Flox Intelligence Brings Edge AI to Keep Wildlife Off Tracks and Runways
Flox Intelligence has reached the Startup Battlefield 200 finals at TechCrunch Disrupt with a system that keeps wildlife away from trains, airports and roads using edge AI. Disrupt runs October 13-15 at Moscone West in San Francisco. TechCrunch
The company spun out of KTH Royal Institute of Technology in 2020. It was founded by Sara Nozkova, who spent 13 years working in transportation and autonomous vehicles before joining the AI lab at KTH in Stockholm to develop the wildlife-deterrent system.
The architecture is local by design, with data processed on site rather than in the cloud. A motion sensor wakes the device. A camera feed is then run through on-device models, small AI systems that run directly on the hardware, to identify nearby animals. Once classified, the unit emits a sound matched to the species to move the animal away.
Flox varies those sounds so animals do not habituate, meaning get used to repeated signals and ignore them. Habituation is the failure mode for sirens, strobes and other fixed deterrents. Randomized, species-specific bioacoustic cues, or animal sound signals, are intended to preserve avoidance behavior over time.
The hardware is battery-operated and designed to mount on poles, trees and other structures along linear infrastructure such as tracks and roads. The unit is called Flox Edge, which detects animals to create species-specific warning sounds to deter wildlife from dangerous areas. Spectrum Local News
Alongside the devices, Flox offers a software platform that maps deployment locations and ground activity and generates reports for regulatory compliance and biodiversity monitoring. The company describes itself as "Wildlife intelligence for infrastructure." Its site frames the approach as combining AI, bioacoustics and behavioral science to decode wildlife communication and use that signaling to direct animals away. Flox Intelligence
The longer-term goal is portability. Flox aims to make its software portable to run on trains, cars and drones, shifting from fixed perimeter nodes to moving platforms.
About 150 Flox devices are now live in North America and Europe. The FAA estimates wildlife strikes cost U.S. civil aviation about $272 million per year on average between 1990 and 2025. In rail, France's SNCF says wandering animals affect more than 8,000 trains per year and rank among its five leading causes of delays. In Brittany in 2023, 120 trains hit animals, disrupting 660 services and causing more than 15,000 minutes of delays.
On funding, Flox has announced a $4 million Seed round led by Tilia Impact Ventures. Flox Intelligence
The broader context here is why the local design fits remote infrastructure. Fixed sensors with cloud backhaul struggle on remote track and airfield perimeters where power and connectivity are constrained. Wake-on-motion plus on-device classification keeps power budgets low and response times predictable. It also keeps raw video local, which simplifies data handling. Species classification at the edge is now routine. Sustained deterrence without habituation is not, and that will determine whether compliance and biodiversity reports reflect real risk reduction or only activity logging.
In my view, the optimistic case is straightforward. If low-power deterrence works reliably at the perimeter, rail and aviation gain a tool that protects schedules and wildlife at once, without fences, culling or constant human patrols. That is a modest intervention with compounding returns.
