When a wildfire ignites, the difference between a contained blaze and a destroyed neighborhood is often measured in minutes. A new constellation of AI-equipped satellites, together with a growing ground network of smart cameras, is starting to close that gap — detecting fires before anyone dials 911.
Reporting published Monday by the Guardian and Climate Central details how the first three FireSat satellites, carried into orbit by a SpaceX rocket in July, form the initial piece of a planned network of 50 satellites built solely to spot wildfires while they are still small enough to stop. For more context on this story, see our ongoing more AI stories.
FireSat: A Constellation Built for Fire
Existing weather and Earth-observing satellites can detect large fires, but they often lack the resolution or revisit the same location too infrequently to catch small fires shortly after ignition. FireSat is designed to fill exactly that gap: the system is built to detect fires as small as a beach bonfire, and once the full constellation is deployed, to scan every point on Earth roughly every 20 minutes.
The satellites use infrared sensors and artificial intelligence to search for heat that could signal a new wildfire. The system compares new images against earlier ones and accounts for factors such as weather conditions and nearby sources of heat before sending alerts to emergency responders — an approach designed to identify real wildfires while reducing false alarms.
Pano AI's Ground Network
In the mountains west of Denver, technicians recently climbed a 150-foot cell tower to inspect a pair of wildfire-detection cameras built by Pano AI, a San Francisco startup that has installed more than 1,400 cameras across 17 states. The cameras continuously scan the landscape from towers and mountaintops while artificial intelligence reviews every image — searching for smoke during the day and heat signatures after dark.
"The AI algorithms are being run continuously on that imagery to say 'smoke, not smoke,'" Pano AI co-founder Arvind Satyam told Climate Central. "At night-time, we're looking at heat signatures, so heat versus cold. And then we're able to zoom in and validate that it is a potential fire start."
If the system spots a possible fire, a human analyst reviews the images before an alert goes to fire agencies — a process that often takes only a few minutes.
Firefighters say that head start matters. "Having those eyes that can alert us well in advance of potentially a 911 call make it a valuable tool for us," said Brendan Finnegan, assistant chief with West Metro Fire Rescue near Denver.
California's System Catches Half of Fires Before 911
California has built one of the largest camera networks in the country through Alert California, a system of more than 1,200 cameras used by Cal Fire. Chief Phillip SeLegue said the network now detects about half of all wildfires before anyone calls 911.
One recent fire near Fresno was spotted roughly 20 minutes before the first emergency call reached dispatchers. Firefighters responded quickly and kept the blaze from growing. "At Cal Fire, one of our missions is to suppress 95% of our fires at 10 acres or less," SeLegue said. "This system helps us in accomplishing that goal."
The detection network has earned the trust of incident commanders. "We trust it enough that when we receive that detection, we're already starting the process of sending resources to it," SeLegue said.
Why Minutes Matter
Wildfires have become one of the most destructive natural disasters in the United States. According to Climate Central, some parts of the western U.S. now experience about two more months of fire weather each year than they did in the 1970s — a consequence of rising temperatures, prolonged drought and more frequent stretches of hot, dry and windy weather.
The fires in Spokane, Washington, this summer — including the Old Trails fire that burned neighborhoods in August — are the latest reminder that when a wildfire is detected can be just as important as where it starts.
Satellites and ground cameras are designed to work together: cameras can only see the landscapes in front of them, while satellites can watch vast stretches of remote forests, mountains and grasslands where there are no cameras and often no people to report a fire. Combined, they give firefighters a faster and more complete picture of where fires are starting.
A fire that is detected quickly and extinguished while still small rarely makes headlines — which is precisely the outcome firefighters are working toward. "The fires that this system has helped us dispatch resources to are the fires you don't read about in the papers," SeLegue said. "The ones you don't see."
An Expanding Playbook for AI in Disaster Response
The wildfire detection buildout is part of a broader trend of AI systems being deployed against natural disasters, from cyclone path forecasting to flood mapping. What distinguishes the fire-detection layer is its closed loop: satellite and camera data flows through machine-learning classifiers, gets validated by human analysts in minutes, and triggers a dispatch decision — sometimes before any human has noticed smoke on the horizon.
For emergency services agencies, the economics are also compelling. Camera networks and satellite data subscriptions cost a fraction of what a single major wildfire consumes in suppression spending, and the systems improve as more imagery accumulates. With the FireSat constellation still years from full deployment, the near-term gains will come from the ground: Pano AI and Alert California are both expanding their footprints, and fire agencies in other states are studying the California model.
The combination of orbital coverage and ground-level validation points to a future in which the question is no longer whether a fire gets reported, but how many minutes after ignition the first engine starts rolling.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
