← The Signal Report Work with me

Report 014 · Defense Tech

How microwaves kill a jam-proof drone

An earlier report left a hard question open: if a fiber-optic drone is off the radio spectrum, is there any electromagnetic answer left, or does the whole fight go kinetic? A December demonstration says there is one, and it is not jamming.

I've written twice now about the fiber-optic drone, the little FPV aircraft that trails a spool of hair-thin glass fiber back to its operator and sends video and commands down that fiber instead of over the air. No radio link means nothing for a jammer to jam. As a former Counter-IED and electronic-warfare officer, the guy whose job was to deny the enemy the spectrum, I've been blunt that this thing is built to make my old job irrelevant. The last report ended on the defender's dilemma: once you can't attack the link, you're pushed into sound, light, and physical mass, and the fight gets expensive because you have to hit each aircraft one at a time.

That left a question I didn't fully answer. Is the electromagnetic spectrum simply out of the fight against a fiber drone? In January, Epirus released video of its Leonidas high-power microwave system disabling a fiber-optic guided drone in a December 2025 live-fire demonstration at a U.S. government test range. Their own language is worth reading carefully, because it is the whole point: the event, they say, "marks the first known instance of electromagnetic interference being weaponized to defeat a fiber-optic guided drone." So the spectrum is not out of the fight. But to see why that isn't a contradiction, you have to be precise about what a jammer attacks versus what a microwave attacks.

Jamming hits the link. A microwave hits the hardware.

A jammer is a communications weapon. It wins by drowning or spoofing the radio signal a drone uses to talk to its pilot: flood the frequency with noise, or feed it a convincing lie, and the link breaks. Everything a jammer does is aimed at that link. Take the link off the air, run it down a glass fiber instead, and the jammer has nothing to point at. That is exactly why Epirus describes fiber drones as operating "without an RF command-and-control link, rendering them immune to jamming, spoofing and other legacy EW counter-UAS measures." That immunity is real, and I don't want to undersell it. Against the link, fiber wins.

A high-power microwave is a different kind of weapon entirely. It does not care about the link. It dumps a burst of electromagnetic energy directly into the target and induces currents inside whatever electronics are there, upsetting or frying them. Epirus says Leonidas "delivers precise, software-defined weaponized electromagnetic interference to induce full kill within critical onboard electronics." Read that again: onboard electronics, not the control link. And here is the thing the "unjammable" headlines quietly skip. A drone is a flying circuit board no matter how it is steered. It has a flight controller, motor drivers, sensors, a camera. Cutting the cord to the pilot doesn't remove any of that. The microwave goes after the parts the fiber trick never protected.

"Off the spectrum" was always imprecise

When people say a fiber drone is "off the electromagnetic spectrum," what actually left the spectrum is the control link. The drone did not. It still flies on electronics that obey the same physics every electronic device obeys, which is that a strong enough field in the wrong place scrambles them. That is the seam directed energy exploits. The U.S. Army's own Center for Army Lessons Learned note on fiber drones, from August 2025, is a good measure of how new this thinking is: it walks through detecting and defeating these drones by sound, by the heat and glint of the cable, by radar, by cutting the fiber, and it does not mention high-power microwave at all. The counter to a fiber drone that stays in the electromagnetic domain simply wasn't on the board yet. That's what the December demonstration is claiming to add.

What the demo does and does not prove

Now the honest part, because this is a manufacturer showing its own weapon working on a range, and I read those the same skeptical way I read a supplement label. What's demonstrated is a genuine capability: at a controlled site, in a live-fire event, a microwave system killed a fiber drone's electronics. That is real and it is new. What is not demonstrated is that the problem is solved. A single video at a test range is a proof of concept, not a fielded, everywhere answer.

And directed energy carries its own tradeoffs that matter on a real perimeter. High-power microwave is power-hungry and comparatively short-legged. Leonidas is a vehicle-mounted system for a reason: it takes serious electrical power and a large phased-array antenna to put a damaging field on a target at useful range. That makes it a point-defense bubble around something you're protecting, not the wide-area, cover-a-whole-sector shield that made RF jamming so cheap and scalable. Timing matters too. Frying the electronics only helps if it happens before the drone completes its terminal dive into the target. And electronics can be shielded and hardened, so this is a measure-countermeasure race like every other fight in the spectrum, not a permanent win.

The upside is just as real, though, and it's the reason this class of weapon is worth taking seriously. Because a microwave attacks the airframe and not a specific signal, it can in principle service many drones in its beam rather than one at a time, which is the exact problem swarms create for one-shot interceptors. Epirus notes the system "utilizes non-ionizing radiation, making it inherently safe for humans" when used as intended, and steers a directional beam onto the target area. A non-kinetic effect that scales against volume is precisely what the kinetic-only endgame of my last report was missing.

The signal

The fiber drone didn't leave the electromagnetic spectrum. Only its control link did. Jamming failed against it because jamming was aimed at the link, and the link was gone. A high-power microwave works against it because it is aimed at the hardware, and the hardware never left. That is the distinction the word "unjammable" flattens: immune to jamming is not the same as immune to the spectrum. This doesn't retire sound, light, and mass from the last report; it adds an electromagnetic layer back on top of them. The lesson from the EW side of the house holds, as it has for twenty years. There is no silver bullet. When the enemy takes away your best tool, you don't quit the domain. You find the seam they left open, and you layer the rest.

Sources

  1. Epirus, Inc., "Epirus' Leonidas Demonstrates Successful Use of High-Power Microwave to Defeat Fiber-Optic Controlled UAS," press release, 13 January 2026. (Primary, and a manufacturer's own claim: Leonidas VehicleKit HPM disabling a fiber-optic FPV in a December 2025 government-range demo; "first known instance," "full kill within critical onboard electronics," "immune to jamming, spoofing," "non-ionizing radiation.")
  2. The Defense Post, "Epirus Demos First Directed-Energy Takedown of Jam-Proof Fiber-Optic Drone," 15 January 2026. (Independent coverage of the demonstration and the operational gap fiber drones opened.)
  3. Army Recognition, "U.S. Demonstrates Microwave Weapon Defeating Fiber-Optic FPV Drones," January 2026. (Independent technical framing: HPM targets the drone's onboard electronics rather than the control channel the fiber removes.)
  4. Alex Braszko, "Fiber Optic Drones: Posing a Significant C-UAS Challenge," Center for Army Lessons Learned, U.S. Army, 12 August 2025. (The counter-fiber-drone playbook as of mid-2025: acoustic, IR-visible cable, radar, and cable-cutting, with no mention of directed energy.)
  5. The Signal Report, "Stopping a Drone You Can't Jam," Report 007, and "'Unjammable' Is a Marketing Word," Report 002. (The defender's dilemma this piece answers, and the attacker's-side companion.)
Onur Oncer
Onur Oncer

U.S. Army combat veteran (Counter-IED / Electronic Warfare), peer-reviewed researcher in microwave spectroscopy, and founder & CEO of Shroombiosis. Consults on laboratory operations, AI, and supplement formulation.

← All reports