A $5,000 open-source radar that sees 20 km, built by one engineer in Morocco

Commercial phased-array radar systems cost between $120,000 and $250,000, and that’s before you pay for the test equipment you need to calibrate them. Nawfal Motii, an electronics engineer working out of Morocco, just posted a complete open-hardware alternative for $5,000 to $7,200 in parts. The repo has 13,100 stars and 2,700 forks. If you squint, it’s the SDR moment for radar.

What Aeris-10 actually is

Aeris-10 is a 10.5 GHz pulse linear frequency-modulated (PLFM) phased-array radar. It comes in two variants: the Aeris-10N Nexus, which has a range of about 3 kilometers and uses an 8-by-16 patch antenna array, and the Aeris-10E Extended, which pushes range out to 20 kilometers using a 32-by-16 slotted-waveguide array and gallium-nitride power amplifiers. Both do full electronic beam steering, plus or minus 45 degrees in azimuth and elevation, using Analog Devices ADAR1000 phase-shifter chips.

The processing stack is off-the-shelf but cleverly integrated. A Xilinx Artix-7 XCA7A50T FPGA handles the FFT math. An STM32F746 microcontroller orchestrates everything else: frequency synthesizers, ADCs, DACs, GPS, barometer, stepper motors for coarse pointing, cooling fans. The software pipeline implements Moving Target Indicator filtering and Constant False Alarm Rate detection, plus Doppler speed estimation. Tom’s Hardware confirms it ships with a desktop GUI for scanning and target tracking.

Why this is the interesting number

95% cheaper than commercial is the headline. The more interesting story is that it’s possible at all. Five years ago, a single ADAR1000 chip was restricted to defense contractors under ITAR-like export controls. Today it’s a $300 part you can buy on Digi-Key. The same thing happened with GaN power amplifiers: military-only in 2015, commodity in 2026. Motii’s accomplishment is less “I invented a cheap radar” and more “I read 40 datasheets carefully and proved these parts compose into something useful.”

That’s the SDR pattern all over again. The HackRF and LimeSDR turned software-defined radio from a $10,000 lab instrument into a $300 dongle. The ecosystem that grew around that (GNU Radio, community plugins, actual field use) didn’t exist before the hardware hit hobbyist price points. Radar is about to get the same treatment, and Aeris-10 is the tentpole project.

The license choice matters

Motii originally released everything under MIT. He switched the hardware components to CERN Open Hardware Licence v2 Permissive after community feedback, keeping the firmware and GUI under MIT. That’s the right call.

MIT is great for code. It doesn’t quite fit hardware, where you have questions about patent rights, mechanical designs, and what “derivative work” means for a PCB layout. CERN-OHL was designed for exactly this problem, and the Permissive variant is the closest equivalent to MIT in the hardware-licensing space. Projects that want to fork Aeris-10 for a commercial variant (a drone-detection startup, say) can do so without tripping copyleft clauses. Projects that want to learn from the design can read and remix freely. This is a small legal detail that has cost other open-hardware efforts dearly.

The bill of materials, explained

The $5,000 floor for the Nexus variant breaks down roughly like this: the FPGA and microcontroller together are about $200, the ADAR1000 phase-shifter chips are the single biggest line item at roughly $1,800 for the 8x16 array, patch-antenna PCBs run $600 to $900 depending on fab, RF front-end components (mixers, filters, LNAs) add another $1,200, and the rest is connectors, power supply, enclosure, cooling, and stepper motors for the coarse-pointing yoke. Nothing exotic. Everything Digi-Key stocks.

The 10E Extended variant jumps to $7,200 mostly because of the GaN power amplifiers, the larger slotted-waveguide fabrication, and more phase-shifter channels for the wider array. The BOM hits what’s essentially the same engineering problem at higher scale: you get more gain, more beam steering, and more range for a roughly linear cost increase. That scaling behavior is what makes the commercial comparison unfair. A commercial $250,000 system isn’t 50 times better than the Aeris-10E. It has a support contract, a certified test report, and a sales engineer. That’s most of the price delta.

Real-world uses (and limits)

Three buckets of users will care.

• Hobbyist drone-detection. At 20 km range and a few thousand dollars, Aeris-10E outperforms every off-the-shelf “drone radar” in the consumer/prosumer price range. Expect somebody to integrate it with ADS-B receivers within a quarter.
• Weather and atmospheric research. Universities and small research groups have wanted affordable phased-array weather radar for years. NOAA’s MPAR program runs into millions per site. An Aeris-10 array won’t match NOAA precision, but it’s good enough for graduate research and field campaigns in places where commercial options don’t pencil out.
• Educational. Phased-array fundamentals are usually taught on cartoons because real hardware is locked behind defense contractors. A $5,000 system that fits on a workbench changes what universities can show undergrads.

The limits are real too. This is not certified aviation equipment, it’s not going into any flight-safety application, and the FCC will care if you light it up in the US without an amateur radio license covering the 10 GHz band. Running radar hardware into a populated area without credentials is a regulatory problem, not a technical one. GitHub readme is clear that operators are responsible for local spectrum rules.

Why you’re hearing about this now

Two reasons. First, Motii announced a Crowd Supply release for Q3 2026, which means pre-built kits shipping to backers who don’t want to source 200 parts themselves. The Crowd Supply factor is what turns an impressive repo into a product people actually build with. Second, geopolitics: drone defense, border sensing, and counter-UAV work are all looking for cheaper sensor packages, and a $7,200 radar is precisely the kind of thing procurement officers notice.

It’s also, bluntly, a great reminder that the open-hardware story isn’t over. Between Aeris-10, this week’s RISC-V Rust 1.95 release, and the general trend of sovereignty-driven open-silicon projects, the long tail of hardware is moving out of proprietary lockdown faster than anyone in 2018 expected. For how that fits into the broader compute and AI-chip conversation, our coverage of Meta and Broadcom’s MTIA deal is the counterweight: open at the edges, hyperscale-custom at the core.

What this means for you

If you’re a maker or RF hobbyist, this is the most interesting repo on GitHub this week. Clone it, read the schematics, order the ADAR1000 eval kit first (it’s the critical part and the biggest risk of a build), and wait for the Crowd Supply kit if you don’t want to hand-solder QFN packages. If you build drone-detection or airspace products commercially, the Aeris-10 variant is now your BOM floor. Any bid you write that doesn’t beat $7,200 for a 20-km phased-array sensor is going to look expensive. If you’re in education, this is a cheap way to teach phased-array concepts with hardware students can actually touch. My read: the project is genuinely useful, the license is clean, the regulatory caveats are real, and Crowd Supply pre-orders will tell us by Q4 whether radar has its HackRF moment or stays a niche curiosity. Check the GitHub repo’s issues page for the kind of contributors showing up. If RF engineers are filing substantive PRs, that’s the signal that this is going to keep getting better, and that the drone-sensor and weather-radar categories are about to get a lot more interesting in the second half of the year.