A Greek startup wants to send a water bomber into the fire with nobody on board. Marcos Aerospace, founded in Athens in January 2026, is developing Quencher, a 4,500-liter (1,189-gallon) unmanned amphibious firefighting aircraft designed to scoop from the sea, drop through the night, and repeat until the fire is dead. It exists today as a preliminary design and an AI-generated render, with a patent application filed. I still think it points at the most consequential shift in aerial firefighting since the first air tanker.
The physics of the argument are hard to beat. Wildfires are at their weakest at night, when temperatures fall and humidity climbs, and night is exactly when most crewed firefighting aircraft park. Fire researchers put the decisive window for a new ignition at roughly the first 20 to 30 minutes. Attack inside it and the fire stays a smudge on a hillside. Miss it and you are evacuating towns.
In six years of running DroneXL I have watched each layer of the wildfire fight go unmanned: first the satellites that find the spark, now the aircraft meant to put it out. The pieces stopped being hypothetical this summer. Three fire-hunting satellites reached orbit last month, and Greece is flying more than 100 drone patrol bases this season.
FireSat Satellites Now Spot A Fire The Size Of A Classroom #
FireSat, the wildfire-detection constellation built by Muon Space for the nonprofit Earth Fire Alliance with Google Research, is designed to spot a fire as small as 5 by 5 meters (16 by 16 feet) anywhere on Earth and revisit every point every 20 minutes once all 50-plus satellites fly. The prototype launched in March 2025 and promptly found a small, cool fire in the western U.S. that existing satellite systems missed. The first three operational satellites reached orbit on July 7, giving fire-prone regions at least twice-daily coverage now and an hourly revisit target by 2029.
The detection layer is already stacking with ground systems. DroneXL covered how California pairs satellite data with 1,260 mountaintop cameras, 915 of them running AI smoke detection that has flagged more than 900 fires before anyone dialed 911.
Patrol Drones Turn A Satellite Ping Into A Confirmed Target #
A satellite alert still needs confirmation at treetop height, and that is patrol-drone work: launch on the ping and hand an incident commander live thermal video showing whether this is a campfire or the first minutes of a catastrophe. Detection firm SenseNet, citing agency research from the U.S., Canada, and Australia, puts the make-or-break window at 20 to 30 minutes. Response has to move at drone speed to fit inside it.
Greece, the country Marcos Aerospace calls home, is running more than 100 drone surveillance bases and three mobile command centers with thermal cameras this fire season, with the stated aim of catching fires in the first minutes. Germany’s Dryad Networks pairs gas sensors in the forest with its Silvaguard drone to catch fires at the smoldering stage, before flames appear. Multirotors excel at hover-and-stare, but fixed-wing VTOL patrol aircraft cover far more ground per battery, and a grid that has to watch an entire region needs that endurance.
The line between watching and fighting is already blurring. Cal Fire spent July field-testing Seneca’s self-launching drone packs, stationed in high-risk terrain, that spot smoldering embers from above and knock down hot spots on their own.
Quencher Is A Bet That The Night Shift Wins The Fire #
Quencher is designed as a pilotless amphibious scooper that lifts 4,500 liters (1,189 gallons) from the sea and keeps flying drop cycles after dark, with a remote pilot in charge of every release while automation handles navigation and terrain awareness. Founder Marcos Caramalengos, an aerospace engineer, laid out the safety case on LinkedIn this week: pilots flying repeatedly at low altitude, in smoke and turbulence, have almost no margin when something breaks. “For the most hazardous missions, the safest cockpit may ultimately be an empty one,” he wrote.
The company’s design targets are built around a Pratt & Whitney PW127 turboprop: a 440 km/h (273 mph) top speed and a 360 km/h (224 mph) cruise, with one operator coordinating up to five aircraft at once. The company is mapping a certification path through EASA and ICAO frameworks, and the concept has already drawn specialist aviation press coverage.
