On paper, “loyal wingman” drones sound like the cleanest deal the Air Force could ever make: more mass, more options, less risk to pilots. But the part that should make everyone sit up is simpler—and more uncomfortable. The U.S. Air Force just brought two prototype combat drones onto an operational test base and started treating them like real teammates before the designs are even locked. That’s either smart realism, or it’s how you accidentally build your whole future around assumptions that won’t survive contact with a serious enemy.
Based on what’s been shared publicly, the Air Force received two Collaborative Combat Aircraft prototypes at Creech: the YFQ-44A Fury and the YFQ-42A Vengeance. The “Y” matters. These aren’t adopted, finished systems. They’re still experimental. And the Air Force is intentionally pulling operators into the process now—so tactics, maintenance routines, arming procedures, and command-and-control get shaped while the aircraft are still flexible.
As a company that builds drone detection radar systems and AI fusion across sensors, we read that and immediately think: good. Not because we’re cheering for one airframe or another, but because the hard part of this program isn’t going to be a glossy spec sheet. It’s going to be the messy middle—how you find things, track things, decide things, and keep doing all of that when the enemy is actively trying to blind you.
The two prototypes reflect different bets. Fury has public performance details: around 2300 kg, near Mach 0.95, up to 9g, and a reported live-fire with an AMRAAM. Vengeance is described as more stealth-focused, with specs kept classified; public estimates put it heavier and slower. That tradeoff is old as air combat: speed and maneuvering versus signatures and survivability. What’s new is that these drones are supposed to fight as part of a mixed team, not as solo acts. That changes what “good” looks like.
Here’s the uncomfortable truth: the drone that’s hardest to see isn’t automatically the drone that’s easiest to use. A stealthier platform can push you into tighter communications windows, stricter routing, and more fragile coordination. And the more you depend on coordination, the more your whole formation depends on sensing and fusion that doesn’t fall apart under stress.
This is where radar drone detection stops being a niche topic and becomes a make-or-break operational question. If you can’t reliably detect small, fast, low-signature drones in clutter—over desert terrain, near friendly aircraft, around decoys—then “loyal wingman” becomes “loyal liability.” And yes, that cuts both ways: we need to detect adversary drones, but we also need to avoid mis-tagging our own, deconflict airspace, and keep the picture coherent when the environment is loud, jammed, and deceptive.
The Air Force is also tying this to dispersed operations—Agile Combat Employment—meaning refuel and rearm on minimally prepared sites with small teams and reduced logistics. That’s the right direction if you believe air bases will be targeted early and often. But it raises a question most people skip: what happens to your sensing and control stack when you’re operating off a rough site with limited gear, fewer trained hands, and a clock that’s always counting down?
Imagine a small team trying to turn jets and drones around quickly at a remote strip. If the drone comes back with battle damage, if its navigation is degraded, if comms are spotty, somebody needs to decide whether it flies again, how it gets armed, and what it’s allowed to do in the next sortie. Those decisions don’t get safer just because there’s no pilot onboard. In some ways they get riskier, because automation invites over-trust when people are tired and the tempo is high.
We like that testing is being consolidated “under one roof” for side-by-side comparisons. That’s practical. But it also creates a trap: you can end up optimizing for the test environment. In the real world, you won’t have clean range conditions, predictable threat sets, and a stable logistics chain. You’ll have missing parts, degraded sensors, and enemy countermeasures designed specifically to break your confidence in what you’re seeing.
And that’s why we keep circling back to production and sustainment. The public framing admits it: success won’t come only from performance numbers. It depends on industrial capacity—airframes, engines, avionics, munitions—at scale. Our view is blunt: a small number of exquisite drones won’t shift the balance. They’ll become high-demand assets rationed by planners, and the whole concept of “take risk off pilots” will quietly shrink back to “use them only when it’s safe enough.” The winner is the side that can field enough systems, keep them supplied, and keep their sensing and fusion resilient when losses happen.
There’s also a debate we don’t dodge: some people will argue this is the beginning of replacing pilots, and that it’s morally and strategically dangerous. We don’t think the near-term reality is replacement. The near-term reality is overload. One crew coordinating with multiple unmanned aircraft in a contested environment is a cognitive and technical burden. If the interfaces, the sensor fusion, and the drone behaviors aren’t rock solid, you don’t get “teaming.” You get distraction at the worst possible time.
So yes, we’re encouraged by the Air Force dragging operators into the prototype phase. That’s how you discover what breaks. But we’re also wary of the storyline that unmanned teammates automatically make airpower safer and simpler. In combat, they can just as easily make it faster, more chaotic, and more dependent on systems that have to work when everything else is failing.
If you had to bet today, do you think these loyal wingman drones will end up reducing risk overall, or will they mostly shift risk into sensing, control, and logistics in ways we’re still underestimating?