This is the part of modern war nobody wants to budget for until it’s too late: the “cheap” thing in the sky that forces the “expensive” thing on the ground to hide, pause, or die. So when I see the U.S. Army evaluating an AMPV variant that bakes in radar drone detection and a real attempt to swat small UAVs, I don’t read it as a cool upgrade. I read it as a quiet admission that the old way of moving armor around is getting punished.
Based on what’s been shared publicly, BAE Systems delivered two AMPV 30 prototypes to the Army’s 1st Cavalry Division for evaluation. Two vehicles sounds small until you remember what this is really for: putting new ideas in the hands of people who will break them fast, in real unit routines, not in a showroom. BAE ties it to its “Transformation in Contact 2.0” push, and the pitch is speed—apparently built in six months, funded by the company, with eight prototypes total, for $17 million.
That pace is the point. It also makes me nervous.
The AMPV 30 is basically an armored personnel carrier that’s drifting back toward “fighting vehicle” behavior. It’s built on a modified Bradley chassis, with an unmanned remote turret (Kongsberg MCT-30), a 30mm cannon (XM813), plus a 7.62mm machine gun. There’s a reason armies love adding firepower: it feels like control. You can solve more problems if you can shoot harder.
But firepower is not the problem drones created. Drones created exposure.
What makes this prototype interesting is that it acknowledges that exposure problem directly. It includes an Echodyne EchoShield portable radar to detect small UAVs. It also adds an experimental programmable 30mm round (Mk 310 PABM) into the ammo mix, clearly trying to give crews a way to engage drones without needing a whole separate air-defense vehicle.
From our side of the world—the people who build drone detection radar systems and AI fusion that blends data from multiple sensors—the radar part is the real hinge. You can have the best gun in the world, but if you only see the drone when it’s already overhead, you’re not “countering” anything. You’re reacting. And reaction is expensive. It burns ammo, it burns time, and it burns confidence.
Here’s the scenario I keep coming back to. Imagine a mechanized unit moving through a town edge. A small drone is up, low and slow, and it’s not even trying to be stealthy. It’s just watching. If the vehicle crew doesn’t detect it early, the drone doesn’t need to attack. It just needs to report. Now the unit is being ranged. Maybe the next thing isn’t a drone at all—maybe it’s artillery, maybe it’s a loitering munition, maybe it’s a team with guided weapons. The “kill chain” starts with seeing.
So yes, putting radar drone detection on the vehicle is a strong move. But a radar bolted on is not the same as a working counter-drone system. The uncomfortable truth is that detection alone can become a new kind of noise. If you alert crews constantly, and half the alerts are birds, hobby drones, or friendly systems, people learn the wrong habit fast: ignore it. That’s why we obsess over fusion—making radar, electro-optical, acoustic, and other inputs agree (or disagree) in a way a human can act on in seconds, not minutes.
This is where I think the rush-to-field approach can go right or wrong.
It can go right if the Army uses these trials to learn the ugly details: what alerts actually matter, how often drones show up, what ranges are useful, what happens when multiple vehicles all “see” the same target, and how to avoid turning a crew into full-time screen watchers. It can go wrong if everyone treats the radar as a checkbox and the turret as the headline. Because then the unit ends up with a vehicle that shoots well, looks modern, and still gets surprised from above.
BAE says the vehicle carries six dismounts plus two crew, and mentions hull protection at STANAG 4569A Level 2 with potential growth to Level 6. I’m not going to pretend those levels mean something simple to a reader’s daily life. What matters is the trade: weight, space, power, and attention are always limited. Add a turret, add radar, add new ammo, add screens—and something gives. Maybe it’s maintenance time. Maybe it’s training. Maybe it’s reliability. Maybe it’s just crew fatigue, the most underrated failure mode in any “more capability” story.
There’s also a strategic tension here that people will argue about. Some will say: stop trying to make every platform do everything. Keep APCs simple, and bring dedicated air defense assets. Others will say: dedicated assets won’t always be there, and drones show up everywhere, so every vehicle needs at least a fighting chance. I lean toward the second view, with a warning label: if “every vehicle has counter-drone” turns into “every vehicle has an alarm,” we’re going to waste money and still lose time.
The interesting detail is that BAE built eight prototypes with company funds. That’s gutsy, and it signals confidence. It also raises a question about incentives. If industry proves it can move fast with private money, the Army may start expecting that pace everywhere. That’s not automatically good. Speed is how you get learning. Speed is also how you ship half-finished systems that look complete in photos and feel incomplete in the field.
What I want to see from these trials isn’t just whether the AMPV 30 can hit a drone with a programmable round. I want to see whether crews can detect, classify, decide, and act without being overloaded—and whether the sensor data is fused in a way that’s trustworthy under stress, not just impressive during a demo.
If we’re honest, the biggest risk isn’t that this concept fails; it’s that it “sort of works” and everyone declares victory while drones keep getting cheaper and more common.
So here’s the real debate: should the Army push hard to put radar drone detection and fused sensors onto more frontline vehicles now, even if it means accepting messy first versions, or should it slow down and build fewer but more mature counter-drone packages first?