In 2024 and 2025, I served for six months as an international volunteer on a first-person view attack drone team in the Armed Forces of Ukraine. My team was deployed in the Donbas region, in one of the hottest sectors of the front. When I joined the team, I was excited to work with a cutting-edge tool. By the end of my deployment, I was a bit disillusioned. Let me tell you why.
First-person view drones are unmanned aerial vehicles with four propellers located at the four corners of the craft, roughly in the shape of a square of seven to 12 inches in length on each side. They are controlled by an operator wearing virtual-reality goggles that receive the image from the drone’s forward-facing camera (hence the name first-person view). The most common types of first-person view drones are single-use: They fly directly into their target, where they detonate an explosive charge of up to 1.5 kilograms. These drones are touted as a cheap and accessible solution that can give troops on the tactical level their own organic precision-strike capability. They can supposedly react quickly and strike moving targets or targets in difficult-to-reach locations, such as bunkers, basements, or inside buildings. Proponents of first-person view drones often repeat the claim that as much as 60 to 70 percent of all battlefield casualties in the Russo-Ukrainian War are now caused by drones. This statistic is probably broadly accurate, though it does not differentiate between casualties caused by first-person view drones and other types of uncrewed aerial systems.
Some authors, including experienced military officers writing in these pages, go even further and claim that first-person view drones will precipitate a revolution in how wars are fought, akin to the introduction of muskets. Among other things, they will make concealment and the massing of troops and equipment in the combat zone nearly impossible. Any concentration of troops or vehicles will supposedly be observed immediately and butchered by swarms of cheap, fast drones. Proponents of drones, especially in Silicon Valley, have claimed that drones might completely replace artillery.
Whether or not we believe these far-reaching claims, we’ve certainly all seen the videos on social media of these drones performing impressive, highly precise attacks. We’ve seen them striking a Russian tank on the move, flying through the open back hatch of an infantry fighting vehicle, or entering a building to surprise the enemy, sometimes literally, with their pants down. But those impressive strikes are rare exceptions. The cases when first-person view drones actually do that are few and far between.
During my time in Ukraine, I collected statistics on the success of our drone operations. I found that 43 percent of our sorties resulted in a hit on the intended target in the sense that the drone was able to successfully fly all the way to the target, identify it correctly, hit it, and the drone’s explosive charge detonated as it was supposed to. This number does not include instances when our higher command requested a sortie but we had to decline because we knew that we could not strike the target for reasons such as weather, technical problems, or electronic interference. If this type of pre-aborted mission is included in the total, the success rate drops to between 20 and 30 percent. On the face of it, this success rate is bad, but that is not the whole story.
I began to notice that the vast majority of our sorties were against targets that had already been struck successfully by a different weapons system, most commonly by a mortar or by a munition dropped by a reusable drone (in other words, not a first-person view drone). Put differently, the goal of the majority of our missions was to deliver the second tap in a double-tap strike against a target that had already been successfully prosecuted by a different weapons system. The proportion of missions when we successfully carried out a task that only a first-person view drone can fulfill — delivering a precision strike on a target that could not be hit by other means — was in the single-digit percent.
There are two reasons why these drones rarely successfully do what they were designed to do. The first has to do with how commanders choose to employ first-person view drones. Presumably, our commanders decided that they had first-person view drones as a capability, so they might as well use them, even if there were other weapons systems that could also do the job. There is a certain logic to this, and the commanders were not paying for the expended drones out of their own pockets. They were more focused on the immediate mission. While first-person view drones are cheap, they are usually not the cheapest option available to commanders. This is the problem with using them in double-tap strikes or for missions that can be achieved by other systems. One of these drone sorties costs about $500 in materiel. A mortar shell costs less than $100. A munition dropped from a reusable drone, usually also something like a modified mortar shell or 40-millimeter grenade, also costs less than $100.
The second reason why these drones rarely do what they were designed to do is technical. They are finicky, unreliable, hard to use, and susceptible to electronic interference. Few first-person view drones have night-vision capability. Those that do are in short supply and cost twice as much as the base model. In Ukraine, in the winter, it’s dark for 14 hours a day. Wind, rain, snow, and fog all mean a drone cannot fly.
A solid quarter of all these drones have some sort of technical fault that prevents them from taking off. This is usually discovered only when they are being prepped for launch. The most common is a fault in the radio receiver that receives inputs from the control panel, or in the video transmitter that transmits the signal to the operator’s virtual-reality goggles. Sometimes this fault can be fixed through a software update in the field. Often, it cannot. Many faulty drones are simply cannibalized for spare parts, because there is no better use for them. Even once a drone is airborne, batteries often die mid-flight. In about 10 percent of sorties, the drone hits the target, but its warhead does not detonate.
Once airborne, operating a first-person view drone successfully is not easy. These drones were originally designed to be toys for rich people. Before they were press-ganged into service as tools of war, they were used either in aerobatic displays or in races where a group of operators would compete in flying through an obstacle course. In either case, the drones were not meant to be easy to fly. They were meant to be highly maneuverable, but also unstable. First-person view drones cannot really hover, fly slowly, or linger above a target. The assumption among hobbyists is that enthusiasts will invest the time and money to become proficient at flying. As a result, training a highly proficient operator can take months. A standard, base-level course for Ukrainian drone pilots takes about five weeks. The quality of operators it prepares is questionable, and graduates of the course need extra on-the-job experience to become truly proficient. Most drone pilots I encountered did not go through this course. Instead, they learned to fly drones on the job. Even experienced operators routinely miss their targets and crash into trees, power lines, or other obstacles.
... continue reading