Ukraine Deploys USD 2,500 AI Drones to Counter 5,000 Monthly Shaheds
Illustrative Image Only (Credit: Jacky Lee)
Ukraine is doubling down on low-cost, AI-assisted interceptor drones to blunt Russia’s expanding campaign of Shahed and other attack UAVs, aiming to preserve scarce missiles while keeping interception rates near 80% despite record-breaking barrages.
Since early November, Ukrainian officials and developers have highlighted new autonomous “kamikaze interceptors” such as STING, Octopus, and General Chereshnya AIR as part of a broader shift toward AI-driven unmanned air defence. Recent Russian strikes have involved hundreds of drones and missiles in a single night, including assaults using more than 800 Shaheds and missiles on 7 September and around 620 air-launched weapons on 20 September.
Battlefield Necessity Sparks Rapid AI Integration
Russia’s Shahed-type drones are cheap, relatively slow and noisy, but their sheer volume and low-altitude flight paths are straining Ukraine’s legacy air defences.
Ukrainian and Western analyses suggest Shahed-type drones cost on the order of USD 20,000 each, while some of the missiles used to shoot them down can cost millions of US dollars per shot.
A recent CSIS review estimates Shahed launches have stabilised at roughly 5,000 per month since May 2025, illustrating how industrial-scale production is now central to Russia’s air campaign.
Ukrainian air force data and independent analyses show that defenders still down or suppress around 75–80% of inbound drones overall, but interception rates have slipped from the very high levels seen earlier in 2025 as attacks intensify.
This cost and volume imbalance is what pushed Kyiv toward cheap, expendable interceptors guided by AI instead of relying solely on traditional surface-to-air missiles.
STING and Other Domestic Interceptors
One of the most visible systems is STING, developed by Ukrainian engineers and fielded in partnership with the armed forces:
STING is a small interceptor drone costing about USD 2,500 per unit, designed to ram and destroy Shahed-type drones.
It uses autopilot and computer-vision tools to track targets at night and under electronic jamming, while leaving humans in the loop for launch and engagement decisions.
Ukrainian officials and local media say several hundred STINGs are already in service and have shot down scores of Shaheds, though precise figures are not independently verified.
Alongside STING, the General Chereshnya AIR interceptor has become another symbol of Ukraine’s wartime innovation. According to the system’s developers and Ukrainian officials, Chereshnya-equipped units helped shoot down more than 500 Russian drones in a single month, although outside observers note that such tallies are difficult to confirm.
An AI-enabled “Drone vs Drone” Shield
The newest addition to this family is Octopus, an interceptor designed specifically to tackle low-flying Shaheds under heavy jamming:
In mid-November 2025, Ukrainian Prime Minister Denys Shmyhal announced that three enterprises had begun series production of Octopus, with 11 more preparing to join.
Octopus is optimised for night interceptions at low altitude, using an AI-assisted autopilot and sensors to hold a course under electronic attack while human operators authorise engagements.
Shmyhal and President Volodymyr Zelenskyy say these interceptors cost only “a few thousand dollars” each, enabling Ukraine to conserve high-end missiles like Patriot and SAMP/T for ballistic threats.
In parallel, Ukraine has rapidly expanded its broader drone ecosystem. Before the full-scale invasion, only seven domestic drone manufacturers were approved for government contracts; by early 2024 that number had grown to around 200 companies, with officials arguing it can support annual output in the millions if funding continues.
Relay Networks Extend Strike Reach in Jammed Skies
AI-enabled interceptors are only one side of Ukraine’s unmanned strategy. On the offensive side, Ukrainian units are also experimenting with relay networks to push low-cost first-person-view (FPV) kamikaze drones far beyond their usual 10–12 km control radius.
A widely reported example involves the Ivan Franko assault unit, which works with the Sternenko Foundation and other volunteers:
In 2025, the unit received a relay-equipped FPV complex that allows attacks over 60 km from the launch point by using intermediate drones as signal repeaters.
These relays, combined with upgraded batteries and multi-band transmitters, helped FPV teams hit Russian positions while remaining outside the range of many front-line jammers. Operators describe relay-based systems as “game-changers” for deep strikes.
Ukraine has also demonstrated the potential of long-range, partially automated attacks in Operation Spiderweb, a 1 June 2025 raid in which the Security Service of Ukraine (SBU) launched swarms of drones from inside Russia:
Spiderweb targeted five bomber bases across Russia and, according to Ukrainian claims, hit 41 aircraft, while US assessments point to at least 20 aircraft damaged, including strategic bombers.
Reporting indicates that pre-programmed routes and AI-assisted guidance were used to navigate around air defences, underscoring how autonomy can enable deep strikes even in heavily contested airspace.
Taken together, these developments show how Ukraine is blending AI, relays, and mass-produced FPVs into a layered drone architecture that both intercepts threats and projects force dozens of kilometres behind Russian lines.
Toward AI-Saturated Drone Fleets
While exact figures are hard to pin down, most analysts agree that AI now touches a large share of Ukraine’s drone operations, from:
Onboard navigation and stabilisation (autopilots and terrain following),
to computer-vision target recognition,
to software that helps operators prioritise targets and manage swarms.
