Ukrainian operators have already demonstrated the problem: a $500 first person view drone can destroy a $10 million armored vehicle, while the legacy interceptor built to defeat it costs $100,000 or more per round.
Ukrainian operators have already demonstrated the problem: a $500 first-person-view drone can destroy a $10 million armored vehicle, while the legacy interceptor built to defeat it costs $100,000 or more per round. That is the framing every counter-drone startup is leaning on, and Neura Defense Systems, a Saint Petersburg, Florida company that emerged from stealth on June 16, is the newest entrant betting it can flip the math.
The cheapest attacker in the air now costs less than the training round of the system built to defeat it. Frontline troops in Ukraine and the Middle East have lived with that asymmetry for years. A wave of American defense startups is now racing to break it.
The specific gap NDS is targeting is the drone that legacy systems cannot see. Fiber-optic-tethered drones, which are attack craft connected to their operator by a thin spool of wire, emit no radio frequency signal. Fully autonomous drones do not need to talk to a controller at all. Both are invisible to the RF detectors and electronic jammers that legacy counter-drone systems were built around.
"The drones that matter most right now are the ones legacy systems can't see," founder and CEO Sam Talari said in the company's launch announcement.
To find them, NDS is building what it calls a sensor-agnostic fusion layer. Radar, optical cameras, infrared sensors, acoustic arrays, and existing military feeds all feed into the same AI model. The system ranks threats and hands engagement decisions back to a human operator. NDS describes the architecture as autonomous detection with a human in the loop for the kill.
The hardware NDS is showing alongside the software is twofold. Raptor, the company says, is a minimum-viable prototype interceptor designed to be cheap enough to fire in volume. Sentinel is a broader family of sensors and effectors meant to be deployed alongside it. The sub-$200-per-round cost figure for Raptor is a target, not a booked price. Provisional patents have been filed. A flight demonstration is planned for 2027.
None of that is independently verified. The Raptor has not been tested in combat. There are no named customers, no fielded systems, and no third-party operator data. For now, NDS is a company press release.
That is not necessarily disqualifying. The category NDS is chasing is real. The Department of Defense has spent the last three years shifting procurement dollars toward cheap, attritable systems. Anduril, Epirus, and DroneShield are all racing toward versions of the same architecture: cheap sensors, fast software, cheap interceptors. The bet is that a flood of small, autonomous defenses can drown out a swarm of small, autonomous attackers.
The harder question is what changes if the bet pays off. Counter-drone operations have been the slowest corner of the U.S. military to automate, in part because the kill decision sits close to civilian infrastructure and friendly forces. A network of autonomous interceptors raises new questions about rules of engagement, allied interoperability, and how a human stays meaningfully in the loop when the engagement window is measured in seconds.
For now, NDS and its peers are pitching products. The real test is whether the new math holds up the first time a fiber-tethered drone appears over a base that legacy defenses were never designed to protect.