Microwave Shields Go Big in 2025: Counter-Drone Hits Scale

The moment the swarm fell

On a clear Indiana morning in late August, a stack of quadcopters lifted together like a flock of startled starlings. Seconds later they dropped, almost in unison, into a safe field at Camp Atterbury. Epirus’ Leonidas, a high power microwave system, had pulsed a curtain of radio frequency energy that scrambled the brains of the drones. Observers from multiple U.S. services and allied militaries watched a 49 drone swarm go down at once, part of a series of runs that tallied 61 drones disabled across the event. Axios, the only outlet on site, reported that the system zapped a 49 drone swarm in seconds.

That spectacle matters because it closes the gap between lab promise and field reality. Over the past five years, directed energy has been a parade of prototypes. In September 2025, it looked like a product. The headline number is not just 49 drones at once. It is that this class of systems can be queued, aimed, pulsed, and made safe in front of commanders and foreign buyers who have to decide what they will put on a base gate or atop a ship.

From demo to delivery

The other half of the story arrived in July. The U.S. Army’s Rapid Capabilities and Critical Technologies Office awarded Epirus a 43.6 million dollar contract for two Generation II Indirect Fire Protection Capability High Power Microwave systems, with tests and spares. This is a follow on to four first generation systems that the company delivered in 2024. The July buy is small as buys go, but it signals a transition from one off experiments to iterative fielding. The Army’s intent is to plug HPM into the same problem set as rockets, artillery, mortars, drones, and subsonic cruise missiles. That means HPM is now being treated as a defensive effector in a layered architecture, not a science project. The contract details are public; you can read the Army’s July award announcement for the GEN II systems.

Put those two threads together and you get the real news of September: high power microwave defenses are hitting scale in 2025. Not mass production yet, but the point where commanders begin to plan around them.

What HPM actually does, in plain language

If lasers are scalpels that burn a target at a point, high power microwaves are a thunderclap that stuns everything in a cone. The weapon pushes a high energy radio wave into the air. That energy couples into the wiring and circuit boards of small drones. Think of drone electronics as tiny antennas even when they are not meant to be. The pulse leaves them confused or permanently damaged. The effect is instantaneous, line of sight, and it does not rely on precise tracking the way a laser does. It is also non explosive, which matters near fuel, ammunition, or civilian areas.

The tradeoffs are different too. Lasers put a lot of energy into a pinpoint and are great against a single drone or the seeker of an incoming munition if you can hold the beam on target. Microwaves flood an area and are best against many small airborne computers at once. The limitation is that HPM is a near field fight. At long range, energy spreads and the effect weakens. That is why you are seeing HPM advertised as part of the inner layer of a defense, the place where a base or a ship needs to stop a dozen or a hundred cheap threats without spending a million dollars on interceptors. For a look at how lasers complement HPM, see Iron Beam laser service.

Where HPM fits in layered air defense

Layered defense is a relay race. Radars and passive sensors detect and classify. Command and control assigns effectors. Kinetic interceptors handle the few big threats that get through the outer layers. Lasers and guns deal with what leaks in closer. HPM is the sweeper for the many small things that slip by, especially when they arrive together. For context on the sensor and data layer overhead, see SDA Tranche 1 LEO network.

For fixed sites and bases: HPM sits with short range radars, passive radio frequency detectors, and jammers around the perimeter. The appeal is its unlimited magazine. If the power plant can keep up, HPM can fire again and again, saving missiles for the one target that truly needs one. Expect to see HPM racks co sited with counter rocket radars and tied into existing command and control software. This is the zone where generators and thermal plans matter most, because base defenses must run all day, not just for a two hour demo.
For ships: A maritime variant matters for two reasons. First, ships already ride with megawatts of power and chilled water, which makes the engineering more approachable. Second, navies now worry about small unmanned surface vessels and low flying drones that can saturate close in weapon systems. HPM offers a non fragmentary way to clear the air and water within a few kilometers of the hull. The hard work is integration with combat systems so the microwave pulse does not blind your own sensors or corrupt friendly radios at the worst moment.
For maneuver units: The Marine Corps and Army both want expeditionary HPM that can ride on a trailer or a light tactical vehicle. Maneuver forces need fast setup, low crew burden, and battery backed shots during generator swaps. HPM gives a platoon leader a way to pop an incoming quadcopter volley without giving away position with tracers or missile blasts. Expect early field use to focus on convoy overwatch, point defense for refuel points, and pop up protection for command posts.

