Breaking

Google Research Details FireSat AI System for Detecting Car-Sized Wildfires From Space

September 17, 2026 at 11:32 EDT

Infra & Chips
Foundation Models

Google Research says its FireSat work is moving from research toward operations, using AI and a growing satellite constellation to detect early wildfires and track how they spread before they become large emergencies. In a September 15 Ask a Scientist interview, the company said the system is designed to spot fires as small as 5-by-5 meters, or 25 square meters, and ultimately image the planet every 20 minutes with roughly 50 satellites around 2030.

Sep 2026 · Google Research

AI spots fires while they are still small

FireSat combines custom infrared satellites with machine learning to detect car-sized wildfires, reduce false alarms and build a global record of how fires spread.

5x5 m

target detection size

20 min

full scan cadence goal

operational satellites in orbit

~50

satellites needed for the goal

From stale imagery to near real time

Each orange block represents 20 minutes. The current-data gap is roughly 11 hours, while the full FireSat goal is one 20-minute interval.

11 h

typical stale imagery

20 min

FireSat full goal

1 · Sense

Infrared instruments capture heat signatures from low Earth orbit.

2 · Compare

AI checks the same place against past imagery and local context.

3 · Filter

Models separate dangerous ignitions from controlled burns and noise.

4 · Inform

Fire agencies receive faster signals for response and planning.

What supporters value

Earlier detection could give incident teams more time, improve perimeter awareness and create a stronger dataset for fire-spread models.

What still has to work

Alerts must reach responders quickly, false positives must stay low and data access has to remain dependable over time.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.