Two Russian soldiers, hands raised, walking out of a treeline toward their captors. That part of the picture is as old as war. What is new is that there were no captors. Not in any shape their grandfathers would recognize. The men were surrendering to a tracked machine the size of a pallet and to the quadcopters circling above it, all piloted by an operator several kilometers away, drinking something warm, staring at a screen.
President Volodymyr Zelensky released the footage on April 14, 2026, though the skirmish itself happened last summer. "For the first time in the history of this war, an enemy position was taken exclusively by unmanned platforms," he said. Mykola Zinkevych, who commands the NC13 unit inside the 3rd Separate Assault Brigade, put it in simpler terms. The position was taken without a single shot being fired.
It is tempting to treat that as a headline and move on. A neat, cinematic moment. The problem is that the moment is not a one-off anymore, and the numbers behind it are the part that should make any military planner outside Kyiv sit down.
The scale Ukraine is already running
In March 2026 alone, Ukrainian units logged more than 9,000 combat and logistical missions using ground robots. In November 2025 that same figure was 2,900. The first quarter of 2026 added up to roughly 24,500 missions. That is not a curve; that is a wall. Defense One's correspondent, who has covered the drone industry for nearly a decade, called it the steepest adoption curve for a new weapons category he has ever seen.
The production side matches the usage side. Ukraine built about 2,000 unmanned ground vehicles in 2024. By the end of 2025, it had pushed 15,000 into frontline units, a 650 percent jump in twelve months. On April 18, Defense Minister Mykhailo Fedorov announced the next target: 25,000 robotic ground systems contracted in the first half of 2026. Roughly double the entire 2025 figure, delivered in six months. Fedorov's stated goal is blunt. One hundred percent of frontline logistics, eventually, will be performed by machines.
Ninety-nine percent of those platforms are domestic. Something like 270 Ukrainian companies are building them. The Brave1 defense marketplace, which is effectively an Amazon for battlefield kit that soldiers order from using unit credits, lists over 600 manufacturers and 2,600 product cards. Fifty-five Ukrainian ground vehicles have been codified to NATO standards. Four years ago, most of the people now running these companies were students, software engineers, or infantry soldiers who came home on leave and decided they could weld something better.
I want to be careful here because the temptation with a story like this is to reach for the Terminator image and run with it. The reality on the ground is messier and more interesting.
Most of what these robots do is not combat. Forty-seven percent of missions are logistics and casualty evacuation. Twenty-five percent are mining and demining. Only about 12 percent involve direct combat. In the 3rd Assault Brigade, UGVs now handle 80 percent of logistics operations. In Pokrovsk and Myrnohrad, two of the most contested cities on the line, the figure is 90 percent. Supplies do not move by truck there anymore. They move by TerMIT, the tracked 300-kilogram workhorse built by Tencore, or by Ratel H, a heavier 600-kilogram cousin used for group casualty evacuation, or by Zmiy Droid, which can pull two or three wounded men out of a position at once and has an armored modular cabin.
The reason this matters is geometric. The kill zone, the area behind the contact line where any movement is presumed dead, is now 15 to 20 kilometers deep. A Ukrainian UGV developer named Vadym Poritskyi thinks it will stretch to 50 kilometers by the end of this year. If he is right, then human-driven resupply becomes physically impossible over most of the front. Sending an armored personnel carrier to pull out a wounded man is, in the words of an unmanned systems commander in the 12th Azov Brigade, "hopeless. You will be engaged on the way to that position, 100 percent."
So instead, you send a 400 kilogram machine carrying a litter and a battery. If a Russian FPV finds it and turns it into scrap, that is a bad afternoon and a $15,000 loss. If a Russian FPV finds the armored vehicle, that is four families getting a phone call.
Where the machines actually stop working
Here is the part that the viral clips leave out. A ground robot is harder to automate than an aerial drone, not easier. A quadcopter flying between two waypoints has nothing to navigate around but air. A tracked platform has mud, tree roots, shell craters, collapsed concrete, fiber optic cable snagging its axles, and the fact that GPS can be jammed down to a few dozen meters of useful range by whoever is shouting loudest on the spectrum that day.
