The XPRIZE Wildfire competition ended with no grand prize winner, but finalists proved that AI-based detection can identify blazes within minutes. The $11 million contest had two tracks: space-based detection over a large area of Australia and autonomous response in rural Alaska. In the detection track, 45 teams submitted technical solutions, with the top prize of $500,000 going to the UK-based SIRIUS Wildfire Alliance, which combined optical and radar satellite data to model fire spread. Team Snuffed of Arizona took $250,000 for its uncertainty-aware detection system, while special recognition went to MyRadar, DeepFire, and Mayday.AI, the latter adapting a volcanic ash algorithm to spot wildfire smoke through clouds.
Because existing satellites are built for general environmental monitoring rather than wildfire detection, teams focused on AI and software that could turn that data into actionable intelligence. Program director Andrea Santy said the goal was to detect fires before anyone else, noting that a 911 call remains the most common detection method and typically takes at least 15 minutes. In the autonomous response track, all three finalists detected the high-risk fire in under 10 minutes, and two deployed drones to suppress it, but none fully put out the blaze. Santy said one team did complete all actions within the time limit and hit the fire accurately, yet the grand prize still went unawarded.
The results highlight a gap between fast detection and effective suppression. While space-based systems can now alert firefighters quickly, autonomous firefighting technology is not yet reliable enough to stop a fire on its own. The competition's sponsors and organizers hope these tools will eventually keep firefighters safer, but the final tests show that practical, end-to-end wildfire response remains a work in progress.