In a first, AI beats human champions in drone racing competition in Abu Dhabi

In a significant milestone for technology and racing, a drone powered by AI from the Netherlands has outperformed top human pilots in an actual racing event. The competition, took place in Abu Dhabi last weekend, was part of the A2RL x DCL Autonomous Drone Championship — a significant event aimed at advancing the boundaries of autonomous flying. The AI drone, developed by a group from MavLab at TU Delft University, triumphed in a direct competition against a leading human pilot in what organizers claim was the toughest head-to-head drone race ever conducted.

The gathering took place at ADNEC Marina Hall in Abu Dhabi and united 14 teams from around the globe, featuring nations such as South Korea, Mexico, Canada, Turkey, China, and the UAE. Every team was required to depend solely on artificial intelligence for controlling their drones — no remote controls, no joysticks, and no human operators. The drones zoomed at speeds exceeding 150 km/h along a challenging, twisting indoor course featuring broad gates, inconsistent lighting, and minimal visual cues.

Team MavLab’s AI drone was not only quick — it was intelligent. It finished two laps of a 170-metre track in just 17 seconds. This sufficed to secure victory not only in the AI Grand Challenge but also in a direct competition against a skilled human drone racer, who had advanced through the DCL Falcon Cup. The victory of the AI drone against the human pilot was the highlight of the competition that everyone discussed.

Every drone was furnished with identical hardware: a camera facing forward, a motion detector, and a Jetson Orin NX compute unit produced by NVIDIA. Equipped only with this onboard technology, each drone was required to make instantaneous decisions in real time. There was no external assistance — the drone had to handle everything itself, from determining the course to modifying speed and direction.

Organizers claim this was among the toughest drone races ever undertaken. The hall’s lighting and the kind of camera employed (a rolling shutter camera) hindered the AI systems from operating seamlessly. Nonetheless, the leading teams managed the challenge remarkably, demonstrating the significant advancement of autonomous technology.

The event featured additional race formats such as drag races and multi-drone competitions, with teams like TII Racing and MavLab achieving significant victories. A STEM initiative conducted during the event educated more than 100 Emirati students in drone operation, highlighting the broader influence of this expanding sector. Now that the drone race has concluded, attention shifts to the upcoming highlight — A2RL’s autonomous car racing series, scheduled to return in late 2025 at Yas Marina Circuit.