Could tech replace firefighters in the battle against wildfires?
AI-assisted cameras from French startups are replacing human lookouts in wildfire detection, spotting fires up to 20km away and cutting response times by 15 minutes.
Intelligence analysis by Llama
France's worst-hit wildfire region has swapped student watchtower lookouts for AI cameras from startups FireTracking and Midgard, which detect smoke at 20km and alert firefighters before any phone call. Suppression still leans on ground crews and 30-year-old Canadairs, but heavy-lift water drones and sensor-fused patrols point to a future of autonomous first response.
Fires are scary, and catching them early matters a lot. In France, smart cameras with computer brains now watch the forests 24/7 from 20km away and text firefighters the second they spot smoke, beating the old human lookouts in towers. Drones might help put fires out soon, but people with hoses still do the real work for now.
Analysis
AI Lookouts Trade Watchtowers for 20km Vision
The most tangible shift in French wildfire response is not a new plane or a new hose but a new pair of eyes. Firefighter and CFDT union branch head Guillaume Millet, who has 24 years on the job, says cameras have replaced the student lookouts who once staffed Gironde's watchtowers. The AI can distinguish smoke from tractor dust with what Millet describes as "almost total certainty," and it scans up to 20km. Two French startups, Midgard and FireTracking, supply the systems, with FireTracking alone now surveying roughly a million hectares. False alarms, long a complaint of fire services, have fallen accordingly.
For FireTracking, the product is more than a camera. CEO Jean-Simon Chaudier describes an integrated dashboard: a real-time video feed with 40x zoom, a map of nearby houses, and a minute-by-minute simulation of how the fire will spread given humidity and wind. The system pushes alerts directly to firefighters' phones. The point is to compress the gap between ignition and first response, and Chaudier claims the AI detected seven fires in ten days a full 15 minutes before the first 18 alert, the difference between a limited burn and a catastrophe.
The Detection-to-Suppression Gap
Detection is only half the battle, and the article is explicit that the rest remains stubbornly analogue. Millet is blunt: "It can help us with detection but it's not technology that puts out fires; you need people to pull the hoses through the forest to attack them." Even a Canadair's 6,000-litre drop only reduces intensity; ground crews finish the job. Dash planes spraying orange ammonium phosphate retardant buy time but do not replace the humans underneath.
A handful of French startups are trying to close that gap. Daidelos is fundraising for a petrol-engine, heavy-lift drone capable of hovering over a fire and dropping 1.2 tonnes of water. Other firms are developing drones that release extinguisher grenades on contact with extreme heat, smouldering the surrounding area with chemical powder. The hope, Chaudier says, is an autonomous first response that extinguishes fires as they start, though he concedes such systems would be useless against an established wildfire.
The Window of Opportunity: Three Minutes
Chaudier's arithmetic captures the stakes: one minute and a glass of water suffices; two minutes needs a tank; three minutes requires a Canadair. That exponential cost curve is why detection AI is winning attention even when suppression drones lag. France's civil protection service is running a programme called Condor, testing solar-powered patrol drones with electro-optical cameras and onboard AI, while Spain's SenForFire project places sensors among the trees to measure gases released before flames become visible. Sensor fusion, cameras, chemical sniffers, and satellite data, is the direction of travel. The article's most forward-looking line, about drones extinguishing fires "as they're starting," gestures at the next frontier of fully autonomous first response, and a model where detection and suppression are bundled into a single automated chain.
Key points
- AI-assisted cameras from French startups Midgard and FireTracking have replaced student lookouts in Gironde watchtowers, detecting fires up to 20km away and sharply reducing false alarms.
- FireTracking CEO Jean-Simon Chaudier says the system detected seven fires 15 minutes before the first phone call, the difference between a limited blaze and a catastrophe.
- Detection AI is saving time, but suppression still relies on ground crews with hoses and 30-year-old Canadair water bombers, according to firefighter Guillaume Millet.
- Startups like Daidelos are fundraising for heavy-lift water drones, and other firms are developing heat-triggered extinguisher grenades, though none are operational.
- France's Condor programme is testing solar-powered patrol drones and Spain's SenForFire is measuring pre-combustion gases, hinting at a future of sensor-fused autonomous first response.
If detection AI scales across the Mediterranean basin as the article suggests, the early-warning window could compress from minutes to seconds, and emerging drone programmes like Daidelos and Condor might eventually automate the initial suppression response. Chaudier's "glass of water in one minute" framing implies that catching fires at ignition could drastically reduce the human and ecological cost of wildfire seasons across southern Europe.
The technology is detection-only for now, and as Millet stresses, wildfires still require ground crews in the forest. Heavy-lift drones like Daidelos are still unfunded, autonomous fire-suppression remains years from deployment, and there is a real risk that over-reliance on detection tech diverts investment from the firefighters, water bombers, and forest management that actually stop major blazes once they take hold.
