Sebastian Asprella

Breaking: Ornadyne has launched out of YC to build robotic birds for military reconnaissance. C-UAS has a potential blind spot: birds. Traditional systems are built to detect rotor and fixed-wing drones, while flapping-wing platforms introduce a different signature set, closer to how birds move, look, and sound in the field. The U.S. has been here before. AeroVironment built the Nano Hummingbird in partnership with DARPA, a miniature drone engineered to look and fly like a hummingbird. It flew for 11 minutes and ran surveillance mission sets in tight indoor environments, a key proof point for urban operations. Earlier still, in the 1970s, the CIA developed a dragonfly-shaped, gas-powered drone for intelligence collection. The platform never made it past crosswind control problems. America has had a few notable ornithopter projects (robotic bird-scale flapping wing), but development has lagged behind countries like China, which has fielded dove look-alike drones reportedly capable of mimicking 90% of a real bird's movement. America has a potential answer via YC. Ornadyne, which just launched, builds "Robot Birds for Reconnaissance." We spoke with the team. Their first target customer is U.S. SOCOM. "Special operations users have an acute need for small, portable ISR systems that can operate close to the target while minimizing visual and acoustic cues," the team told us. "That mission profile maps well to what a bird-like Group 1 aircraft is designed to do." Law enforcement is a longer-term market expansion. Below is our Q&A with the team, lightly edited for clarity: Question: How does Ornadyne's bird-like form factor help it evade today's counter-UAS systems compared to quadcopters or FPV drones? Ornadyne's advantage is not that a bird-like drone is "invisible," but that it presents a very different signature than the drones most counter-UAS systems are optimized to find. Most small drones today have obvious cues: exposed propellers, strong tonal acoustic signatures, rigid-body flight paths, and visual profiles that look like quadcopters or fixed-wing aircraft. A flapping-wing platform changes that signature. It removes propeller-driven flight as the dominant acoustic and visual cue, uses a more natural silhouette, and moves in a way that looks closer to biological flight. In many counter-UAS systems, birds are not the target. They are the noise. This is because counter-UAS systems have to manage false positives. Birds, clutter, and harmless objects can overwhelm detection pipelines, so many systems are tuned to classify or filter those out. Ornadyne's thesis is that by building ornithopters that look, move, and sound more like birds than drones, we can operate in a harder-to-classify signature space: falling into that noise instead of standing out as another quadcopter, FPV drone, or fixed-wing UAS. Question: The U.S. had DARPA’s Nano Hummingbird over a decade ago. Why has there been so little visible progress since? DARPA clearly understood the value of biomimetic flight early. The Nano Hummingbird was an incredible technical achievement. The issue was not that the concept lacked value. The issue was that the timing was early. At that scale, endurance, payload, manufacturability, control, and cost were not yet aligned with what the military needed. A drone that looks exactly like a bird is awesome, but if it only flies for minutes and is difficult to build or ruggedize, it's not really productizable. The market just rewarded much simpler architecture. Quadcopters, fixed-wing Group 1 UAS, and systems like the Black Hornet by Teledyne, which ultimately took the place of the Nano Hummingbird. The U.S. helped pioneer bird-like flight with the Nano Hummingbird, but after that lead we didn't really pursue it much further. Now, Europe and China are far ahead in both academia and industry with bird robotics. Ornadyne’s mission is to put the U.S. back at the forefront of bird-like aerial systems starting with the longest-flying ornithopter produced in the states. Question: How much processing can happen on the edge given the size, weight, and power limits? For the first version, we're prioritizing flight performance and payload over heavy onboard AI. The near-term architecture is to stream imagery to the ground station, where more compute-intensive processing can happen offboard. Over time, as chips get more power efficient, and model distillation gets better, we'll begin to integrate basic perception, navigation, and lightweight autonomy features. The key is not to overload the first aircraft with every autonomy feature on day one. Our near-term goal is to build a useful low-signature ISR platform. Then progressively add onboard intelligence as the aircraft’s power and payload margins improve. Question: What new / unique mission sets does this unlock? The first mission set is low-signature close-range reconnaissance in environments where conventional drones are too obvious. A bird-like aircraft creates a different option: persistent or semi-persistent observation from a platform that looks and sounds less like a traditional drone. That unlocks missions where the goal is not just to see the target, but to observe without immediately changing the target’s behavior. Examples include reconnaissance in contested areas, surveillance around sensitive facilities, route observation, perimeter monitoring, etc. Basically, anywhere where drone presence would compromise the mission. Our birds can also be configured with mission-specific colorways and proportions based on bird species common to the operating region. Question: If Ornadyne works, how does it change the battlefield? If Ornadyne works, it changes the assumption that every useful aerial ISR platform has to look like a drone. Today, the battlefield is adapting quickly to quadcopters, FPV aircraft, and fixed-wing UAS. Fighters can hear them, spot them, jam them, and increasingly treat anything drone-shaped as a threat. Ornadyne is taking a different approach. Instead of making a conventional drone slightly quieter, slightly smaller, or slightly harder to see, we are building aircraft that blend into the natural aerial environment and are quieter via first principles. There is no camouflage or “invisible paint” for something flying in the open sky. The best camouflage is to look like something that already belongs there (especially at low altitudes). Birds are already everywhere, and most people do not treat them as suspicious. If we can make that practical for ISR, bird-like aircraft become a new class of reconnaissance platform: one designed around ambiguity, low observability, and naturalistic flight from the beginning. Question: Are you building a better small drone, or a new class of ISR platform current counter-UAS systems were not built to recognize? We are building a new class of ISR platform. A better quadcopter is still a quadcopter. A better fixed-wing drone is still a fixed-wing drone. Ornadyne is building around the assumption that the final frontier of close-range ISR should not just be smaller, faster, or cheaper. It should be harder to notice, harder to classify, and easier to mistake for something already present in the environment. So the product is not just a better small drone. It is a robot bird designed for a world where conventional drones are becoming easier to detect and counter.

in the next 10 years ROMULLUS will manufacture, build and deploy more strategic infrastructure than London has in 50 years. in the following 10 years it will manufacture, build and deploy more strategic infrastructure than Britain has in 50 years.

Europe’s defence budgets are rising. That does not reduce buyer diligence. In C-UAS and site protection, the bottleneck is still documented performance in relevant operating conditions.

YC’s latest Requests for Startups includes Counter-Swarm Defense. The important signal: counter-drone defence is shifting from standalone hardware to real-time distributed systems. Not one drone. Swarms. Not one sensor. A fused picture. Not a weapon problem. An integrated sensing + decision stack. That is the architecture shift we are building for at ÆTHERION.

This kind of smart, walkable, anti-drone urbanism will become mandatory to build in many American cities

DJI unveiled its new FC200 transport drone. Standout feature: a four-drone coordinated formation that has a max payload of 600 kg. This will change lots of industries.