When BAE Systems unveiled the Shadow EW product family at the Air & Space Forces Association’s Air, Space & Cyber Conference in National Harbor, Maryland this week, it did something unusual for the world’s dominant electronic warfare contractor: it named the commercial semiconductors powering the system. The NVIDIA Jetson Orin NX AI processor, the AMD-Xilinx Kintex KU15P field-programmable gate array, and a Texas Instruments TDA4 automotive-grade system-on-chip are not custom defense silicon — they are the same commercial components powering autonomous robots, industrial cameras, and high-volume automotive systems. That specific choice is the engineering architecture that makes Shadow EW’s “affordable mass” promise possible — and it is also the detail that the press release buries.

Why Every Airborne Platform Now Needs Electronic Warfare

For two decades, the U.S. military flew in electromagnetic environments so uncontested that electronic warfare was essentially optional for smaller platforms. The war in Ukraine ended that assumption with brutal efficiency. An Air Force Special Operations Command official described the shift in blunt terms: AFSOC’s aircraft “operated in largely benign electromagnetic combat environments, often guided by special tactics airmen on radios, in virtually uncontested airspace against insurgent groups possessing no electronic warfare capabilities” — an era that Ukraine’s war has closed.

Ukraine demonstrated what happens when that equation reverses. Both sides saturated the electromagnetic spectrum so thoroughly that each eventually resorted to fiber-optic cable-guided first-person-view drones — wire-guided precision weapons that cannot be jammed because they carry no radio link at all. Ukrainian forces report that electronic warfare now neutralizes more than half of enemy aerial targets (that figure comes from Ukrainian government sources and has not been independently confirmed by Western intelligence agencies). The broader lesson that defense planners have internalized is this: the electromagnetic spectrum is a contested warfighting domain for every airborne asset, regardless of size.

The problem is that electronic warfare has historically been expensive. Systems built for the F-35 and B-21 cost tens of millions per unit and take years to develop. The helicopters, legacy fixed-wing aircraft, and proliferating drone fleets that make up the majority of any large force’s airborne assets have gone largely unprotected because no affordable, mass-produceable EW option existed.

Shadow EW: What the Product Family Is

Shadow EW addresses that gap with three named hardware variants. The Shadow EW 100 is the smallest and most affordable, built specifically for drones and systems with severe size, weight, power, and cost — or SWaP-C — constraints. It weighs 7.2 pounds (3.3 kilograms), measures 4.4 by 3.34 by 6.94 inches (11.2 by 8.5 by 17.6 centimeters), draws fewer than 150 watts, and covers RF frequencies from 70 megahertz to 6 gigahertz. Production begins in 2027 per the product datasheet.

The Shadow EW 120 is a hybrid software-defined radio transceiver — meaning it can both receive and transmit — covering the same 5 megahertz to 6 gigahertz band while adding dual 1.5 gigahertz instantaneous bandwidth transmit channels. Its core processing combines a TDA4 automotive processor with the Kintex KU15P FPGA, with an optional Jetson Orin NX module available for additional AI inference capability. This configuration optimizes the 120 for radar EW operations while keeping the door open to signals intelligence and communications intelligence roles via its streaming in-phase and quadrature interface. The 120 datasheet confirms BIG IRON framework compatibility.

The Shadow EW 500 is the high-performance option in the family, stretching coverage to 18 gigahertz and adding dual 4-gigahertz instantaneous bandwidth receive channels for precise direction-finding and geolocation. It uses an Orin AGXi system-on-module and UltraScale+ ZU47 for higher-performance edge processing.

All three variants are software-defined radios: their core functions — signal detection, waveform generation, threat classification, countermeasure selection — are implemented in software running on programmable hardware, rather than in fixed analog circuits. That means they can be updated in the field as new threat signatures emerge, without swapping hardware.

“The modern battlefield is changing at an unprecedented pace and all platforms, regardless of size, require adaptable electronic warfare capabilities to survive,” said Rebecca Cruz, Director of Shadow EW products at BAE Systems. “Shadow EW solutions provide flexible, configurable hardware paired with iterative software updates to counter rapidly evolving threats.”

What Makes the Chip-First Bet Work — and What It Costs

The specific chip architecture matters because the engineering problem being solved is not just “make EW smaller” — it is “separate the three distinct computational jobs inside an EW system and assign each to the commercial chip that performs it most efficiently.”

