Tornyol's mosquito-hunting micro-drone scores its first air-to-air hit

An AI micro-drone built to kill mosquitoes scored its first air-to-air hit on July 14, 2026. The startup behind it, Paris-based Tornyol, went through Y Combinator’s Fall 2025 batch.

How it hunts without cameras

The system uses no cameras at all. A base station carrying a 380-microphone ultrasonic phased array sends out pulses and tracks the echo of a mosquito’s specific wingbeat frequency in 3D, out to about 8 metres, in complete darkness.

The 40-gram drone is purely the interceptor: it flies the predicted path and hits the insect with one of its four propellers. Founders Clovis Piedallu and Alex Toussaint built it, with Toussaint coming from missile maker MBDA, which explains a lot about the architecture. This is air-defense doctrine, miniaturized for the garden.

Price is around $1,100, or $50 a month.

Now the honest part

The first confirmed kill was a moth, not a mosquito. The demo also ran on external motion-capture tracking rather than the finished onboard stack, so the full camera-free pipeline has not yet closed the loop on its own. Flight time is about three minutes.

The company’s bigger claim, that 10 drones per square kilometre could cut mosquito-control costs 100x and help eradicate malaria, is its own projection, with no field trials and no third-party validation yet.

Why it is still worth watching

Caveats aside, the pieces are genuinely novel: acoustic-only tracking of insect wingbeats, a sub-50-gram interceptor, and a defense-grade engagement model applied to public health. If the onboard stack catches up to the motion-capture demo, and if the economics survive contact with real fields, this becomes one of the more interesting attempts to point autonomous hardware at a problem that kills hundreds of thousands of people a year.

For now, it is one moth down, and a very long way from malaria. Would you put one in your garden?