The best drone killer in the world exists — the Pentagon won't buy it

THE FIRST AMERICAN combat deaths of the Iran war came not from a ballistic missile or a fighter jet but from a Shahed-136 drone — an Iranian weapon that costs somewhere between $20,000 and $50,000 to produce. On March 1st, one struck a temporary operations center at the Port of Shuaiba in Kuwait, killing four Army Reserve soldiers and wounding dozens more. Three weeks into Operation Epic Fury, roughly 200 U.S. troops have been wounded, the majority by drone strikes, according to CENTCOM.

Iran has launched more than 2,000 one-way attack drones and 500 ballistic missiles since February 28th, hitting targets from Dubai International Airport to Saudi Aramco's Ras Tanura facility to the American consulate in Dubai, which was set ablaze by a Shahed. On Wednesday, Iranian missiles struck Qatar's Ras Laffan Industrial City — the world's largest LNG export facility, responsible for roughly 20% of global supply — causing extensive damage and fires that QatarEnergy now says could take years to repair. Qatar expelled Iran's military attachés within hours. Brent crude surged past $111. The drone threat, in other words, is no longer merely a battlefield problem; it is an energy-market crisis, a diplomatic accelerant, and — as of this week — a homeland-security question. On Wednesday night, the Washington Post reported that unidentified drones had been detected over Fort McNair, the Washington Army base where Secretary of State Marco Rubio and Defense Secretary Pete Hegseth reside, prompting a White House meeting and discussions about relocating both officials. No one has determined where the drones came from.

The Pentagon reckons it has degraded Iran's drone capacity by 83% since the war began. That still leaves a lot of drones in the air — and the chairman of the Joint Chiefs told Congress in a closed-door briefing this month that one-way drones were "posing a bigger problem than anticipated," and that U.S. air defenses could not intercept all of them.

Somewhere in the CENTCOM theater — the Army will not say where, exactly — sit a handful of Epirus Leonidas systems, the U.S. military's only operational high-power microwave weapon purpose-built to down drone swarms. At a live demonstration at Camp Atterbury near Indianapolis last year, Leonidas disabled 49 drones simultaneously with a single electromagnetic burst, in front of observers from multiple military branches and allied nations. In January 2026, the company achieved what it calls a landmark: the first known defeat of fiber-optic-guided drones, the autonomous variety immune to conventional jamming that has become ubiquitous on the Ukrainian front and is now proliferating to state militaries globally. The technology, by most accounts, works. The problem is not the weapon. The problem is scale.

The Army has four Generation I prototypes deployed for testing and has contracted for just two Gen II units at $43.5 million. At a moment when thousands of cheap drones are reshaping the geometry of a live war, the most promising counter-swarm weapon in the American arsenal is functionally a science project.

The economics of drone warfare are brutally simple, and they favor the attacker. A Patriot PAC-3 interceptor costs upward of $3 million per shot. A Stinger missile runs several hundred thousand dollars. The Shaheds raining down on American bases across the Gulf cost a fraction of either. Washington's use of the expensive to neutralize the cheap is, as Foreign Affairs recently put it, unsustainable: Operation Rough Rider against the Houthis in 2025 cost nearly a billion dollars, driven in large part by the price of defensive missiles fired at drones built in relatively simple factories. The Iran war has only deepened the asymmetry. In the first 24 hours of the conflict alone, Iran launched 541 drones at the United Arab Emirates.

The Pentagon has known about this imbalance for years — since at least late 2016, when ISIS fighters began strapping grenades to DJI Phantom quadcopters over the Battle of Mosul. "I went, 'Oh, this is going to be bad,'" the Army's chief technology officer, Alex Miller, has said of that moment. "Basically it's an airborne IED at that point." The danger has built steadily since, with advances in machine vision, AI-navigation software, and suicide-drone tactics accelerating after Russia's invasion of Ukraine. The Pentagon has studied directed-energy weapons as a potential answer since the Cold War, when a high-altitude nuclear test over the Pacific unexpectedly knocked out streetlights in Hawaii and set off decades of classified research into electromagnetic pulse effects. Yet the institutional machinery that produces $13 billion aircraft carriers has proved remarkably slow at manufacturing $16.5 million microwave emitters.

Epirus, for its part, has not lacked for capital or credentialing. The Torrance, California-based startup has raised $595 million in venture funding across seven rounds, including a $250 million Series D last March led by 8VC and Washington Harbour Partners that values the company north of $1 billion. General Dynamics Land Systems, one of the Pentagon's largest prime contractors, participated as a strategic investor — a signal, in defense-industry semaphore, that the grown-ups are paying attention. The founding lineage traces through Palantir (co-founder Joe Lonsdale started both companies) and, more importantly, through Raytheon, the microwave technology pioneer cofounded by MIT's Vannevar Bush in 1922. Raytheon's Next Generation Jammer program supplied the gallium-nitride semiconductor architecture at the heart of Leonidas, and CEO Andy Lowery was the chief engineer on that very program before decamping for the startup world. His current CTO, Matt Markel, is ex-Raytheon. So are Epirus's chief engineer for defense, its VP of engineering, its VP of operations, and a healthy share of the rank and file. The company, in other words, is Raytheon's radar technology reversed — the same phased-array fundamentals, repurposed to fry drones at close range rather than detect missiles at long range.