The engineering bottlenecks that now matter

HPM is no longer a science fair. That makes the boring stuff the most important.

Power budget. The headline is unlimited magazine, but the meter still spins. A battlefield shot might cost a few cents in electricity, yet the source of that electricity drives the whole design. At a base, that is a grid tie, a generator, and a flywheel or capacitor bank to smooth the draw. On a truck, it is a diesel alternator with peaks handled by ultracapacitors. In both cases, you need power reserve for sustained pulses when a swarm approaches in waves. Program managers now ask not only about peak power, but about how many pulses per minute you can fire at 90 degrees Fahrenheit without tripping your own breakers.
Thermal management. Every joule that does not go into the air becomes heat in the box. Heat soaks gallium nitride amplifiers and shifts performance. The solution is boring and heavy: heat exchangers, bigger fans, chilled water on ships, and thermal safeties in software. The trick is to keep the box small enough for a lift gate while keeping the electronics happy in a dust storm.
Software and waveforms. The microwave is sculpted in time and frequency. That waveform is a software product. It must adapt to the target set without interfering with friendly radios. The best systems pull from libraries of pulses, learn from new drone behaviors, and let an operator set a no hurt zone so downed drones fall where they will not hit people or vehicles. Software is also where selective targeting lives. The Camp Atterbury runs showcased the ability to drop specific drones while leaving others untouched, and to nudge a crippled one toward a predefined drop zone. That kind of discrimination will be the difference between permissive range demos and daily use near an airport.
Electromagnetic compatibility. HPM must play nicely with everything else on the net. A base is saturated with radios, datalinks, and navigation aids. A ship is worse. That means guard bands, timing, and coordination with spectrum managers. It also means hard earned certifications so an operator can pulse without killing the base’s own mesh network.
Reliability and sustainment. Microwave emitters do one job until they do not. The maintenance model is moving from field replaceable modules to line replaceable units. The more a system can self diagnose a failing amplifier and swap it in minutes, the more it will be used. The logistics chain must stock radios, amplifiers, capacitors, and fans the way a battery unit stocks fuses.

The rules of engagement problem

Rules of engagement for HPM are new enough that lawyers and operators are writing them together. The technical challenge is straightforward. The policy challenge is not.

Safety of people and ordnance. HPM uses non ionizing radiation, but there are exposure limits. Commanders will demand clear diagrams that show safe arcs near flight lines, fuel points, and ammunition storage. Expect an ocean of signage, training, and checklists.
Collateral effects. A wide beam can also tickle the electronics in nearby civilian devices. That risk is manageable with power settings, aim, and timing, yet it demands coordination with airport authorities and local regulators when bases operate near cities. Selective modes shown this year help here. The more the system can steer and shape the pulse, the easier it is to prove that collateral risk is low.
Spectrum and deconfliction. HPM must live inside crowded radio frequency neighborhoods. That means automated spectrum management that senses where friendly emitters are and gates pulses accordingly. The right answer is integration with the air defense command system so the HPM fires when the rest of the kill chain is ready and quiet.

Exports and controls in 2026

Allies want this. The late summer demo was designed to be watched. But exporting HPM raises hard questions. The United States will treat high power microwave waveforms and amplifier designs as sensitive. Expect Foreign Military Sales for close allies first, with hardware locked to approved waveforms and a black box approach to the most valuable software. End use monitoring will be strict. The pitch to allies will be simple: this is how you defend an air base, a logistics node, or a ship channel against massed quadcopters. The catch is that deliveries will be gated by both production capacity and licensing.

The United Kingdom’s Ministry of Defence has already tested its own high power microwave concept in 2025, and other European and Indo Pacific governments are not far behind. Over the next year, watch for joint demonstrations tethered to major exercises. The timeline will be familiar: demo, small buy, accelerated fielding at one or two critical locations, then a bigger decision in late 2026.

The cost curve is bending

Everyone talks about cost per shot. Here is what it means in practice.