What is clearer is the macroscale trend:
Ukraine’s defenders intercepted or suppressed about 4,242 of 5,312 drones of all types in October 2025 – just under 80% – even as Russia increased the size and frequency of its attacks.
In some individual barrages, such as a September assault involving more than 800 drones and missiles, Ukrainian air defences still managed to shoot down the majority of incoming Shaheds and other UAVs.
President Zelenskyy has repeatedly called for scaling domestic production to “thousands of drones per day”, and officials say that interceptor output alone could reach around 1,000 units daily once Octopus, STING and similar systems are fully ramped up, although these are political goals rather than verified output figures.
How Ukraine’s Low-Cost Systems Compare Globally
Ukraine’s home-grown interceptors now sit at the centre of a broader, international race to field cheap, AI-assisted counter-drone systems.
Ukrainian systems such as STING, Octopus and General Chereshnya AIR share several common traits:
Unit cost: typically in the low thousands of US dollars, significantly cheaper than both traditional air-defence missiles and many Western interceptor drones.
Concept of operations: launched from simple rails or vehicles, flown largely autonomously toward radar- or camera-designated targets, with humans confirming final engagement.
Performance: Ukrainian officials attribute hundreds of Shahed and other UAV shootdowns to these systems collectively, but independent observers caution that public tallies are difficult to audit in wartime conditions.
NATO’s AI-powered Counter-drone Import
On Ukraine’s frontlines – and now on NATO’s eastern flank – the US-developed Merops system is emerging as the most direct Western analogue:
Merops pairs ground sensors with the Surveyor interceptor drone, a small aircraft costing roughly USD 15,000 that can home in on enemy drones and destroy them mid-air.
In Ukraine, Merops has reportedly achieved over 1,000–1,900 kills against Russian Shahed-type drones and other UAVs, according to the system’s developers and NATO officials. These numbers have not been independently verified but are broadly consistent across multiple reports.
In late 2025, Poland and Romania began deploying Merops, with Denmark also slated to field the system, and US, Polish and Romanian troops are now training on it at Nowa Dęba in Poland.
Compared with Ukrainian interceptors, Merops is roughly an order of magnitude more expensive per drone, but it offers multi-mission flexibility and mature integration into NATO air-defence networks.
Russian and Chinese Efforts
Russia and China are also racing to add more autonomy to their unmanned arsenals, but in ways that differ from Ukraine’s “cheap swarms plus AI” approach.
Russia relies heavily on Lancet loitering munitions and Shahed-derived drones for both frontline and deep-strike missions. These systems incorporate basic electro-optical seekers and, in some variants, rudimentary image-based homing, but open-source analysis indicates they remain highly vulnerable to jamming and decoys and depend on human operators for most decisions.
China’s Feihong FH-97A is a prototype “loyal wingman” unmanned combat aircraft designed to fly alongside J-20 fighters, carrying air-to-air missiles or loitering munitions and coordinating swarms of smaller drones with AI assistance. However, the FH-97A remains a testbed, with no public evidence of operational deployment or combat use, and is vastly more complex and expensive than the expendable drones used in Ukraine.
In other words, while Russia and China are pushing AI into drones as well, Ukraine’s niche is the ultra-cheap, rapidly iterated, combat-hardened end of the spectrum, rather than large, exquisite platforms.
Implications for Asymmetric Warfare
The combination of AI-enabled interceptors, relay-extended FPVs, and foreign systems like Merops is reshaping the economics of air defence:
Instead of firing USD 1–4 million missiles at USD 20,000 Shaheds, Ukraine can increasingly rely on interceptors and FPVs that cost a few thousand US dollars each, making sustained defence more viable.
Despite Russia launching thousands of drones per month, interception rates near 80% mean many barrages now inflict less damage than their scale suggests, especially where critical infrastructure is well-protected.
At the same time, drones now account for the majority of battlefield equipment losses on both sides, up to 75–80% in some assessments, underscoring how unmanned systems have become the main attritional tool of the war.
The picture is not one-sided. Russia is scaling production of Shahed-type drones to what Ukrainian intelligence estimates could be tens of thousands per year, and continues to adapt its own electronic warfare and decoy tactics. Fog, snow and smoke also degrade the effectiveness of visual AI, forcing Ukrainian crews to constantly tweak algorithms and tactics for winter conditions.
Even so, the Ukrainian experience is now shaping global doctrine. NATO’s rapid adoption of Merops, the EU’s push for a “drone wall”, and China’s investment in loyal-wingman prototypes all build on lessons from a conflict where small, cheap, AI-assisted drones are routinely deciding whether shells land, bridges fall, or power stays on.
For Ukraine, the stakes are existential. Its AI-driven drone revolution is not a technology demonstration but a survival strategy – trading silicon and code against steel and explosives, one interception at a time.
TheDayAfterAI News
We are a leading AI-focused digital news platform, combining AI-generated reporting with human editorial oversight. By aggregating and synthesizing the latest developments in AI — spanning innovation, technology, ethics, policy and business — we deliver timely, accurate and thought-provoking content.