The pedigree is formidable. The production capacity is not. Epirus can build roughly 20 to 30 units per year at its current facility, a figure that would need to multiply several times over to meet the kind of demand a sustained regional conflict generates.

But the deeper issue is not manufacturing floor space — it is the Pentagon's acquisition apparatus. The Leonidas entered the Army's Indirect Fire Protection Capability pipeline in January 2023, when Epirus won a $66 million contract after outperforming six competing systems. It has been proceeding through the methodical cadence of prototyping, testing, user feedback, and (eventually, perhaps) a program of record ever since. The Army topped up the contract with another $17 million and, last July, awarded $43.5 million for two Gen II systems featuring 30% more power and double the effective range. A critical field test at Naval Air Weapons Station China Lake was scheduled for October 2025; Lowery called it "a defining moment" for the company. Formal Army acquisition, if the tests go well, would follow — on defense-procurement timelines, which is to say, not quickly.

That cadence made a certain sense in peacetime. It makes considerably less sense when Shaheds are striking American air bases in Saudi Arabia, damaging five KC-135 refueling tankers at Prince Sultan Air Base in a single attack, and killing service members at forward operating posts across the Gulf.

The competition, meanwhile, is crowding in. Thales has tested its RapidDestroyer microwave system with the British army in Wales, downing swarms with what the U.K. Ministry of Defence called "near-instant effect." Leidos expects to deliver its Mjölnir system — named, with characteristic defense-industry subtlety, for Thor's hammer — to the Air Force Research Laboratory by early next year. RTX is developing two container-sized microwave platforms, PHASER and CHIMERA; Lockheed Martin is refining MORFIUS, a microwave weapon packaged in a 14-kilogram drone designed to fly into an enemy swarm and fire. China displayed a competing system called Hurricane at a major military exhibition in late 2024. None of these are fielded at scale either. The entire category remains stuck in what defense procurement veterans call the "valley of death" — the gap between a successful demonstration and a production contract with actual volume behind it.

What distinguishes Epirus from the pack is its approach to the underlying physics. Most competing systems still rely on magnetron vacuum tubes — the same technology inside your kitchen microwave, albeit vastly more powerful — which are hot, bulky, and require regular maintenance. (The Air Force's THOR system, for instance, fills an entire shipping container and needs a dish antenna to aim.) Epirus instead uses bespoke gallium-nitride semiconductor chips that produce sustained microwave pulses in the millisecond range, compared with the nanosecond bursts of magnetron emitters. The result is a weapon that fits on a trailer towable by a standard Army truck, can flick between targets at the speed of software via phased-array beam-steering, and — critically — works against autonomous, unjammable drones by hitting them broadside with enough electromagnetic energy to short-circuit whatever wire or circuit the beam finds first. Copper tape, propeller shafts, exposed antennas: in testing, the microwave always finds a way in.

The Pentagon's senior leadership is now publicly acknowledging what the battlefield has made obvious. Assistant Secretary of Defense Michael Dodd told a defense-industry conference in Honolulu on March 9th that the Department of Defense plans to field directed-energy weapons at scale within 36 months. President Trump touted lasers and microwaves at a White House press conference the same day. The UK has dispatched Ukrainian counter-drone specialists to the Gulf to help allies cope with the Shahed barrage — a remarkable admission that Europe's hardest-won battlefield expertise is now needed to defend some of the world's richest cities.

Lowery has said Epirus is exploring expanded production near Fort Sill, Oklahoma, and that the Air Force is likely to begin leasing Leonidas units this year. International sales are now permitted — a policy change from 2025 — and Japan, Australia, and several other allies have expressed interest. Epirus even requested permission to send systems to Ukraine but was initially denied; recently, a U.S. official encouraged the company to resubmit, signaling that the Trump administration's posture may be shifting. At least two Ukrainian firms, Transient Technologies and First Contact, are developing their own microwave weapons, albeit at the proof-of-concept stage. The race to counter the drone swarm is global. The question is whether anyone can run it fast enough.

Whether Epirus can convert its technological edge into production at wartime speed will depend less on the physics of gallium-nitride semiconductors than on the bureaucratic physics of Pentagon procurement — a system optimized for deliberation, not urgency. The U.S. military spent decades and hundreds of billions preparing for peer conflict with a great power. What it got instead, for now, is a swarm of cheap Iranian drones overwhelming expensive defenses at forward bases across the Gulf, killing soldiers with weapons that cost less than a mid-range sedan. Epirus built a machine that can drop 49 of them at once with the push of a button. The Army, at last count, has bought four. ■

For more, join 75,000 subscribers getting tech's favorite brief here