An electronic warfare system must perform three functionally separate tasks. The first is real-time signal processing: monitoring a wide band of the RF spectrum, detecting radar pulses, and generating Pulse Descriptor Words — a structured data format that captures a pulse’s frequency, width, amplitude, time of arrival, and direction of arrival. This task requires parallel processing at deterministic, low-latency speeds that neither a CPU nor a GPU can reliably deliver. It requires an FPGA. The AMD-Xilinx Kintex KU15P is an FPGA from one of the world’s leading defense-grade FPGA suppliers — formerly Xilinx, before AMD’s $35 billion acquisition closed in 2022. Xilinx FPGAs, and their successor AMD-Xilinx devices, have dominated defense EW and radar signal processing precisely because of this capability.

The second task is AI inference: taking the stream of identified signals, matching them against threat libraries, classifying the threat type, and selecting the optimal countermeasure response. This is the task that has changed most dramatically in the past decade — it is no longer a simple lookup table but an increasingly machine-learning-driven classification problem. The NVIDIA Jetson Orin NX 16Gb is an edge-AI processor delivering up to 100 TOPS (tera-operations per second) of AI inference performance at configurable power modes between 10 and 25 watts — fast enough for onboard real-time threat classification in a package roughly the size of a credit card.

The third task is platform integration: managing communications interfaces, timing, control, and data routing between the other components and the host platform. The Texas Instruments TDA4, an automotive-grade system-on-chip used in advanced driver-assistance systems, handles this role in the Shadow EW 120. The automotive sourcing is deliberate — TDA4-family chips are produced in high volumes for the automotive industry, are qualified for harsh environments under the AEC-Q100 standard, and carry commercial pricing because automotive demand dwarfs defense demand.

That architecture — FPGA for real-time parallel signal processing, AI accelerator for threat classification inference, automotive SoC for interface management — is what makes high-volume production economically viable. Defense-custom ASICs that combine all three roles into a single chip cost tens of millions of dollars to design and qualify, require years of development, and are produced in quantities of hundreds or thousands. Commercial chips cost far less per unit, can be ordered off the shelf, and can be updated through software as the threat environment evolves.

There is a structural tension embedded in that bet, however. Both the NVIDIA Jetson Orin NX and the AMD-Xilinx Kintex KU15P are produced primarily by TSMC, the Taiwanese foundry that manufactures the majority of the world’s advanced semiconductors. The Department of Defense has spent billions to reshore semiconductor manufacturing — the CHIPS and Science Act appropriated $52.7 billion for domestic semiconductor manufacturing — specifically to reduce this fab-concentration risk. Shadow EW’s commercial chip strategy accelerates capability delivery and reduces unit cost, but it does so by leaning on the same supply chain geography that the CHIPS Act is trying to change. That tradeoff is real and worth noting, even if it is not visible in any product announcement.

All three Shadow EW variants are MIL-STD-810H qualified — the U.S. military’s environmental durability standard covering shock, vibration, temperature extremes, humidity, and altitude — and all are compatible with BIG IRON (a defense open-standard framework for EW software integration), WOSA (Weapon Open System Architecture), and Open Mission Systems standards. That compatibility is not incidental: it positions Shadow EW directly within the U.S. Air Force’s Modular Open System Architecture acquisition paradigm, the DoD-mandated procurement approach that the FY2021 National Defense Authorization Act formalized for all major defense programs.

Where Shadow EW Fits in a Crowded Market

BAE Systems is not the only defense company that has identified the small-platform EW gap. Rafael Advanced Defense Systems unveiled STORM SHIELD in May 2026 at the AOC Electronic Warfare conference in Helsinki. Rafael said STORM SHIELD targets the gap in UAV self-protection: “the absence of market available and combat-proven, active EW systems on unmanned platforms.” Anduril Industries rolled out Pulsar-L, a compact EW variant designed for signal disruption on small platforms, in April 2025 — the Pulsar-L launch announcement describes a system about the size of a shoebox already fielded in operations. Northrop Grumman demonstrated its Tactical Edge Electromagnetic Solutions system — which the company described as smaller than a business card — on robots, unmanned surface vessels, and drones at the Silent Swarm 2025 exercise. Thales debuted a mini-EW payload weighing under 5 kilograms (11 pounds) and drawing under 40 watts at the 2025 Paris Air Show.

