
Would you like to see more automated decisions with AI handling most track-limit calls, or do you prefer keeping a larger role for human judgment in final rulings? Share your thoughts in the comments.
