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

When a train in Mumbai slammed to a halt after colliding with a wild bison, the incident left a mangled tail dangling from a branch and a costly engine tow in its wake. That vivid memory sits at the heart of Flox Intelligence, a Swedish AI startup that is trying to keep animals out of the way of trains, airports and roads—not with louder sirens but with species‑specific sounds that speak the language of wildlife.

Why wildlife‑infrastructure clashes matter

The clash between nature and human‑built corridors isn’t a niche problem. The U.S. Federal Aviation Administration estimates that wildlife strikes have cost civil aviation roughly $272 million a year on average between 1990 and 2025. In France, SNCF reports more than 8,000 trains each year are affected by wandering animals, ranking wildlife among its five leading causes of delays. A single region in Brittany saw 120 train‑animal collisions in 2023 alone, disrupting 660 services and adding over 15,000 minutes of delay.

These figures illustrate a hidden cost chain: damaged equipment, service interruptions, regulatory headaches and, inevitably, animal casualties. The traditional toolbox—fences, lights, sirens, falcons and even culling—often fails to scale or adapt to the diversity of species that roam near critical infrastructure.

From a personal shock to a tech solution

Sara Nozkova, the founder of Flox, turned that personal shock into a mission. After 13 years in transportation and autonomous vehicles, she recognized a pattern: machines excel in engineered, static environments but go blind when nature intrudes. “We have taught machines to be brilliant, but only inside the environments we engineered for them,” she told TechCrunch. That insight drove her to the AI lab at KTH Royal Institute of Technology in Stockholm, where she began building a perception system that could not only spot an animal but also read its behavior and respond accordingly.

That research spun out into Flox Intelligence in 2020. The startup entered the spotlight as a finalist at this year’s Startup Battlefield 200 at TechCrunch Disrupt in San Francisco, showcasing a hardware‑first approach that mirrors autonomous‑vehicle tech: on‑device perception, edge processing and adaptive deterrence.

How Flox’s edge devices work

Flox’s hardware lives on poles, trees or any structure near a potential conflict zone. A motion sensor wakes the unit, triggering a camera that runs an on‑device model trained on wildlife behavior. Once the animal is identified, the device emits a sound tuned to that species—whether it’s a distress call for birds or a low‑frequency rumble for larger mammals. The system also learns from each encounter, varying its signals so animals don’t become habituated.

Beyond the hardware, Floz offers a cloud‑connected platform that maps device locations, streams real‑time detections and generates compliance reports. The software can also be repurposed for biodiversity monitoring, giving regulators and conservation groups a new data source. Long‑term, the company aims to embed its AI stack into moving machines—trains, trucks, drones—so they can “understand nature” on the fly.

Early traction and business model

Today about 150 Flox devices are live across North America, Europe and Australia, protecting airfields, rail corridors, national parks, farms, urban zones, mining sites and even electrical substations. The startup sells the edge unit upfront and then charges an annual software and AI subscription after the first year. It also licenses its software to OEMs; train maker Alstom is already a partner that pays an annual fee to run Flox’s AI on its rolling stock.

Financially, Flox expects to close 2026 with roughly $1.5 million in revenue. The company has raised about $5 million across pre‑seed and seed rounds, with the seed round led by Unconventional Ventures and Tilia Impact Ventures. Investors also include GENIUS NY, Norrsken, Satgana and a slate of angels. A $10 million Series A is now open, earmarked for scaling the device count from the current 150 to several thousand and expanding the platform beyond wildlife deterrence.

Competitive landscape and moat

Flox isn’t alone in the animal‑deterrent arena. Robin Radar’s Scarecrow also uses species‑specific distress calls, while iCHASE deploys computer‑vision‑driven lasers that track and chase birds. Flox differentiates itself through the breadth of its data set and the “ground truth” it has accumulated in the field—real‑world recordings of how moose, grizzly bears and countless other creatures react to specific sounds. Nozkova argues that this data depth creates a moat, giving Flox a richer model library than rivals that focus on narrower use cases.

The startup’s edge‑first architecture also offers a latency advantage. By processing video and audio on‑device, Flox sidesteps cloud latency and privacy concerns, a selling point for regulated sectors like aviation and rail where data sovereignty is critical.

Implications for the broader “physical AI” market

Flox’s story signals a broader shift: AI moving out of clean factories and onto the messy, alive outdoors. As autonomous trucks, delivery drones and field robots proliferate, each will need to navigate not just roads but forests, savannas and wetlands. Understanding animal behavior becomes a safety prerequisite, not a nice‑to‑have add‑on.

For everyday commuters and travelers, the payoff could be fewer delays, lower ticket prices and a reduced risk of wildlife casualties. For farmers and conservationists, the same technology could double as a monitoring tool, feeding data into habitat‑management plans. And for investors, Flox’s $10 million Series A round points to a nascent market where hardware, AI and environmental stewardship intersect.

What’s next for Flox and the industry

With a modest revenue runway and a clear growth plan, Flox’s next milestone is scaling deployment. The company aims to push devices into more high‑traffic corridors—especially fence ends and wildlife‑crossing points where animals are most likely to breach barriers. Simultaneously, it will refine its models to cover additional species, expanding the deterrent catalog beyond mammals and birds to reptiles and even insects that threaten power lines.

If Flox can hit its deployment targets, it could set a template for “nature‑aware” AI that other sectors will emulate. The ultimate test will be whether the technology can reduce the costly wildlife‑strike statistics that have plagued aviation and rail for decades. Success would mean a future where a train whistles past a herd without a single animal or engine harmed—a future that started with a mangled bison tail and a determined engineer who refused to settle for louder sirens.

This article was produced with AI-assisted research and editorial support. Reporting is based on the source material cited below. Sources: TechCrunch; techcrunch.com; Global1.News (14 October 2026).

By Nova Chen, Staff Writer