On paper, arming loyal wingman drones sounds like clean, modern deterrence. In real life, it’s a fast way to flood the sky with weapons platforms that are easier to build than they are to control, track, and trust. That gap—between “we can bolt it on” and “we can use it safely at scale”—is where people like us live.
From what’s been shared publicly, Anduril just showed air-to-ground weapons integration on its YFQ-44A Fury under the Collaborative Combat Aircraft idea. The munitions shown were familiar ones: Hydra 70 rockets, JDAM guided bombs, and the GBU-39 Small Diameter Bomb. The U.S. plan, as described, is to procure at least 150 of these drones, which is not a boutique experiment. That’s a fleet.
But there’s a detail that should slow everyone down: the weapons shown haven’t yet flown in tests on this platform.
That matters more than the marketing momentum suggests. Integration isn’t just hanging a bomb on a hardpoint and calling it done. The ugly truth is the last 10%—safe release, separation, timing, targeting, and making sure the drone still flies like it should—is where programs stumble. If the munitions haven’t flown, then we don’t yet know how “real” the strike promise is, or what corners will get cut to hit a schedule once procurement pressure kicks in.
And the pressure will kick in, because “150+” changes behavior. When a force thinks it’s about to get a lot of something, it starts planning around it. Training plans, basing concepts, tactics, expectations—those harden early. It becomes psychologically hard to admit later that the weapons piece is still messy or limited. That’s how you end up with a fleet that exists on slides and parades, but struggles in the specific, high-stakes moments it was bought for.
We also saw talk of future integrations: Anduril’s Barracuda 500 cruise missile, more air-to-air missiles beyond the AIM-120, and targeting pods to tighten “sensor-to-shooter.” And the production FQ-44 is expected to have four hardpoints instead of two on the prototype. More hardpoints sounds like more options, but it also multiplies the detection and control problem. Every extra payload type brings different flight profiles, different emission patterns, different mission planning habits, and different ways a drone can be misused—or misunderstood.
Here’s where our perspective gets blunt: the real arms race is not only “who can field the most unmanned strike jets.” It’s who can see them, sort them, and decide what to do about them faster than the other side. If you can’t do radar drone detection reliably—especially in clutter, at odd angles, in mixed groups of crewed and uncrewed aircraft—then the rest is just wishful thinking with explosives attached.
Imagine a defensive unit staring at a messy air picture: birds, decoys, small drones, a few larger drones, maybe a crewed aircraft nearby. Now add armed CCAs into that same space. If we classify too aggressively, we risk engaging the wrong thing and escalating for no reason. If we classify too cautiously, we let a strike package slip through because it looked “low threat” until it wasn’t. That’s not a theoretical tradeoff. That’s the kind of decision that gets made in minutes, by tired people, with incomplete information, while somebody is jamming and spoofing.
This is why we keep pushing sensor fusion that doesn’t fall apart when the environment stops being polite. Radar alone is powerful, but radar alone is not a full story. Passive sensors alone are quiet, but they can be fooled. What wins is combining what each sensor is good at, then making the output usable by humans who are trying to avoid a catastrophic mistake. If armed drones become common, the penalty for a wrong label goes way up.
There’s another angle people don’t love to talk about: armed drones push commanders toward “cheap” strikes. Not cheap in money—cheap in political risk and personnel risk. If you can send an unmanned aircraft to take a shot, it becomes easier to justify taking that shot. That might save lives on one side. It might also make violence more frequent because the human cost feels distant. The consequences aren’t abstract. They show up as faster escalation, more cross-border “messages,” more tit-for-tat attacks, and less time for diplomacy to catch up.
To be fair, there’s a strong case for this path. If a peer fight ever happens, a force that can put weapons on many platforms, spread out risk, and keep pilots farther from danger has an advantage. And it’s not wrong to try to get there. What I don’t buy is the casual assumption that scaling armed drones is mainly an industrial task. It’s a detection, identification, and command task. It’s an airspace management task. It’s a “how do we prevent a bad day from becoming a tragic week” task.
And yes, we should be honest about our own stake: the more the sky fills with uncrewed platforms carrying real munitions, the more nations will invest in the ability to detect, track, and interpret them. That’s our world. But that doesn’t mean we should cheer every integration demo. If anything, we should be the ones saying, “Prove it in flight, prove it under stress, and prove you can keep it accountable.”
If the U.S. really moves toward 150+ armed CCAs, what standard should we demand for radar drone detection and multi-sensor identification before those weapons are treated as operationally routine?