None of it has flown. Marcos Aerospace says so itself: preliminary design phase, with testing and certification to follow. Renders are not aircraft, and plenty of concepts die between the render and the runway. Certifying a pilotless aircraft this size to scoop among boats and operate inside active fire airspace will take years. The reason to take this one seriously is the problem it names, in the country that named it.
On July 25, 2023, a Hellenic Air Force CL-215 clipped a tree pulling out of a water drop near Karystos on the island of Evia. Commander Christos Moulas, 34, and co-pilot Periklis Stefanidis, 27, died in the crash. Three days earlier, Prime Minister Kyriakos Mitsotakis had called Greece’s 1960s-era scoopers “old, difficult [to fly] and prone to malfunction.” Two more Greek pilots died fighting fire on the same island in 2007. Taking the crew out of that flight profile does not save the aircraft. It saves the two people who used to be inside it.
Heavy-Lift Quadcopters Already Fly The Precision Missions #
While pilotless water bombers wait on certification, heavy-lift multirotors are working fires now. The [ DJI FlyCart 100 lifts up to 100 kg (220 pounds)](https://www.dji.com/flycart-100/specs) at sea level and lowers loads on a 30-meter (98-foot) winch, a precision tool for putting water and gear exactly where a crew needs them. DJI has pitched firefighting and emergency rescue as core missions
[since the aircraft’s debut](https://dronexl.co/2025/06/30/dji-flycart-100-unveiled-heavy-lift-drone/).
The logistics case is proven at the harshest job site on Earth. A FlyCart 100 moved 10,073 kg (22,207 pounds) of cargo between Everest Base Camp and Camp 1 this season, crossing in eight minutes terrain that costs a human six to eight hours. A quadcopter will never replace a 4,500-liter drop. Its job is the surgical one: a night hot-spot knockdown, or water lowered to a crew cut off from the road. And the machine best suited to that work is not for sale in the United States, a fact American fire chiefs may want to raise with Washington.
DroneXL’s Take #
The most famous drone at an American wildfire is still the DJI Mini that punched a hole in a Super Scooper over the Palisades Fire, and 2025 closed with 218 drone incursions over active fires, the most the U.S. Forest Service has ever recorded. That is the reputation this industry gets to live with. The fastest way out of it is drones doing the firefighting, visibly and at scale, until a drone over a fire means the ignition got caught in minute 12 rather than an air tanker got grounded.
Watch what our own archive did in eight weeks this summer. Greece scaled its detection grid past 100 drone bases in June. Cal Fire flew Seneca’s self-launching suppression packs in July. Sacramento put FireSat and 900-plus AI camera detections at the center of a state wildfire briefing days later. Canada’s prime minister named thermal drones in a national fire mobilization. Nobody coordinated those stories. The direction coordinated itself.
Two things make the thesis testable soon. The XPRIZE Wildfire finals flew in Nenana, Alaska in June: five teams attempting to autonomously detect and suppress a fire somewhere in 1,000 km² (386 square miles) of terrain within ten minutes, a pace the foundation calls four times faster than current best practice, with winners due later this year. And FireSat’s operational satellites are scheduled to begin regular fire tracking over the United States, Australia, and Portugal before the end of 2026. If both deliver, unmanned initial attack stops being an argument and starts being a scoreboard.
Aerial firefighting crews are among the bravest pilots flying, and the bill for that courage keeps arriving: two Greek airmen in 2023, two more on the same island in 2007. Every fire nation has its own list. Honor the crews by retiring the mission profile that kills them, the low pass through smoke into rising terrain. Keep people in command of every drop. Find the fire from orbit, confirm it with a thermal drone, and kill it before minute 30, in the dark if that is when it is weakest. Quencher may never fly. The argument it makes already holds.
Sources: Marcos Caramalengos on LinkedIn, Marcos Aerospace, Wildfire Today, Google Research, XPRIZE, Dryad Networks, Al Jazeera, DJI, SenseNet DroneXL uses automated tools to support research and source retrieval. All reporting and editorial perspectives are by Haye Kesteloo.