A wildfire the size of a beach bonfire. That is what the new FireSat satellite network is designed to detect, from orbit, faster than your neighbor can find their phone. California's existing ground-camera system is already beating 911 callers to roughly half of all fires in the state, and that was before anyone put 50 dedicated fire-watching satellites into space.
50 Satellites Pointed at the Ground, Looking for Fire
In July, a SpaceX rocket carried the first three FireSat satellites into orbit. According to The Guardian, they are the opening act of a planned 50-satellite constellation built for one specific purpose: catching wildfires while they are still small enough to stop.
Existing weather and Earth-observing satellites can see a large fire just fine. The problem is they do not revisit the same patch of ground often enough, and they lack the resolution to catch something small. FireSat is engineered to close that gap. Once the full network is up, it will scan every point on Earth roughly every 20 minutes using infrared sensors and AI that compares new images against older ones, filters out industrial heat sources, and accounts for local weather before it sends an alert to emergency responders.
Twenty minutes. That is the detection window the full constellation is aiming for, anywhere on the planet. If you have been paying attention to what wildfires have done to the western United States over the last decade, you understand why that number matters.
The Cameras Already on the Ground Are Working
FireSat is not arriving into a void. There is already a ground-level AI surveillance network spreading across the country, and it is quietly doing something remarkable.
Pano AI, a San Francisco startup, has installed more than 1,400 cameras across 17 states, The Guardian reports. The cameras sit on cell towers and mountain tops and scan constantly. During the day, the AI is hunting for smoke. At night, it switches to heat signatures. When it flags something, a human analyst reviews the images before any alert goes out. That whole process, from detection to dispatch, often takes just a few minutes.
Arvind Satyam, co-founder of Pano AI, describes the system's logic simply: "The AI algorithms are being run continuously on that imagery to say 'smoke, not smoke.'" It is not a complicated premise. It is just one that nobody had the infrastructure to execute at scale before now.
California Is Beating 911 to Half Its Fires
California has built one of the most extensive ground-camera networks in the country through Alert California, which feeds more than 1,200 cameras to Cal Fire. The Guardian visited one of Cal Fire's control centers, where Chief Phillip SeLegue stood in front of a wall of live feeds and explained what the system has actually delivered.
About half of all wildfires in California are now detected by the camera network before a single 911 call comes in. One recent fire near Fresno was spotted 20 minutes before emergency dispatchers heard from the public. Crews were already moving before most people in the area knew there was a fire.
"The fires that this system has helped us dispatch resources to are the fires you don't read about in the papers," SeLegue told The Guardian. "The ones you don't see." Cal Fire's official benchmark is suppressing 95% of fires at 10 acres or less. Early detection is how they get there.
Brendan Finnegan, assistant chief with West Metro Fire Rescue near Denver, put the operational value plainly: having eyes that can alert responders well before a 911 call is the difference between a contained incident and a catastrophe.
The Fire Weather Is Getting Worse. The Detection Window Is Getting Shorter.
Here is the context that makes all of this urgent rather than just impressive. Climate Central reports that parts of the western United States now experience roughly two additional months of fire weather per year compared to the 1970s. Not two more bad days. Two more months.
Rising temperatures, prolonged drought, and longer stretches of hot dry wind have made fires easier to start and harder to stop once they get going. The fires in Spokane, Washington this summer are only the latest proof of that equation. The window in which a fire can be caught and killed before it becomes a disaster is shrinking. The technology is racing to compensate.
Satellites cover what cameras cannot see: remote forests, distant mountain ranges, vast stretches of grassland with no towers and no people to call anything in. The two systems are designed to function together, giving firefighters a faster and more complete picture than either could provide alone.
The Dingo Take
You are supposed to believe the news about wildfires is always the catastrophic ones. The ones with names. The ones that burn a hundred thousand acres and make the evening news for two weeks. Those are the fires the country has trained itself to process as the baseline, as the new normal, as the thing we wring our hands about and then do nothing to structurally address.
What is actually happening here is different, and it deserves more attention than it is getting. Engineers and firefighters have quietly built a system that is winning battles the public never finds out about. Half of California's wildfires caught before the first 911 call. A near-Fresno fire contained because crews had a 20-minute head start on public awareness. These are not feel-good footnotes. They are proof that early detection infrastructure saves land, structures, and lives at a scale we cannot fully measure because the disasters are prevented before they generate headlines.
The hard part is that none of this fixes the underlying problem. Two extra months of fire weather per year is not a detection-technology problem. It is a climate problem, and 50 satellites watching every bonfire on Earth cannot change what is happening to the atmosphere. What this technology can do is buy time, contain damage, and protect communities while the broader political argument about whether climate change is real continues to be a thing that adults apparently need to have in 2026. So: good on the engineers. Considerably less good on everyone else.
Comments