I remember when I was eight years old, the train to Mumbai came to a screeching halt. About eight hours later, as we started moving again, someone pointed out a dismembered tail hanging from a branch near the railroad tracks.
As you can see, the train apparently hit a wild bison at full speed, and the impact damaged the engine so badly that it had to be towed away for repairs. For a long time, I thought this kind of thing only happened in places like rural India, but a few years ago I learned that wildlife and infrastructure don’t mix anywhere on the planet.
The FAA estimates that from 1990 to 2025, wildlife strikes cost U.S. civil aviation an average of approximately $272 million annually. French railway operator SNCF says wandering animals affect more than 8,000 trains across the country each year and are among the top five causes of delays. In Brittany, 120 trains collided with animals in 2023 alone, disrupting 660 services and causing more than 15,000 minutes of delays.
Reading about all this damage to both infrastructure and animals, Sarah Nozkova (pictured above, left) was inspired to develop a wildlife deterrent system that would work differently than the sirens, lights, fences, falcons, and culling traditionally used to keep animals away from roads, railways, airports, and other infrastructure.
“I was confident that the answer wouldn’t be much of a deterrent,” she told TechCrunch.
After working in the transportation and self-driving vehicle fields for 13 years, Nozkova learned how quickly the systems he built to navigate roads and other controlled environments can break down when nature intervenes.
“We have taught machines to be good, but only within the environment we designed them to be,” she said. “Cities, factory lines, warehouses. Fixed geometry, known objects. When I stepped outside into nature, I saw how the same machines could quickly go blind, because nature is complex and always alive.”
So she instead joined the AI lab at the KTH Royal Institute of Technology in Stockholm and began working on a system that could change deterrent responses based on what nearby animals were doing, and even the type of animal they were.
That effort became Flox Intelligence, which was spun out of KTH in 2020 and is one of the finalists in TechCrunch Disrupt’s Startup Battlefield 200 pitch competition, which will be held October 13-15 in San Francisco at Moscone West this year.
Flox makes edge devices that process data on-site rather than in the cloud. Similar to self-driving cars, they use sensory systems to sense their surroundings, but their systems are trained based on the behavior of wild animals. When a motion sensor detects movement, it wakes up the device and activates a camera that uses a model on the device to identify nearby animals. The device then emits a species-appropriate sound to scare the animal away.
Nozukova said the system also learns from experience and changes its response to prevent animals from becoming accustomed to repeated signals.
The company currently sells battery-powered devices that can be attached to utility poles, trees, and other structures. We also provide a software platform that shows clients where their devices are located and what is happening in the field. The platform can also generate reports for applications such as regulatory compliance and biodiversity monitoring. The long-term goal is to make the software portable enough to run on machines that operate outdoors, such as trains, cars, and drones.
Nozkova said about 150 Flox devices have been installed around airfields and airports in North America, Europe and Australia, in critical infrastructure such as road and rail corridors (particularly at intersections along fence edges and wildlife corridors), national parks, farmland, urban areas, mining sites and electrical substations.
The company sells Edge units upfront and charges annual software and AI subscriptions after the first year. Flox also has OEM partnerships with companies that pay an annual fee to deploy the software on trains and other vehicles. Train manufacturer Alstom is one such partner.
The company expects to have revenue of about $1.5 million by the end of 2026, Nozkova said.
Although Flox’s product and approach are unique, the startup is not without competition. The Robin Radar Scarecrow also uses species-specific distress signals to scare away birds and record their activities. iCHASE, on the other hand, sells a system that uses computer vision to detect birds, targets them with a moving laser, and adjusts the response as needed.
Nozkova said her moat is the breadth of the system and the data Phlox collects in the field, which she calls “the truth about actually moving elk and grizzly bears safely.”
“Physical AI is coming out of factories and into the cluttered, vivid outdoors,” Nozkova says. “All machines that operate outdoors, such as trains, trucks, drones, tractors, and robots, need to understand nature. Natural intelligence is the next layer in the perception stack. Very few people are building it.”
The company has raised approximately $5 million in pre-seed and seed rounds to date, with the seed round led by Unconventional Ventures and Tilia Impact Ventures. Co-investors throughout the round include GENIUS NY, Norsken, Satgana and angel investors, Nozkova said.
The company has now opened a $10 million Series A, which will primarily go toward scaling from approximately 150 devices to thousands of deployed units and expanding the underlying system beyond its current wildlife use case.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
