Flox Intelligence is using edge AI to identify wildlife near railways, airports and roads, then trigger species-specific sounds intended to move animals away from danger. The Swedish startup says about 150 devices are operating across North America, Europe and Australia, and it expects roughly $1.5 million of revenue in 2026.
The system runs locally rather than depending on a constant cloud connection. A motion sensor activates a camera, on-device models identify the animal, and the unit selects a deterrent sound. Flox says it varies responses to reduce habituation, the tendency of animals to ignore a repeated signal.
Commercial evidence is emerging
Alstom confirmed extensive field tests of the technology along Swedish railways. The train manufacturer said early tests showed potential to identify animals accurately and reduce collision risk. That independent partner confirmation matters because many of Flox's deployment and performance claims otherwise come from management.
Flox sells hardware upfront and charges an annual software subscription after the first year. It also licenses software to original-equipment manufacturers, with Alstom named as a partner. The company says customers use its platform for incident reporting, regulatory compliance and biodiversity monitoring.
Management reports that wildlife strikes cost U.S. civil aviation an average of $272 million a year from 1990 through 2025, while animal incidents disrupt thousands of French trains. Those figures define a real operational problem, but they do not establish how much of the loss Flox can prevent.
Scaling remains the test
Flox has raised about $5 million across pre-seed and seed financing and is opening a $10 million Series A. Management wants to move from roughly 150 devices to several thousand. That expansion will require proof that detection accuracy holds across different species, weather and geography, and that recurring software revenue grows after hardware installation.
The investment significance lies in adoption, not the novelty of an AI wildlife detector. Verified collision reductions, renewals and larger fleet contracts would support the business model. Until those data arrive, Alstom's trials provide a credible validation step, while the revenue and device counts remain management-reported figures.