When I was eight, our train to Mumbai ground to a halt, not from a malfunction, but from a wild bison. I learned then that wildlife and infrastructure don’t mix. But as Sara Nozkova from Flox Intelligence points out, nature is complex and always alive, and machines built for controlled environments often struggle to adapt.

Flox’s AI system, which uses edge devices to process data on-site, is designed to emit species-appropriate sounds to scare away animals. It learns from experience, varying its responses to avoid habituation. This system is already deployed in North America, Europe, and Australia, showing the potential for this technology to reduce wildlife strikes on infrastructure.

While Flox is not the only player in the field, Nozkova believes the breadth of their system and the data they have collected in the field give them a competitive edge. As physical AI moves into the outdoors, every machine that operates outside—trains, trucks, drones, tractors, robots—will need to understand nature. Nature intelligence is the next layer of the perception stack.

Flox is now raising $10 million to grow, extending their system beyond its current wildlife use cases. With the potential to reduce delays and damage from wildlife strikes, this technology could be a game-changer for infrastructure around the world.