Cal Fire spent July 15 field-testing fully autonomous firefighting drones built by California startup Seneca, flying alongside nonprofit partner FireWERX and a team of engineers. The drones operate in packs of five and convert onboard water into hundreds of gallons of foam aimed at young fires before they grow. It’s the third time in nine months this company has crossed my desk, and the test moves Seneca from demo footage to a state agency’s own proving ground.
Firefighters With Brooms Get a Force Multiplier
Cal Fire’s Fire Innovation Unit ran the July 15 field test together with FireWERX, the nonprofit that anchors the agency’s technology work, to find out whether Seneca’s autonomous drones can take some of the load off hand crews who still attack young vegetation fires with backpack pumps and hand tools.
Deputy Chief Jack Worden leads that effort. “(Firefighters) go out there with the backpack pumps and the brooms and the shovels and they can put it out really fast,” Worden told KMPH FOX 26. “The Fire Innovation Unit is trying to force multiply that. Take some of the weight off their backs.”
According to the station’s report, each aircraft weighs about 400 pounds (181 kg), sits on a carbon fiber frame, and carries roughly 20 gallons (76 liters) of water. Mixed with firefighting concentrate, a five-drone pack can lay down 500 to 1,000 gallons (1,893 to 3,785 liters) of foam.
The Fire Innovation Unit itself is a young experiment. Cal Fire runs it in partnership with FireWERX, with philanthropic backing from the Gordon and Betty Moore Foundation, as the agency’s hub for testing outside technology before it reaches the fireline.
Seneca Sells a Strike Team, Not a Single Drone
Seneca packages its product as a strike team: five autonomous aircraft plus a launch station, each drone carrying more than 100 pounds (45 kg) of Class A foam suppressant and discharging it at over 100 PSI (6.9 bar), with a stated goal of putting foam on a fire in under ten minutes.
The company is nine months old as a public entity. It launched in October 2025 with a $60 million round led by Caffeinated Capital and Convective Capital, the largest first raise I’d seen for a firefighting drone startup, which I reported in Firefighting Drone Startup Seneca Secures Record $60M.
Founder and CEO Stuart Landesberg came from consumer goods, where he took Grove Collaborative public. His stated mission for Seneca is bigger than one state: protect 500 million acres (202 million hectares) across the US and allied nations by 2035, starting with systems fielded during the 2026 fire season.
Agencies keep lining up to look. San Bernardino County Fire got a demonstration in December and named the aircraft in its release: the Argo-1, a semi-autonomous UAV with thermal tracking and precision targeting. Aspen’s fire district bought a five-drone strike team slated to go operational this summer, a deployment I dug into in Seneca’s Aspen Firefighting Drones Face the Real Test.
Here’s the wrinkle in the numbers. Aspen Public Radio described drones hauling 12 gallons (45 liters) of water each, or roughly 60 gallons (227 liters) once turned into foam, light enough to lift by hand and stage in a pickup bed. Either the platform grew a lot between February and July or the two reports describe different aircraft, and Seneca hasn’t published a spec sheet that settles it.
Autonomy Here Means Launch Without a Pilot’s Hands
Autonomous, in Seneca’s design, means the drones launch themselves when a fire call comes through local dispatch, navigate to an approximate location, and use infrared and video targeting to find and hit the heat, while a single remote operator supervises the flight instead of flying it.
That matches what the company markets on its own site: systems stationed in high-risk locations or carried on utility vehicles, auto-launching on a fire indication, spotting smoldering embers from above, and knocking down hot spots on their own.
The paperwork is the unfinished part. Back in February, Aspen Public Radio reported that details with the FAA, the Department of the Interior, and the US Forest Service still needed to be worked out before Aspen’s drones could fly real missions.
There’s also an uncomfortable backdrop I’ve reported before. The US Forest Service logged 218 drone incursions over active wildfires in 2025, the most ever recorded, and 184 of them happened during the Eaton and Palisades fires in Los Angeles. Every unauthorized drone forces manned firefighting aircraft to land.
Cal Fire now wants authorized drones launching themselves into that same sky. Airspace integration is the problem that decides how far this goes, and it’s a harder one than foam capacity.
DroneXL’s Take
Wildfire response is turning into the proving ground for autonomous drone infrastructure in the US, ahead of package delivery and ahead of inspection work.
Look at the ladder in DroneXL’s own archive. October brought a record $60 million round for a company with no public product. December brought a San Bernardino County demo, February brought a paid strike team in Aspen, and July brought Cal Fire flying the system with its own innovation unit. Nine months, four rungs, zero large-agency contracts.
The funding structure matters as much as the hardware. The Fire Innovation Unit runs on philanthropic money through FireWERX, which means Cal Fire gets to test unproven hardware without betting procurement dollars on it. That setup strips the usual career risk out of a failed public-sector pilot, which is exactly why these tests keep happening at this pace.
I keep coming back to the 218 incursions, though. The same airspace where a hobbyist’s drone grounds an air tanker is where California wants machines launching on their own authority. Regulators will have to hold both lines at once, banning one drone while clearing the other to fly itself, and a one-day field test can’t answer that.
Seneca made two public commitments: first systems in the field during the 2026 fire season, and an operational strike team in Aspen this summer. Both are now due, so the scoreboard reads itself. Whether Cal Fire converts a field test into a contract is an open question, not a prediction.