What distinguishes Shadow EW’s positioning is less the technology itself — software-defined radio for small platforms is a widely pursued approach — and more the manufacturing claim. BAE Systems is asserting that Cedar Rapids, Iowa can produce Shadow EW in high volumes using established production lines, and that it has done the supply chain work to make that assertion credible. The company has positioned Cedar Rapids as the manufacturing hub for the product line, with software and design work running through Nashua, New Hampshire, where BAE’s existing EW engineering workforce is concentrated.

No independent validation of the manufacturing claim exists yet — Shadow EW 100 production does not begin until 2027, and no contracts have been publicly announced. The competitive test will come when the Air Force and other services translate their evident interest in affordable, modular EW for small platforms into actual procurement requirements.

Does It Cover Everything the Modern Threat Environment Demands?

Shadow EW’s highest-performance variant, the 500, covers frequencies from 10 megahertz to 18 gigahertz. That spans the vast majority of tactical radar threats, GPS bands, most communication frequencies, and the X-band and Ku-band radar systems used in many advanced air defense platforms.

It does not cover millimeter-wave frequencies above 40 gigahertz, which is where precision guidance seekers, active imaging systems, and some newer air defense radars operate. The defense industry has recognized this pressure: trade publications covering EW system design note that modern platforms face requirements to support frequencies beyond 40 gigahertz within the same constrained payload envelope that already has to cover the legacy 2-to-18 gigahertz range. Shadow EW’s frequency ceiling at 18 gigahertz means it addresses the most common current threats without covering the emerging upper end of the spectrum. Whether the 500-series or a future variant extends into millimeter-wave has not been publicly stated by BAE Systems.

The “collaborative effects” capability in the Shadow EW mission menu — the ability to deliver coordinated effects across multiple platforms — points toward the DoD’s larger Collaborative Combat Aircraft concept, in which autonomous or semi-autonomous drones operate alongside crewed aircraft in large-force employment. Shadow EW’s MOSA and open-architecture compliance means it is designed to interoperate in that construct, but realizing that capability in operational settings requires far more than a compliant payload.

Frequently Asked Questions

How does Shadow EW protect a drone that cannot carry a large EW system?

Shadow EW achieves compact EW capability through a division of computational labor. An FPGA (AMD-Xilinx Kintex KU15P) handles the deterministic, low-latency task of identifying radar signals by processing pulse descriptor words in real time. A commercial AI accelerator (NVIDIA Jetson Orin NX) handles threat classification using onboard machine learning inference. Together, in a 7.2-pound (3.3-kilogram) package drawing under 150 watts, they replicate functions that previously required much larger, heavier systems — though the size reduction comes with tradeoffs including reduced frequency coverage compared to larger platform EW suites.

Why are commercial chips like NVIDIA’s Jetson Orin NX appearing in a defense EW system?

The Department of Defense shifted toward commercial off-the-shelf (COTS) electronics beginning with the Perry Memo of 1994, which mandated COTS adoption wherever possible to reduce cost. For Shadow EW specifically, commercial AI chips deliver more compute per watt and more compute per dollar than custom defense silicon because they are produced in far higher volumes for commercial markets. The tradeoff is that commercial chips carry less environmental qualification than mil-spec parts (Shadow EW compensates with MIL-STD-810H system-level testing) and their supply chains run primarily through Asian foundries — a concentration risk the DoD has spent billions to mitigate through the CHIPS and Science Act.

Does Shadow EW address the fiber-optic drone threat that has neutralized conventional EW in Ukraine?

No EW system currently addresses the fiber-optic FPV drone threat directly, because those drones carry no radio link to jam. Shadow EW provides electronic support (detecting and classifying threats), electronic protection (self-defense jamming against radar-guided threats), and electronic attack (degrading adversary systems). It helps a drone equipped with Shadow EW survive in a contested electromagnetic environment. But a fiber-optic-guided adversary drone flying toward the same drone would be entirely unaffected by Shadow EW — it requires kinetic countermeasures to defeat.

When will Shadow EW be available, and who is the intended customer?

The Shadow EW 100, the compact UAV-focused variant, has a production start date of 2027 per BAE Systems’ own datasheet. The Shadow EW 120 and 500 production timelines have not been publicly disclosed. The intended customer base is U.S. and allied military purchasers looking to equip drones, legacy helicopters, and older fixed-wing aircraft with EW capability. No public contracts have been announced as of this article’s publication.