This is the part that should make everyone a little uneasy: the Pentagon is trying to buy strike drones the way people buy consumer tech — fast trials, quick winners, big orders — and it might actually work. But if it works, the pace of change is going to outrun the parts of the system that keep people safe, including how we detect and stop these things when they come back the other way.
Based on public reporting, the US just kicked off Phase III of a program called Drone Dominance. The point is simple: rapidly buy low-cost strike drones, mostly expendable one-way attack drones, and do it in tight competitive cycles instead of multi‑year development programs. There are two tracks: Deep Strike (long-range) and Close-Quarters Battle (short-range use, including urban and underground environments). Vendors aren’t just allowed to show a flying prototype and call it progress. They have to complete flight tests, prove warhead or payload integration, and then sprint into production to deliver a batch for Gauntlet III testing. Submissions run until Oct 12, 2026.
The details matter because they reveal the mindset. In Phase II this summer, dozens of manufacturers took part, and 19 made it to Gauntlet II. Those systems were tested out to 15 km, indoors, at night, and under electronic warfare and counter‑drone pressure. To even qualify for final trials, each participant reportedly had to deliver about 120 armed drones in roughly five weeks. Then comes the big carrot: Phase III could place orders totaling up to 60,000 drones among top performers, on top of an additional 30,000 ordered after Phase I. There’s also a separate effort for reusable “dropper” drones.
From where we sit — as a company that builds drone detection radar systems and AI fusion across sensors — this is both rational and dangerous.
Rational, because the old procurement model is too slow for a battlefield where drones are disposable and tactics evolve weekly. If you can’t test in near-combat conditions, you’re basically buying PowerPoint. And if you can’t scale production quickly, you’re buying a boutique weapon in a world that rewards mass.
Dangerous, because the same “move fast” logic doesn’t automatically carry over to defense against drones. Strike drones are getting cheaper, more plentiful, and more adaptable. Detection and protection have to keep up, and that is not just a hardware problem. It’s a “what do you believe is in the air right now, and how confident are you” problem. That’s where sensor fusion earns its keep. A single sensor can be fooled. Multiple sensors, fused well, are harder to trick — but only if you treat data quality, integration, and testing as seriously as you treat the drone airframe.
Imagine a base that suddenly has to plan for waves of one-way drones, not one or two. The question isn’t “Can we shoot one down?” It’s “Can we see the second and third wave early enough, with enough confidence, while everything is noisy and jammed?” Radar drone detection is a core piece of that, especially when you can’t rely on clean radio signals or perfect visuals. But radar alone isn’t magic. Put it next to other sensors and fuse it properly, and you can start making decisions faster, with fewer false alarms. Without that, you get the worst outcome: defenders either panic and overreact, or stop trusting alerts entirely.
Now zoom in to the Close-Quarters track: urban blocks, tunnels, underground spaces. That should set off alarms for anyone who thinks counter-drone is just an outdoor perimeter problem. In tight spaces, you have clutter, reflections, and weird flight paths. Visual cameras can be blind at night or in smoke. Radio-frequency cues can vanish or be spoofed. This is exactly where a multi-sensor approach can make the difference between “We think something is there” and “We know something is there, here’s where it’s going.”
There’s another tension people don’t like to say out loud: the Pentagon’s approach is going to reshape the drone market. If you dangle potential orders up to 60,000 units, you invite a whole ecosystem to optimize for passing the Gauntlet. That’s not evil. It’s just how incentives work. Companies will tune designs to the tests, and the tests will try to anticipate that. It becomes an arms race between evaluation and adaptation. The real worry is what happens when the next opponent watches the same playbook and copies the mass-drone model with their own constraints removed.
And yes, there’s a strong counterpoint: mass-buying “cheap” drones could save lives by giving troops better tools and reducing riskier manned missions. That’s real. It’s also why this program is likely to expand, not shrink.
But if offense scales faster than defense, the bill comes due somewhere else. It comes due when airbases have to pause operations because they can’t separate real threats from noise. It comes due when a convoy can’t tell the difference between a bird, a hobby drone, and an incoming one-way system until it’s too late. It comes due when civilian infrastructure becomes a tempting target because the tools are abundant and hard to trace.
What we don’t know yet is whether the same “Gauntlet” discipline will be applied to the full kill chain on the defensive side — detection, tracking, identification, decision, and response — under the same ugly conditions: jamming, darkness, indoor spaces, and swarms.
If the US is about to normalize buying tens of thousands of expendable strike drones at speed, are we putting the same urgency into proving that our detection and protection systems can keep up when the sky stops being predictable?