General Radar Corp.

A 14-person shop on Constitution Drive, quietly arguing with the entire defense-radar industry.

Walk into General Radar Corp. and you will not find a factory floor humming with spinning dishes. You will find engineers staring at point clouds - the kind of 3D data most people associate with self-driving cars - except these were drawn by radio waves, not lasers. The company builds active electronically scanned array (AESA) radar: panels that never move, that sweep the sky by steering beams electronically, and that lean on GPUs and software for the parts a traditional radar solves with expensive, single-purpose metal.

It is a small team for such a large claim. Roughly fourteen people, headquartered in Menlo Park, California, sitting in a market historically owned by primes with five-figure employee counts. General Radar's pitch is not subtle: take the architecture that costs aerospace programs a fortune, and rebuild it from commercial parts you can actually buy.

Radar got good a long time ago. Then it mostly got expensive.

High-performance radar has long been a luxury good. A capable aerospace radar can cost millions, take years to field, and lock you into hardware designed for exactly one mission. That worked when the customers were governments with deep budgets and a single job to do. It works less well in a world that now wants radar for wind farms, drone defense, airfield safety, autonomous vehicles, and weather - all at once, and all on a budget.

Two stubborn facts sat at the center of the problem. Custom radar silicon is costly. And legacy systems are rigid: built around fixed waveforms and mechanical scanning, they resist the kind of software upgrades that have transformed almost every other piece of modern electronics. Wind turbines, famously, confuse old radar - their spinning blades throw back clutter that conventional systems struggle to ignore.

An MIT engineer who built antennas for spacecraft decided radar was a software problem.

Dmitry Turbiner founded General Radar in 2016. Before that, he worked on RF and antenna engineering at NASA's Jet Propulsion Laboratory, holds a BS in electrical engineering from MIT, and did graduate work at Stanford. His bet was contrarian in the way good deep-tech bets usually are: that the expensive, bespoke parts of radar could be pushed into software and commercial silicon, and that doing so would not just cut cost - it would unlock missions nobody bothered building radar for because the economics never closed.

Kleiner Perkins backed the idea early, leading a $3.4M seed in 2016. Turbiner then surrounded himself with people who knew exactly how the incumbents built things - and therefore where the bodies were buried. The roster includes former Raytheon engineering fellows David Cushing and Jeff Peacock, decades of DoD radar and software experience between them, plus a former Naval Intelligence officer running the government side. The advisory board reaches all the way up to a retired four-star admiral who once commanded both INDO-PACOM and NORAD/NORTHCOM.

One architecture. A GaN front-end, digital beam-forming, and a supercomputer doing the thinking.

The recipe is consistent across the lineup: a gallium-nitride (GaN) AESA front-end to transmit and receive, digital beam-forming to shape and steer the signal in software, and a GPU supercomputing backend to turn raw returns into usable, high-resolution data. Because so much lives in software, the same hardware can chase different missions by changing its waveforms - including AI-assisted imaging and target identification.

The whole argument fits in one chart.

General Radar's central claim is economic, not just technical: that its approach cost-reduces an aerospace-grade radar by roughly twenty times. Treat the exact figure as the company's own, but the direction is the entire thesis - take cost out, and radar stops being a luxury reserved for the biggest budgets.

Democratize the radar architecture - then watch what people build with it.

General Radar describes its mission, plainly, as democratizing the aerospace radar architecture. Strip the cost and the rigidity out of radar, and you do not just sell a cheaper version of an old thing. You change who gets to use it. A wind-farm operator. An airport worried about debris on the runway. A base that needs to see small drones. A weather service. Each of these is a market that conventional radar economics quietly priced out.

That is what makes the company genuinely dual-use rather than just defense-adjacent. The same software-defined panel that matters to national security also matters to clean energy and aviation safety. It is a tidy thesis - perhaps too tidy, which is the standard objection to every ambitious hardware startup. The honest answer is that the proof is in the fielding, and General Radar is still early.

The sky is getting crowded. Cheap, sharp radar is about to be in demand.

Drones are multiplying. Airspace is busier. Wind farms keep going up in places radar needs to see through. Autonomy wants sensors that work when cameras and lasers fail in fog and rain. Every one of those trends points the same way: toward more radar, in more places, at prices that legacy hardware can't hit. If General Radar's bet pays off, the company is positioned at exactly that intersection - and if it doesn't, it will have at least proven how far commercial silicon can push a famously conservative field.

Return, for a moment, to that room on Constitution Drive. The panels still don't spin. The engineers are still staring at point clouds drawn by radio. But the thing on the bench is no longer just a clever demo of what radar could cost. It is a small company's argument, made in hardware, that the most expensive sensor in the sky doesn't have to be - and that the next decade of radar might be written in software by a team of fourteen, rather than handed down by the primes.