CX2

A drone takes off, and the invisible battlefield lights up

Two people, a few minutes, a thirty-two-pound aircraft. It climbs into airspace where GPS has been jammed into uselessness and radios go to static. Most equipment up there goes blind. This one does the opposite. It listens. It draws a map of every hostile signal below - the jammers, the control links, the ground stations quietly running someone else's drones - and hands that map to the people who need it. The aircraft is called Wraith. The company that built it is CX2.

CX2 is a defense-technology startup in El Segundo, California, with roughly 58 employees and a fairly specific obsession: the electromagnetic spectrum, and who controls it. Not the trenches, not the airspace, but the invisible layer of radio energy that every modern weapon now depends on. CX2 builds AI-enabled hardware and software that detects, disrupts, and defends that layer. They call the goal "spectrum dominance." It sounds abstract until you realize that almost nothing on a battlefield works without it.

The math of modern war stopped adding up

Here is the uncomfortable arithmetic. A cheap first-person-view drone costs about two thousand dollars. The missile traditionally used to shoot one down can cost two million. You can win every engagement and still lose the war, simply by going bankrupt one interception at a time. Ukraine made this visible to anyone paying attention: the front line is increasingly defined by who jams whom, who sees whom first, and who can afford to keep doing it.

Legacy electronic-warfare systems were not built for this. They are expensive, slow to field, and tied to programs that take years and nine-figure budgets. They were designed for a world of large platforms fighting other large platforms, not swarms of disposable drones flooding the spectrum. The threat got cheap and fast. The defense stayed costly and slow. That gap is the thing CX2 exists to close.

It is a tidy irony of the current arms race that the most advanced militaries on earth can be drained by hardware you could assemble from a hobby catalog. CX2's wager is that the answer is not a bigger missile. It is better information, delivered cheaply, at the speed the spectrum actually moves.

Four founders who had each seen the gap from a different angle

The genesis is almost too neat. Porter Smith, a military veteran and former investment partner at Andreessen Horowitz, went to Ukraine in 2023 and watched electronic warfare decide engagements in real time. He came back convinced the spectrum was the decisive domain - and called Nathan Mintz. Mintz is a Stanford-trained electrical engineer who spent roughly fourteen years on EW and radar systems at Raytheon and Boeing, then co-founded the directed-energy company Epirus, which became a defense unicorn, and later the automotive-radar startup Spartan Radar. He had spent his career on exactly this problem.

They were joined by Lee Thompson, who worked on radio and RF systems for SpaceX's Starship program and became CX2's CTO, and Mark Trefgarne, who had built the video ad exchange LiveRail and sold it to Meta for around $500 million before becoming the company's president. It is an unusual mix - directed-energy physics, rocket-grade radio engineering, and consumer software scaling - and that is rather the point. CX2's bet is that electronic warfare should look less like a defense program and more like a fast-moving hardware-plus-software company.

A sensor that hunts, and a payload that finishes

CX2's core insight is a pairing it sums up as "Hunt the Archer, not the Arrow." Wraith is the hunter. It is a Group 2 unmanned aircraft, about thirty-two pounds, NDAA-compliant, deployable by two people in minutes. In environments where GPS has been jammed, it navigates by visual odometry and multiband GNSS rather than going dark. It detects, classifies, and geolocates hostile emitters and produces spectral heatmaps of the battlefield - turning invisible radio energy into something a commander can actually point at.

The economics are the real product. A Wraith finds the emitter; a cheap drone carrying Vadris flies to it. The expensive part is the information, not the explosive - and the information is reusable. It is a quietly subversive design: instead of matching an adversary's cheap drones with costly missiles, CX2 tries to make the defense cheaper than the offense. The spectrum, for once, on a budget.

Money, contracts, and a 300-watt jammer that did not work

In May 2025, CX2 closed a $31 million Series A led by Point72 Ventures, with participation from Andreessen Horowitz, 8VC, 201 Ventures, Pax Ventures, and Upfront Ventures - bringing total funding to roughly $46 million. That capital follows real customer signal. The Defense Innovation Unit awarded CX2 a roughly $2 million DRM3 contract and downselected Vadris for continued evaluation. Wraith was developed under a cooperative R&D agreement with a U.S. government partner, evaluated by the U.S. Army, and flown at NATO's Joint Multinational Readiness Center in Hohenfels, Germany.

The most telling data point is not financial. In testing, Wraith held accurate geolocation against a 300-watt GPS jammer - the kind of brute-force interference designed to make exactly this sort of aircraft useless. It kept working. For a company whose entire thesis is "we operate where others go blind," that is the proof that matters.

Make sure the right side can always see

CX2's stated mission is to ensure its allies always have the upper hand in the electromagnetic spectrum: to illuminate the unseen battlefield, outmaneuver adversaries, and enable decisive action anywhere, anytime.

Strip away the defense vocabulary and the idea is simple. The spectrum has become contested space, and most of the systems guarding it were designed for a slower, more expensive era. CX2 wants the people it works with to keep their senses when an adversary tries to take them away - and to do it without spending a fortune per engagement. It is a hardware company, but its real output is awareness.

Cheap drones are not going away. Neither is the spectrum.

The trend that created CX2 is not reversing. Drones get cheaper and more numerous every year, and every one of them lives or dies by radio. Whoever can read and shape the spectrum - quickly, cheaply, at scale - holds the advantage. CX2 is betting that this advantage belongs to companies that ship like software firms and think like RF engineers, rather than to programs measured in decades. Whether they are right is the open question. The contracts, the field tests, and the jammer that failed to blind them suggest they are at least asking it seriously.

Return to that drone over the jammed field. Two years ago, the spectrum below it was a fog: everyone's signals, hidden, equal, unreadable. Now a thirty-two-pound aircraft turns that fog into a map, hands it to the people who need it, and points a cheap interceptor at the thing that was hiding. The invisible battlefield is still invisible. CX2 just decided that should not be an excuse.

Interviews and conversations