Cheap drones broke the math on air defense, and everybody in the Pentagon knows it. A quadcopter that costs a few hundred bucks can pull a six-figure interceptor out of a launcher, and if fifty of them show up at once, the defender runs dry long before the attacker does.
The Marine Corps just put $11 million against that problem. The contract went to Epirus, a Torrance, California company that builds high-power microwave systems, and it buys a vehicle-mounted weapon called HAVOC.
The pitch is one-to-many. Instead of assigning an interceptor to each drone, HAVOC puts out a cone of microwave energy and scrambles the electronics of everything inside it at once.
Delivery is set for the second quarter of 2027. So this is a procurement story, not a fielding story, and a fair amount of the coverage this month has been sanding down the difference.
HAVOC descends from ExDECS, not straight from Leonidas
The family tree matters here, and most writeups collapse it.
Epirus builds Leonidas, the high-power microwave platform the Army has been testing since 2023. From that came the Leonidas Expeditionary, and from that came ExDECS, the Expeditionary Directed Energy Counter-Swarm system the company delivered to the Marines in 2025 under a $5.5 million contract, also funded through the Office of Naval Research.
HAVOC is the next rung. Epirus describes it in its own announcement as an autonomy-enabled evolution of ExDECS, not a Leonidas derivative.
That distinction carries weight because ExDECS was trailer-mounted. HAVOC arrives on what Epirus calls a universal sled mount, meant to move between manned and unmanned ground vehicles without custom integration work for each platform.
The Army fed the design. Its Leonidas units went to the Middle East and the Pacific on an experimental basis, and CEO Andy Lowery has said that deployment experience shaped what the Marines are getting.
Nobody will tell you the range, and there’s a reason
Here is where the reporting gets loose. Several outlets this month have printed a figure of about two kilometers for the current system, sometimes with a claim that Epirus is working to double it.
Lowery declined to give specifics. He told Marine Corps Times that the range increase is significant and then said he could not put a number on it.
The Defense Department has been firmer than that. When ExDECS was delivered last year, TWZ reported that its maximum effective range is classified, with the department confirming only that Epirus systems work at tactically relevant distances.
What Lowery did offer was an analogy. He described the generational jump as going “from a 50-kilowatt to 150-kilowatt type laser in the same size package.” Take that for what it is, which is a vendor explaining a classified improvement in terms a reporter can print.
He also flagged the unglamorous half: reliability, weather tolerance, dust, and general ruggedness for a Corps that drags equipment through bad places.
MADIS gets a third engagement layer
The Marines already field short-range air defense. MADIS, the Marine Air Defense Integrated System, went to operational units in December 2025 and runs as a networked pair of Joint Light Tactical Vehicles.
One truck is the kinetic half, carrying Stinger missiles and a 30mm chain gun. The other is the sensor and electronic warfare node, with a 360-degree radar, jamming gear, and a minigun.
MADIS handles a handful of drones well. Against fifty, it becomes a magazine problem, which is the same wall every gun-and-missile system runs into.
Lowery told DefenseScoop the new weapon is designed to run alongside MADIS rather than replace any piece of it. Stack them and you get three tiers: jam what can be jammed, microwave what shrugs off jamming, and shoot whatever survives both.
Epirus executives have been careful not to oversell it. Speaking to The Jerusalem Post in January, they described Leonidas as one layer in a stack rather than a solution on its own.
Fiber-optic drones are the specific problem
One threat category explains a lot of the urgency behind this contract.
A fiber-optic guided drone trails a physical spool of cable back to its operator. It receives no radio commands and emits no control signal, which means conventional radio-frequency jamming has nothing to grab onto. Ukraine and Russia have both been using them heavily.
A microwave weapon doesn’t care how the drone is being steered. It goes after the electronics themselves, so the guidance method stops mattering.
That is the gap the Marines are buying into. Their existing electronic warfare gear covers the drones that listen for radios. HAVOC is meant to cover the ones that don’t.
Tower 22 is the scenario Epirus keeps pointing at
Lowery has been explicit about the use case, and it isn’t hypothetical.
In January 2024, a drone attack struck Tower 22, a U.S. outpost in northeastern Jordan, killing three soldiers and wounding dozens more. Lowery has repeatedly cited that attack as the exact scenario his systems are built for, and argues a properly employed Epirus system would have changed the outcome.
That is a vendor’s claim about a hypothetical, and it should be read as one. It is also a reasonable description of what the Corps is trying to protect: refueling points, radars, communications nodes, generators, and the air defense vehicles themselves.
There is something recursive about parking a microwave weapon next to a Stinger battery so drones can’t kill the thing that kills the bigger drones. That is what layered defense looks like now.
The cost exchange is the entire argument
A Stinger runs into six figures. A Shahed-type one-way attack drone runs a few thousand dollars. Every planner staring at Ukraine and the Red Sea has been trying to invert that ratio for three years.
Directed energy inverts it by turning ammunition into electricity. The capital cost is front-loaded into the hardware, and the marginal cost of an engagement drops toward the price of running a generator.
The Pentagon is betting on that broadly, not just here. In May it named five installations, including Fort Bliss, Fort Huachuca and Grand Forks Air Force Base, as operational test sites for laser and high-power microwave counter-drone systems.
Epirus says it is ready for volume. Lowery told DefenseScoop the company has “the capacity to get up to about one system per week” and has prepared its facilities for low-rate production. Capacity is not the same thing as output, and no one has ordered at that rate yet.
What happens next is the part worth watching. If the Corps exercises its option for a second unit after testing in 2027, microwaves stop being an experiment and start being kit. If it doesn’t, this stays an $11 million answer to a problem that keeps getting cheaper to create.
