American Autonomous Vehicles Face Real Combat Tests in Ukraine

Field Testing Under Fire

More than 100 autonomous all-terrain vehicles built by Forterra have been operating in Ukrainian conflict zones since last October, marking what appears to be the largest combat deployment of uncrewed ground vehicles by any U.S. defense technology company. The vehicles have traveled over 2,500 miles across 1,100 missions, transported more than 777,000 pounds of cargo, and evacuated 52 casualties. At DailyTechWire, we've tracked the evolution of military robotics for years, but this deployment represents a watershed moment: the point where venture-backed autonomy meets the chaotic variables of live combat.

The scale and duration of this field test offer something commercial autonomous vehicle developers rarely access: data from environments where failure carries lethal consequences. Scott Sanders, Forterra's chief growth officer and former Marine officer, frames it bluntly. Until systems encounter the realities of combat, he notes, developers are operating on assumptions. The Ukrainian deployment has shattered several of those assumptions while validating others.

Why Ground Autonomy Matters Now

Aerial drones have dominated coverage of the conflict, but their proliferation has created a secondary effect that ground robotics are uniquely positioned to address. Extensive surveillance from above has turned large swaths of terrain into kill zones where any visible movement invites immediate attack. Sergeant Major Corey Wilkens, who leads U.S. Army programs on autonomous vehicles and tactics, describes the operational constraint: soldiers become extremely vulnerable to first-person-view drones, artillery, and munitions the moment they're exposed.

Ukrainian forces have been building their own uncrewed ground vehicles to address logistics and casualty evacuation, but these systems face technical limitations. According to a Ukrainian soldier who has worked with multiple UGV platforms and spoke on condition of anonymity for security reasons, domestically produced vehicles are typically battery-powered with payload capacities around 250 kilograms. Forterra's Lancer vehicles, built on Polaris ATV platforms and equipped with custom sensor and compute stacks, run on gasoline and can carry 750 kilograms. That threefold increase in capacity transforms the range of missions these systems can support.

The Ukrainian soldier was emphatic about the operational value: for logistics and maintaining defensive positions, Forterra's platform has become the most important UGV in Ukraine. The demand for additional units far exceeds current supply.

Adapting Silicon Valley Tech to Wartime Constraints

Initial reception was less enthusiastic. Ukrainian forces have encountered Western contractors before, and early versions of Forterra's system felt over-engineered for U.S. Army requirements rather than optimized for Ukrainian battlefield conditions. The turning point came when Forterra added Starlink satellite internet connectivity, enabling reliable remote operation in areas where terrestrial networks had been destroyed or jammed.

That modification highlights a broader tension in defense technology transfer. Systems designed to meet the specifications of the Pentagon's acquisition process often carry assumptions about infrastructure, logistics chains, and operational doctrine that don't map cleanly onto the resource constraints and tactical realities of a grinding land war. The companies that adapt quickly gain credibility; those that insist on their original design philosophy lose relevance.

Forterra has learned hard lessons about electronic warfare, over-the-air software updates under hostile conditions, and the importance of mechanical reliability when maintenance windows are measured in hours rather than scheduled depot visits. Some vehicles have been lost, particularly when stuck in deep mud where Russian forces can engage them without risk. The company, which has raised over $500 million from investors including XYZ Venture Capital and Moore Strategic Partners, is leveraging these lessons as it competes for larger U.S. national security contracts.

The Autonomy Gap

Despite being built for autonomous operation, the Lancers are primarily being teleoperated by Ukrainian soldiers in combat zones. This represents a significant gap between the technical capability Forterra has demonstrated in controlled environments and what combat conditions allow.

Two factors drive this reliance on remote human control. First, the vehicles are too valuable to risk in fully autonomous mode given current attrition rates. Second, and more fundamentally, the autonomy stack isn't ready for dynamic threat response. While the vehicles can navigate diverse terrain autonomously, they cannot yet identify unexpected enemy forces and react appropriately in real time. The Ukrainian soldier explained the constraint: responding to live enemy threats requires judgment the autonomy doesn't yet possess.

This limitation is not unique to Forterra. It reflects a broader challenge in military robotics: the gap between path-planning algorithms and true tactical decision-making. Forterra has been working on autonomous vehicles for two decades, and the company is now exploring how to combine classical robotics approaches with generative AI models that can handle novel situations in a more generalized way.

Sanders points to a data collection problem that sets military autonomy apart from civilian self-driving development. Many critical tasks have no equivalent in open-source training datasets because humans don't perform them. Navigating minefields and operating weapon systems require training data that must be generated specifically for military contexts. The solution, according to Forterra, involves knowing when to rely on deterministic robotics methods and when to deploy AI where adaptability is essential.

The Competitive Landscape

Other companies are converging on similar technical approaches. Scout AI raised $100 million earlier this year to train foundation models for military applications and develop a suite of autonomous platforms including UGVs. Field AI and Overland AI are conducting trials with the U.S. military, each bringing different architectural choices to the problem of ground autonomy under fire.

The Ukrainian deployment gives Forterra a data advantage that's difficult to replicate in peacetime testing. Scott Philips, Forterra's chief innovation officer, visited Ukrainian operations centers to observe the vehicles in use, an experience that revealed precisely where manual processes still dominate and where automation could relieve pressure on operators working in real time. Those insights, gathered in range of Russian artillery, carry weight that simulated environments cannot match.

Cost as a Tactical Constraint

Ukrainian operators have issued a clear challenge to Forterra: make it cheaper. The Lancers are cost-competitive within their category, benefiting from Polaris' commercial supply chain for the base vehicle. But they remain too expensive to deploy with the same expendability as aerial drones. Attrition is inevitable on this battlefield, and each loss hurts operationally and financially. The Ukrainian soldier was direct: more units are needed, which means lower unit costs are essential.

This demand reflects a fundamental shift in how military robotics must be evaluated. In a high-intensity conflict with sustained attrition, the calculus shifts from performance per dollar to acceptable loss rates. A system that costs twice as much but can only be fielded in half the numbers may be operationally inferior to a cheaper, more numerous alternative, even if individual units are less capable.

That cost pressure will shape the next generation of military UGVs. Companies that can drive down unit economics while maintaining core functionality will win larger orders. Those that optimize for performance without regard to affordability will struggle to scale beyond niche applications.

What Combat Deployment Reveals

Wilkens, the U.S. Army program lead, is convinced that ground autonomy is achievable now, not in some speculative future. The Ukrainian deployment validates that assessment while clarifying the work still required. Autonomous navigation across complex terrain has been solved well enough for operational use. Autonomous threat response and tactical decision-making remain open problems.

The distinction matters because it shapes where defense technology investment should flow. Incremental improvements to path planning and obstacle avoidance deliver diminishing returns at this point. The frontier is in perception systems that can distinguish threats from neutral actors in ambiguous situations, and decision architectures that can execute tactically sound responses without human intervention.

For Forterra and its competitors, the Ukrainian field test has compressed years of development into nine months of operational feedback. The vehicles that emerge from this crucible will be fundamentally different from the systems that entered it, shaped by the unforgiving constraints of real combat rather than the assumptions of peacetime engineering.