What are Forterra's Lancer vehicles used for in Ukraine?

They're primarily used for logistics—hauling supplies and ammunition—and for casualty evacuations, removing humans from high-risk missions in drone-saturated combat zones.

Are the vehicles fully autonomous in combat?

Not in active combat situations. Ukrainian soldiers have largely been teleoperating the vehicles in hot zones, as the autonomy stack isn't yet capable of making real-time tactical decisions against enemy forces.

How much has Forterra raised, and who are its backers?

Forterra has raised over $500 million in venture funding, with backers including XYZ Venture Capital and Moore Strategic Partners.

What was the key modification that made the vehicles more effective in Ukraine?

Adding a Starlink satellite internet antenna, which enabled reliable teleoperation and remote software updates in the field.

Home›Robotics & Autonomy›Article

Robotics & Autonomy

American Autonomous ATVs Are Already Fighting in Ukraine—Here's What They've Learned

Forterra's Lancer autonomous ground vehicles have completed over 1,100 missions in Ukraine, hauling hundreds of thousands of pounds of cargo and evacuating casualties. But the real story is what the battlefield revealed about the limits of autonomy.

Dispatch desk·July 8, 2026·4 min read

TL;DR

Forterra's autonomous ATVs have logged 2,500+ combat miles in Ukraine, proving genuine utility—but also exposing exactly where self-driving tech still falls apart when bullets start flying.

What changed

Forterra publicly disclosed that more than 100 of its Lancer autonomous ground vehicles have been operating in Ukrainian conflict zones for nine months, marking what it claims is the largest combat deployment of autonomous ground vehicles by a U.S. defense tech company.

Why it matters

This is the most significant real-world stress test of autonomous ground vehicle technology ever conducted, and the lessons—about electronic warfare, terrain limits, and the gap between demo-day autonomy and battlefield autonomy—will shape military robotics development for years.

Editorial record

Drafted by autopilot. Reviewed when marked.

Let's skip the press release version and talk about what's actually happening here, because it's genuinely interesting—and a lot more complicated than "robots are going to war."

Forterra, a U.S. autonomous vehicle company with two decades of development behind it and north of $500 million in venture backing, has confirmed that more than 100 of its Lancer autonomous ATVs have been operating in Ukrainian conflict zones for the past nine months. Over 1,100 missions. More than 2,500 miles driven. Nearly 800,000 pounds of cargo hauled. Fifty-two casualty evacuations. Those are not demo numbers. That's a real operational footprint.

The Lancers are built on Polaris ATV platforms, layered with Forterra's custom sensor and compute stack, and critically—gas-powered. That last detail matters more than it sounds. Ukrainian-built uncrewed ground vehicles (UGVs) are mostly battery-electric and top out around 250 kilograms of payload. The Lancer carries 750 kilograms. When you're moving ammunition and wounded soldiers under drone surveillance, that capacity difference isn't trivial.

The Drone Problem Nobody Talks About Enough

Here's the tactical context that makes this deployment make sense: aerial drones haven't just changed the battlefield, they've made the surface of it extraordinarily dangerous. FPV drones, munition-dropping quadcopters, artillery-spotting platforms—they've created what one U.S. Army sergeant major described as an environment where "there's nowhere to hide." Soldiers moving supplies, evacuating casualties, or repositioning equipment become immediate targets.

That's the threat model autonomous ground vehicles are actually solving for. Not some sci-fi vision of robot armies—just the grinding, lethal logistics problem of keeping people alive and supplied when every human movement is a potential kill signal from above. If you can remove the human from the resupply run, you've just made an important class of mission survivable again.

What Actually Works (And What Doesn't)

Here's where I'll ask you to pump the brakes on the triumphant narrative, because the operational reality is messier than the headline suggests.

Ukrainian soldiers are mostly teleoperating these vehicles in active combat zones—not running them in full autonomous mode. That's a meaningful distinction. Teleoperation means a human is watching a camera feed and driving remotely. It reduces exposure, yes. But it's not the autonomous system making decisions under fire; it's a human doing that from a safer location.

Why not full autonomy? Because the software isn't there yet for the hardest scenarios. Navigating a known route across open terrain? Manageable. Detecting an unexpected enemy patrol and making an appropriate tactical decision in real time? That's a different problem entirely. As one Ukrainian soldier put it, the vehicle needs to respond to enemy threats live, while directly in front of hostile forces—and the autonomy stack simply doesn't know how to do that yet.

Some vehicles have been lost in combat. Deep mud is apparently as lethal as enemy fire: a stuck vehicle is a stationary target, and Russian forces are perfectly happy to take their time with those.

The Data Problem Is the Real Problem

Forterra has been doing this work for twenty years, which gives them a legitimate edge over the defense-tech-bro startups that discovered autonomy last Tuesday. But even they're running into the fundamental wall that stops every autonomous system: you need training data that matches your operational environment, and the operational environment here is genuinely novel.

Navigating a minefield. Operating weapons systems. Responding to adversarial tactics that no open-source dataset contains. These aren't edge cases—they're the core mission. And you can't just scrape the internet for this data or fine-tune a general-purpose model on it.

Forterra's approach, as described by their chief growth officer and former Marine officer Scott Sanders, is essentially a hybrid: classical robotics algorithms (the kind that gave us self-driving car pipelines) blended with newer generative AI components that can handle more open-ended situational reasoning. In other words, rule-based systems where the rules are knowable, learned systems where they aren't. That's a sensible architecture. Whether it scales to actual combat decision-making is an open question.

The Starlink Detail Nobody Should Skip

Initial Ukrainian reception to the Lancers was lukewarm. Western contractors showing up with tech optimized for U.S. Army specs—not Ukrainian field conditions—is apparently a recurring frustration. The fix that turned sentiment around? Bolting on a Starlink antenna.

Think about what that implies. The limiting factor wasn't the autonomy stack or the payload capacity—it was connectivity. Without reliable communications, you can't teloperate effectively, you can't push software updates, and you can't maintain situational awareness. Starlink solved a problem that made everything else functional. It's a reminder that the supporting infrastructure around autonomous systems often matters as much as the autonomy itself.

What This Means Beyond Ukraine

Forterra, backed by funds including XYZ Venture Capital and Moore Strategic Partners, is now sitting on something extremely valuable: combat-validated data and operational credibility that no competitor has matched. That positions them well for major national security contracts—they've already secured production awards alongside Oshkosh Defense for the U.S. Marine Corps.

The broader lesson here isn't that autonomous ground vehicles are ready to replace soldiers. They aren't. The lesson is that a specific, constrained version of autonomy—logistics support, casualty evacuation, removing humans from high-risk resupply runs—is deployable today and demonstrably useful under fire. That's not hype. That's a real capability gain with real operational evidence behind it.

The gap between "drives autonomously across terrain" and "makes sound tactical decisions against an adaptive adversary" is still enormous. Filling that gap is going to require data that can only come from environments like the one Forterra has been operating in. Which means Ukraine isn't just a humanitarian crisis with a technology subplot—it's the world's most consequential proving ground for autonomous systems, and the lessons being learned there will define this field for the next decade.