AivexaNewsSearch
AI news for builders and product teamsChecked every hour

Flox Intelligence is building AI to keep wildlife away from trains, airports, and roads

Collected Oct 13, 2026

Flox Intelligence, a startup spun out of the KTH Royal Institute of Technology in Stockholm in 2020, is building wildlife-deterrent hardware and software for infrastructure like railways, airports and highways. The company is one of the finalists in this year's Startup Battlefield 200 at TechCrunch Disrupt, held at Moscone West in San Francisco from October 13 to 15. Founder Sara Nozkova developed the system after 13 years in transportation and autonomous vehicles, and after reading about the damage wildlife causes to both infrastructure and animals.

The problem is measurable. The FAA estimates wildlife strikes cost U.S. civil aviation about $272 million a year on average between 1990 and 2025. French rail operator SNCF says wandering animals affect more than 8,000 trains nationwide each year and rank among its five leading causes of delays. In Brittany, 120 trains hit animals in 2023 alone, disrupting 660 services and causing more than 15,000 minutes of delays.

What Flox sells is a battery-operated edge device that can be mounted on poles, trees and other structures. A motion sensor wakes the unit when it detects movement, triggering a camera whose on-device models identify the nearby animal. The device then emits a species-appropriate sound to drive it away. Because processing happens on-site rather than in the cloud, the system works where connectivity is poor and avoids shipping video off-site, which matters for deployments along remote track and perimeter fences. Nozkova says the system also learns from experience and varies its responses so animals do not habituate to repeated signals.

Nozkova said the traditional toolkit — sirens, lights, fences, falcons and culling — was not the right answer, and that her conviction was that a louder deterrent was not the solution. Her argument is that perception systems trained for engineered environments such as cities, factory lines and warehouses fail in nature, where conditions are variable and the subjects are alive.

About 150 Flox devices are live across North America, Europe and Australia. Reported sites include airfields and airport perimeters, road and rail corridors — especially fence ends and crossings along wildlife corridors — national parks, agricultural land, urban areas, mining sites and critical infrastructure such as electrical substations. Flox also offers a software platform showing clients where devices are deployed and what is happening on the ground, and generating reports for regulatory compliance and uses such as biodiversity monitoring. The stated longer-term ambition is to make the software portable enough to run on machines that work outdoors, including trains, cars and drones.

Commercially, the model has two parts. Customers buy edge units up front, then pay an annual software and AI subscription after the first year. Separately, Flox has OEM partnerships under which companies pay an annual fee to deploy its software on vehicles such as trains; train manufacturer Alstom is named as one such partner. Nozkova said the company expects to end 2026 with revenue of about $1.5 million.

The startup has raised about $5 million across pre-seed and seed rounds. The seed round was led by Unconventional Ventures and Tilia Impact Ventures, with co-investors across the rounds including GENIUS NY, Norrsken, Satgana and angel investors, according to Nozkova. Flox is now opening a $10 million Series A, which would mostly fund growth from roughly 150 devices to several thousand deployed units and extend the underlying system beyond its current wildlife use cases.

Competition exists on adjacent approaches. Robin Radar's Scarecrow uses species-specific distress calls to disperse birds and records the activity, while iCHASE sells a system that detects birds with computer vision and targets them with a moving laser, adjusting its response as needed. Nozkova points to breadth and field data as the moat, describing Flox's asset as ground truth on what actually makes a moose or a grizzly bear move safely.

The framing Nozkova offers for the category is that physical AI is leaving the factory for the outdoors, and that every machine operating outside — trains, trucks, drones, tractors, robots — will need to understand nature. In practice, the hard part of that claim is data rather than model architecture: species recognition is a well-trodden computer-vision problem, but knowing which sound reliably moves a particular animal, at a particular distance, without training it to ignore the signal, is an outcome question that only field deployments answer. This suggests the value sits as much in the accumulated response data and the reporting layer for compliance and biodiversity monitoring as in the detectors themselves.

A likely trade-off is that edge inference keeps latency low and bandwidth costs down but caps model size, so the range of species and conditions a single unit handles will be constrained by the compute it can carry on battery. Coverage is another constraint: 150 units spread across airfields, rail corridors, parks, farms, mines and substations is a wide footprint for a small fleet, which makes the jump to several thousand units a test of manufacturing, installation and support rather than of the model alone. The vehicle OEM route, with Alstom named, is arguably the larger opportunity, since software on a moving train covers far more track than a fixed pole, but it depends on partners shipping it.

Why it matters: teams working on edge perception, outdoor robotics and transport safety now have a concrete example of a wildlife-behavior model sold as hardware plus subscription, with an OEM channel alongside it. For product and platform builders, the interesting pattern is the split between up-front units and recurring software fees, and the compliance and biodiversity reporting layer that turns raw detections into something customers must buy annually. Whether the Series A closes and deployments scale to the thousands will indicate if that packaging travels outside airports and rail corridors, which is inference from the stated plans rather than a reported outcome.

Read at TechCrunch · AI

Based on reporting from the original publisher. Visit the source for full context and later updates.

Publisher excerpt

Flox's edge devices can identify animals and use species-specific sounds to drive them away from critical infrastructure.