Offense. A basic quadcopter with an improvised warhead can cost a few hundred dollars. The cheapest way to scale is quantity. Offense wins when defense shoots a three hundred thousand dollar missile at a thousand dollar flyer. The Pentagon’s push for massed autonomy is covered in our Replicator program reset analysis.
Defense. HPM flips the math. One pulse can drop a formation. The marginal cost is the electricity for the shot plus wear on the amplifiers. Imagine a dozen pulses to clear a complex attack and you are still in the tens of dollars in energy. Even if you bake in maintenance and crew time, the cost per defeated drone is tiny compared to interceptors.

The offense adapts. Over the next 12 to 18 months, expect to see common countermeasures: better shielding on flight controllers, fiber optic isolation between subsystems, replaceable transient suppressors on power inputs, and navigation algorithms that tolerate electromagnetic spikes. Each hardening step adds dollars and grams. Add up a hardened autopilot, better cabling, and a new power module and you may push a cheap quad from five hundred dollars toward one thousand. That narrows the offense’s volume advantage. More importantly, it reduces swarm size at the outer edge of battery performance. If your hardened drone runs hotter and heavier, its range and loiter shrink.

Defense adapts too. Expect HPM software that automatically varies pulse width and frequency to hunt for the weakest parts of a hardened board. Expect mixed loads where HPM clears the cheap stuff and a laser trims the few hardened stragglers. Expect bases to make space and power for two or three HPM units so they can overlap cones and fire in sequence to handle multi vector swarms. That depth, not a single box, is what changes the cost curve for good.

What to watch through 2026

Fielding to units. The July Generation II buy will flow into testing and then initial operational use in 2026. The tell will be which units receive them first. If a continental United States air base gets an HPM section on the perimeter, the message is homeland defense. If a forward operating location gets one, the message is deterrence by denial.
Maritime trials. A shipboard demonstration with realistic power and integration would be the inflection point for navies. The success metric is dull: zero interference with the ship’s other emitters while still clearing the low altitude picture.
Expeditionary packages. The Marine Corps and Army will push for lighter, faster set up kits. Watch for trailer borne systems with integrated generators and ultracapacitors that can sprint with a convoy or park by a refuel point.
Software releases. The waveforms and target libraries are the secret sauce. Look for software drops that advertise better classification, selective effects against mixed swarms, and automated drop zone management.
Rules and training. The busiest units next year will be spectrum managers and safety officers. The more HPM gets used at real installations, the faster the checklists and range cards mature into doctrine.
Exports. If allies sign letters of offer and acceptance in 2026, it will be for small numbers at first. Watch the language. If waveforms are locked or geofenced, that tells you how the United States intends to keep the crown jewels protected even as it sells the capability.

What commanders and program managers should do now

Plan the power. Treat HPM as a power plant project. Map circuits, buy better generators, and add ultracapacitors. Without that, your magazine is not unlimited.
Design the thermal. In hot climates, assume performance will drift unless you budget for bigger heat exchangers or chilled water taps. Put the radiator where it will actually get airflow, not behind a grill full of mud.
Write the playbook. Integrate HPM into the base or ship’s air picture and fire control now. Run drills where HPM clears a path for guns and lasers to finish what is left.
Train the spectrum. Assign a spectrum manager to the HPM team. Teach crews how to time pulses between friendly transmissions and how to operate selective modes near civilian infrastructure.
Collect the data. Every shot is a lesson. Log waveforms, downed drone models, and failure modes. Feed that back to software updates. This is what will keep you ahead of countermeasures.

The bottom line

The late summer swarm shootdowns made microwave defense feel real. The July Generation II buys made it official. High power microwave is moving from prototypes to fielding, carving out the inner layer of base, ship, and maneuver defense where volume attacks used to be cheap. There is still plenty of work ahead. Power and thermal budgets have to be tamed. Software has to be trusted. Rules have to be written so crews can fire pulses without collateral surprises. But the direction is clear. In 2025 the question shifted from whether HPM works to how fast commanders can wire it into the everyday air defense stack. The winners over the next 18 months will be the teams that do the unglamorous jobs first: power, cooling, software, and training. Do that, and the next time a hundred drones rise together, they will drop together too.