Forterra, a US defense tech company with roots going back over 20 years in autonomous vehicle research, confirmed this week that more than 100 of its self-driving cargo vehicles have been operating in Ukrainian conflict zones since October 2024. The vehicles have run over 1,100 missions, carried nearly 800,000 pounds of supplies, and completed 52 evacuations of wounded soldiers.
That sounds impressive. The more useful framing is: what does this actually tell us about where autonomous vehicles are today, and where they are not?
The context for why ground robots matter in this war is important. Aerial drones have turned the Ukrainian front lines into a surveillance grid where any moving person or vehicle can be spotted and targeted within minutes. As a result, the most dangerous job on the front is now basic logistics: carrying ammunition, food, and medical supplies to the people fighting. Robots filling that role means fewer humans exposed to that risk.
Forterra's Lancer vehicle is built on a commercial Polaris ATV chassis. That commercial base is what keeps costs manageable and what makes spare parts easier to source. The vehicle runs on gas rather than battery power, which gives it a longer range and more carrying capacity than most Ukrainian-built alternatives. Ukrainian-built ground robots typically carry up to 250 kilograms; the Lancer carries 750. That difference is operationally significant when you are trying to sustain a frontline unit.
The Ukrainian forces were skeptical at first. Western contractors have brought technology to this war that was engineered for US Army specifications, not for the specific realities of a grinding front. What changed Forterra's standing was a simple modification: adding a Starlink satellite internet antenna so the vehicle could be operated and updated remotely, even in areas where local communications are jammed or destroyed.
Here is the part the company is more careful about: these vehicles are mostly being driven by humans via remote control, not operating on their own. The autonomous navigation works well enough across difficult terrain. The hard problem, reacting to an unexpected enemy soldier or threat in real time, is not solved yet. Soldiers kept human hands on the controls partly because losing a vehicle to bad autonomous decision-making was too costly, and partly because the software genuinely is not ready for that level of judgment.
That gap between "can navigate terrain" and "can respond to combat threats" is the defining challenge for the entire field right now. Forterra is working on combining traditional robotics with newer AI software to close it, but the company's own team acknowledges they need data from scenarios that simply do not exist in any public training set, such as navigating minefields or responding to weapons fire.
The scale of demand coming from Ukraine makes this commercially significant. Ukraine plans to procure 25,000 ground robots in the first half of 2026 alone, aiming to shift all frontline logistics to unmanned systems. In 2025, Ukrainian domestic industry delivered 15,000 ground vehicles to frontline units, up from around 2,000 the year before. The Ukrainian defense minister's stated goal: 100 percent of frontline logistics handled by robots, not people.
Forterra is not alone in this space. European competitor ARX Robotics is scaling to hundreds of deployed vehicles in Ukraine. Startups like Overland AI and Field AI are running trials with the US military domestically. The competition for defense contracts in autonomous ground vehicles is now real and accelerating.
The main feedback from Ukrainian soldiers is not a technical request. It is a price request. Losing a Lancer in deep mud, where Russian forces can target it at leisure, costs too much. The vehicle needs to become cheap enough that losing one is an acceptable outcome, the way losing an aerial drone already is. That cost pressure is the signal that tells you this technology is transitioning from "impressive demonstration" to "operational tool": the people using it in life-and-death situations want more of them, and they want them cheaper.