Because of that, a UGV operator almost never runs a robot the way a drone pilot runs a quadcopter. They run it with constant attention, with an FPV drone pilot overhead coordinating every turn, swapping intel on which treeline is hot. John Hardie of the Foundation for Defense of Democracies put it plainly earlier this year: autonomous navigation on the ground is still not reliable enough to trust. Most Ukrainian UGVs last about 24 hours in the field before the battery dies or Russian fire finds them. A machine that captures a trench still needs human soldiers to hold the ground it took. A robot with a jammed machine gun is just a metal box waiting for someone to walk up and fix it.
A single UGV reportedly held a position for 45 days, kept alive by a DevDroid machine gun turret and a battery change every two days. That story is astonishing, and it is also an outlier. It is the kind of success that survives because someone wrote it down. Plenty of robots died quietly in craters and nobody filmed them.
And the economics are less one-sided than they look. A ground platform runs $30,000 to $40,000, depending on the model. That is 30 to 50 FPV drones for the same money. The machine lasts longer per flight hour, carries more weight, and offers a stable firing platform that no quadcopter can match. But it is also slower, louder, and harder to hide, especially when you bolt a .50 caliber Browning on top. Every advantage has its price tag.
Russia is doing some of this too, though not the same way. Russian industry has focused overwhelmingly on one way attack drones, the Shahed loitering munitions, and the Lancets from ZALA, and far less on ground platforms for taking positions. The Courier, an electric UGV that carries 250 kilograms and does some electronic warfare, is in the field. So is the Kuryer, which can mount a flamethrower and run for about five hours. Uralvagonzavod is still testing an unmanned version of the T-72. But there is no Russian equivalent of Brave1, no open marketplace where a platoon commander in Pokrovsk can compare payloads and order a Murakha from an engineering bureau in Dnipro by Friday. The asymmetry is not in the hardware. It is in the supply chain feeding the hardware.
This is the part that I find most interesting and most under-discussed. Ukraine has not built a better tank. Ukraine has built a better procurement system for a kind of weapon that did not exist industrially five years ago. The Defense Procurement Agency signs contracts with longer timelines so factories can plan. The Brave1 Market lets a sergeant with unit credits order directly from a manufacturer, skipping the usual year of paperwork. Victory Robots, run by the Dignitas nonprofit, funds training centers and mobile workshops that repair UGVs within a few kilometers of where they broke. It is an ecosystem in the literal ecological sense, a set of organisms adapted to each other and to the conditions they grew up in. Nothing about it would work in peacetime. Nothing about it is copy-pasteable into the Pentagon, which canceled its own $3 million Robotic Combat Vehicle program and is starting over.
The reason I want to push back on the usual framing, which says robots are the future of warfare, is that this framing lets the rest of us off the hook too easily. Robots are the present of this particular war, fought under these particular conditions, by a country that has no choice but to invent a way to fight with fewer humans because it does not have the humans to spare. "We will never have more personnel," Zinkevych said. "We will never have a numerical advantage over the enemy. So we need to achieve this advantage through technology." That is not a prophecy. That is a population chart with a gun in its hand.
When Robert Tollast at the Royal United Services Institute was asked about the surrender footage, he gave what I think is the most honest read available. Ground drones will struggle to actually hold territory. Using them without infantry is a bit like using tanks without infantry, which every army learned the hard way in the 20th century and keeps having to relearn. But they are saving lives in casualty evacuation, mine clearance, dangerous resupply, and increasingly in fighting. Both things are true. Neither cancels the other out.
Ukraine's General Staff claims robotic platforms have reduced personnel casualties by up to 30 percent. I cannot independently verify that number, and a 30 percent reduction in casualties in a war of this intensity would be one of the most significant operational shifts of the decade. If it is even half true, it explains why Zelensky's 2026 target is to replace a third of Ukrainian infantry positions with machines. Not to make a science fiction point. Because a third of the bodies currently standing in those positions would otherwise come home in pieces, or not at all.
There is a moment in the Zinkevych footage, which I rewatched twice before writing this, where one of the Russian soldiers looks up at the drone circling above him. He is trying to figure out who he is talking to. Whether there is anyone to talk to. The camera does not answer him. It just hovers, patient, and waits for him to walk in the direction the operator miles away has already chosen.
That image is not the future of war. It is last July.