Russia has deployed the fully autonomous V2U drone with AI target selection, while Ukraine is developing similar systems
According to the source, Russia has deployed the V2U drone with a chip made by Nvidia, which independently selects and strikes targets after launch without an operator. Ukraine is developing similar AI-assisted drones that can also operate without a connection to an operator, but they are not yet in mass production.
According to an analysis cited by the source, Russia deployed a drone designated V2U in early 2025 that can independently select and strike a target after launch without operator intervention. A Ukrainian analysis of recovered units revealed a small computer made by Nvidia inside the drone, designed to run AI decision-making software; the onboard AI uses computer vision for object recognition, tracking, navigation and target selection. According to the source, variants recovered in late 2025 no longer contained the communication systems typically used for operator control.
Russia has also used another AI drone in combat with a less complete form of autonomous targeting — in July 2025, during an attack on a petrol station in Zaporizhzhia that killed three Ukrainian civilians, the target area was selected by a human, while the onboard software independently determined the final point of impact. According to the source, Russia is also integrating similar technology into the Geran series of drones, based on the Iranian Shahed-136 drone.
Ukraine is focusing on modular AI packages (chips with software, sometimes also cameras) that can be installed in various platforms; one module can guide a drone to its target during the final 500 metres of flight. Former minister of digitalisation Mykhailo Fedorov said in July that Ukraine had used a drone with no radio link to an operator that selects its own targets in strikes on occupied Crimea, but that it was not yet in series production. In November 2025, a Ukrainian FPV drone completed a strike on a suspected Russian tank using AI guidance after its connection to the operator was cut — but the target had previously been selected by a human. Earlier claims about “fully autonomous” quadcopters deployed two years earlier near Bakhmut are challenged by reports that these were preprogrammed machines, rather than drones that selected targets independently.
The source states that reliable estimates of the number of drones deployed in this way are lacking and that they apparently account for only a small fraction of the millions of drones used by both sides. It also highlights the concern that the pace of development of this technology is outstripping the ability of Russia and Ukraine to maintain some form of human control over the deployment of weapons.
Why it matters
The source describes a shift towards drones that independently decide to strike a target without human approval, weakening the existing principle of human decision-making about the proportionality of force and the risk of civilian casualties — as illustrated by the aforementioned strike on a petrol station in Zaporizhzhia. For tracking the development of AI and its regulation, this is a concrete example of military deployment of autonomous decision-making systems outpacing existing frameworks for accountability and oversight.
Relevant practical impact
What this means
For a business
The deployment of AI systems that decide whether to strike without human oversight shows how quickly military applications are outpacing existing frameworks for controlling autonomous weapons. This provides relevant context for the broader debate about regulation and accountability for high-risk AI systems.
Risks and compliance More business impacts →Check the original
Event sources
only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.