How AI and satellites help fight wildfires


In July, hundreds of firefighters are battling flames in Spain, Portugal and France.

A map fromĀ NASA’s Fire Information for Resource Management SystemĀ (FIRMS) shows more wildfires: Red dots are scattered all over the world — essentially only the white areas of Greenland and Antarctica are excluded. Each of theĀ dots marks a satellite-detected hotspot or active fire.

A satellite image from NASA's Fire Information for Resource Management System showing the entire world
NASA’s Fire Information for Resource Management System shows red dots that indicate a satellite-detected hotspot or active fireImage: NASA FIRMS

With heat waves and droughts, wildfires are occurring more frequently worldwide.

They often starts quite unspectacularly — a discarded cigarette or a spark blown away from a campfire is enough to set dry leaves and small branches on fire. After that, quick action is important to extinguish the fire before it can set larger areas ablaze.

SpottingĀ fires early with satellites

To act quickly, however, small fires need to be detected early. This is a problem in more remote areas. But before the human eye can see rising flames or smoke, satellites may have already sounded the alarm.

Much satellite data is freely available, such as that from FIRMS. Global data is publicly available within three hours of satellite observation — in the US and Canada, some of it is even available in real time.

The European Forest Fire Information System (EFFIS) also provides free data for Europe, the Middle East and North Africa, drawing on information from NASA,Ā Copernicus andĀ the European Earth observation program.

Despite all this free data, Munich-based startup OroraTech has developed a business model for wildfire detection. The formula: satellite data, artificial intelligence (AI), service — and all of it fast.

Real-time wildfire monitoring from space

“The systems of NASA andĀ the European Space AgencyĀ (ESA) usually only take images at a specific time of day,” explains Thomas Grübler,Ā chief strategy office atĀ OroraTech. For example, only in the morning or at noon.

Typically, however, wildfires tend to break out from the middle of the day onward. OroraTech fills this “afternoon gap” with its own satellites.

In addition, the data provided by conventional satellites is not geared towardĀ the detection of fires. If the resolution of a camera is not high enoughĀ small fires might be overlooked.

As a result, important data may only reach firefighters with a delay. During that time, fires can spread undetected.

A many holding up one of German startup OroraTech's satellites in Munich, Germany, Thursday, Sept. 13, 2023
Seeing through dense smoke from fires makes it difficult to survey the affected areas, no matter how big or small the satelliteImage: Matthias Schrader/AP Photo/picture alliance

In spring 2022, OroraTech launched its first satellites into orbit. The satellites are equipped with thermal infrared sensors that detect temperature differences, enabling them to detect heat at night and through smoke, although dense cloud cover can still limit observations.

What makes them special, says Grübler, who cofounded the company in 2018, is that the cameras have been miniaturized to fit inside a shoe-box-sized satellite. This makes them cheaper and they only require a fraction of the energy of larger satellites.

HowĀ AI turnsĀ wildfire data into action

According to the company, theyĀ currently operateĀ more thermal imaging satellites than anyĀ country or other company in the world. By 2028, the number of daily scans is set to reach a stage where every point on Earth isĀ scanned every 15 minutes.

Two firefighters holding a firehose in front of huge flames near the village of Cercosa during a wildfire in Vouzela, Portugal, July 4, 2026
How is the fire spreading? Where and how should firefighters respond? Real-time data helps to manage wildfiresImage: Pedro Rocha/REUTERS

However, detecting fire sources is not enough.

“We make this data actionable,” adds Grübler.

To achieve this, OroraTech analyzes not only its own data using AI, but also that of other satellites, such as those operated by ESA and NASA.Ā 

Customers thus receive information about fire sources within minutes, along with continuous updates on the potential progression of the fire and an assessment of its consequences.

Who uses OroraTech’s wildfire data?

Among its customers in 25 countries are government agencies,Ā fire departments, forestry companies, infrastructure and energy providers, insurance companies and environmental organizations.

OroraTech now also sells data from its own satellites to ESA and NASA.

The Greek government has been won over by the company, too.

In collaboration with ESA, a system was developed specifically for Greece,Ā consisting of four satellites, a ground station in Greece and OroraTech’s Wildfire Solution platform. This enables the continuous delivery of real-time wildfire detection data across Greece.

A man in the middle of the remains of his home after it burned down in a wildfires in the Los Angeles area in 2025
In 2024–2025, wildfires burned an area larger than India across the globeImage: Ringo Chiu/REUTERS

Systems that help better organize emergency response teams are particularly useful in countries where fires develop and spread very rapidly.

OroraTech monitors 407 million hectares of forest worldwide — an area almost as large as the European Union — including regions in Brazil, Chile, the US, Canada and Australia, according to its own figures.

New uses for thermal satellites

“I would not have expected the entire market to shift from a niche to the mainstream like this,” says Grübler. “By now, we are deeply connected and experts in the wildfire sector,” and due to the increasing number of fires, the market will likelyĀ grow.

In addition to wildfire detection, further potential business fields are emerging.

After all, it is not only fires that can be measured, but all types of heat emissions. For instance, information about how much water evaporates from fields could be collected for agriculture, heat islands in populated areas could be identified, oil spills at sea could be detected, and thermal data could also prove useful in the defense sector.

Among other applications, ships can be detected from space even when they have switched off their transponder.

“I suspect that in five years there will be use cases that we don’t even know exist today,” says Grübler.

This piece was originally published in German.

Leave a Reply

Your email address will not be published. Required fields are marked *