Iranian designed, Russian built attack drones now arrive in thousand drone salvos. Ukraine's small, AI guided, often 3D printed interceptors are an attempt to make defending a city affordable again.
A low buzz cuts across the Ukrainian night sky, and a small drone, often little more than a 3D-printed airframe with an autopilot and a small explosive, tilts upward into the trail of a much larger aircraft. That larger aircraft, a Russian-launched Shahed-136, an Iranian-designed propeller-driven attack drone now mass-produced in Russia as the Geran-2, is on its way to a Ukrainian city. The interceptor's job is not to outfly it or outgun it. Its job is to be cheap enough that Ukraine can afford to lose it, and smart enough that, more often than not, it does not have to be aimed.
This is the new shape of the air war over Ukraine: a fight being waged largely inside the cost equation rather than the fire-control equation. According to EE Times, citing Royal United Services Institute (RUSI) analyst Robert Tollast, Russian forces have begun launching more than 1,000 drones in a single night as a sustained tactic, mixing Shahed-type munitions with decoys designed to drain Ukrainian interceptors and surface-to-air missiles. Conventional air-defense arsenals were designed around aircraft and cruise missiles, not cheap propeller drones, and their per-shot economics cannot scale to a thousand-drone raid. That is the problem Ukrainian engineers are trying to outflank.
Their answer is not a bigger missile. It is a smaller drone.
Across the country's defense-tech cluster Brave1, a Ukrainian government-backed hub that funds and curates domestic counter-drone development, a generation of low-cost interceptors has moved from prototype to combat use. The most public of them is STING, a Shahed-class interceptor developed by Wild Hornets, a volunteer engineering group that describes the system as combat-proven. Ukraine has also fielded the Bagnet interceptor, designed to hunt and ram incoming Russian drones midair, and the UEB-1 interceptor built by OSIRIS AI, reported to fly at roughly 315 km/h in its public unveiling. Wild Hornets, OSIRIS AI, and outlets such as United24 Media and Militarnyi that describe these platforms are manufacturer- or government-aligned sources; their performance claims, especially that 315 km/h figure, are vendor-supplied numbers and should be read as snapshots, not steady-state specifications.
What matters for the air war is that these systems exist at all. Most are built around small airframes, off-the-shelf components, and onboard computer-vision guidance that lets them pursue a target without GPS. In an environment where Russian electronic warfare (EW) and GPS jamming have made satellite navigation unreliable, that autonomy is the feature that lets a cheap interceptor still find a Shahed. The result is a defender's weapon priced closer to the drone it is trying to destroy than to the multimillion-dollar missile it replaces.
The industrial layer underneath is now visible from outside Ukraine. The Pentagon has awarded roughly $500 million to Perennial Autonomy for counter-drone interceptors whose autonomy stack draws on Ukraine-proven designs, according to the defense-industry trade outlet Inside Unmanned Systems. That figure has not yet been cross-checked against a published Department of Defense contract announcement and should be treated as reported rather than confirmed. Inside Ukraine, the Ministry of Defence has stood up the Brave1 Dataroom, a secure environment for training military AI models, institutionalizing the pipeline that feeds autonomy into the interceptor fleet. Both moves suggest the cost-curve reframe is no longer a Ukrainian improvisation. It is beginning to show up in U.S. defense procurement.
None of this settles the air war. The interceptor programs are recent, the kill ratios against Shahed-type targets are not publicly verified, and Russia is itself iterating: layered strikes now mix slow propeller drones with faster jet-powered designs and ballistic missiles that cheap interceptors cannot catch. Vendor-reported numbers, including UEB-1's 315 km/h, may drift as field performance comes in, and the same Ukrainian officials who speak on background about autonomy breakthroughs are also the first to admit that defending a city from a thousand-drone raid is not a problem any single interceptor solves.
What is no longer in doubt is the direction. The traditional air-defense model, built around multimillion-dollar surface-to-air missiles aimed at multimillion-dollar aircraft and cruise missiles, was already uneconomic against cheap drones. Ukraine's small, AI-guided, often 3D-printed interceptors are a deliberate attempt to push the cost curve back in the defender's favor, to make a Shahed expensive to use, not just expensive to shoot down. Whether that equation holds at scale, in weather, and against Russian countermeasures is the open question. The bet is being placed in hardware, and in the night sky, right now.