$49.1B defense tech funding in 2025 fuels DC RF engineer boom

Why Distributed Sensing Became the New Baseline

Guardian RF's palm-sized Scout sensors are now fused into the air defense picture at Vandenberg Space Force Base, a satellite launch and missile testing facility, under a $1.2 million Direct-to-Phase II SBIR from AFWERX. That deployment, alongside 33 U.S. airports and active combat zones in Ukraine, has made passive RF sensing the foundational layer of modern counter-UAS defense and catalyzed a specialized hiring boom for RF and DSP engineers in the Washington defense corridor.

Every drone that isn't fully autonomous betrays itself. It must talk to its controller, sending telemetry, receiving commands, streaming video, and that conversation radiates into the spectrum. Passive RF sensing doesn't transmit; it only listens. In doing so, it answers the one question no radar, camera, or acoustic array can reliably answer: where is the pilot standing?

The U.S. Department of Defense formalized this logic in its Joint Counter-small Unmanned Aircraft Systems Office strategy, which mandates a layered, fused detection architecture rather than reliance on any single technology. The threat set spans hobbyist quadcopters through Group 2 fixed-wing systems, varying in size, altitude, RF behavior, and flight profile. No one sensor covers it all. But passive RF catches the overwhelming majority of incursions at commercial facilities: consumer and prosumer aircraft with active control links, video downlinks, or, since the 2024 FAA mandate, Remote ID broadcasts carrying the drone's identity, position, and control-station location. All of it visible to a receiver that never reveals its own presence.

That covertness is operationally decisive. A radar emits to detect, announcing its position and operating frequency to any adversary with a spectrum analyzer. In contested environments, that emission is a liability — a beacon for anti-radiation munitions or a cue for electronic attack. Passive RF contributes no signature to the local air littoral. It powers on and operates in roughly 15 minutes, drawing little enough power to run indefinitely on a solar trailer's 400-watt array and battery bank; a surveillance radar generally cannot. The power math alone dictates the deployment sequence: passive RF first, radar only where the requirement justifies the cost and consumption.

The operator-location capability changes what a response team can do. A drone's control link often carries operator position data. When present, a capable passive RF system resolves a bearing to the controller as well as the aircraft; with multiple sensors in a mesh network, it triangulates an estimated operator location. In many scenarios, the person controlling the drone is just as much an actionable target as the aircraft itself. A detection architecture that locates both fundamentally expands the response envelope — interdiction, apprehension, or attribution become possible before the drone reaches its objective.

Modern signal processing has removed the old dependency on protocol libraries. Data-driven RF signature analysis treats emissions as distinct physical-layer patterns in the time-frequency domain, such as burst duration, hopping behavior, and spectral occupancy, bypassing the need for bit-level decoding. This allows reliable detection and classification even when communication protocols are proprietary, partially unknown, or constantly evolving, and it maintains robustness in congested RF environments and low-SNR conditions where standard demodulation fails. When combined with deep learning, the approach achieves sub-second classification at frame rates suitable for real-time cueing.

The doctrine is cumulative: passive RF as the persistent, low-signature foundation; radar or acoustic sensors to cover RF-silent or autonomous platforms; electro-optical/infrared for positive identification and engagement cueing; an AI-enabled command-and-control layer translating multi-sensor inputs into real-time threat awareness. Multi-sensor correlation improves detection confidence and resilience by validating threats across modalities, making it harder for an adversary to defeat the architecture by countering a single layer. But the foundation remains passive RF — not because it solves every threat, but because its failure modes don't overlap with the others, and it operates without ever announcing itself.

The Signal-in-Noise Problem: Why Counter-UAS Demands a New Breed of Engineer

The electromagnetic spectrum over a modern battlefield or a busy airport looks nothing like a textbook. Thousands of emitters (Wi-Fi, 5G, tactical radios, commercial drone links) crowd the same bands. A Group 1 quadcopter hovering at 200 feet transmits a control link that hops across dozens of frequencies per second, buried in urban multipath and thermal noise. Detecting that signal, classifying it, and geolocating both the drone and its operator before the aircraft reaches a sensitive perimeter is not a detection problem. It is a signal-processing problem at the edge of what current hardware and algorithms can do.

Frequency-hopping spread spectrum (FHSS) has become the default for commercial and military drone links precisely because it defeats narrowband scanning. The carrier jumps over a wide band in a pseudorandom sequence known only to the radio pair. Traditional super-heterodyne receivers, the backbone of defense RF for a century, rely on fixed intermediate-frequency filters that define a static channel bandwidth. That architecture cannot retune fast enough to follow a hop sequence, and its image-frequency and carrier-leakage artifacts (inherent to analog mixing) raise the noise floor exactly where weak drone signals live. As recently documented in defense RF literature, imperfect local-oscillator isolation and process-voltage-temperature variation make it impossible to maintain the 90-degree phase offset between I and Q channels, degrading image rejection in ways that software alone cannot fix.

The industry's answer is direct-conversion, zero-IF (ZIF) transceivers that integrate the entire RF, analog, and digital chain on a single CMOS die. Analog Devices' AD9361 and AD9371 exemplify the shift: they embed DSP blocks that run quadrature and carrier-leakage correction in real time across all process, frequency, and temperature corners. The AD9371 adds an on-chip ARM core for calibration control, an observation receiver for power-amplifier linearization, and a sniffer receiver for white-space detection — features that matter when a sensor the size of a palm must scan 100 MHz of spectrum continuously. Guardian RF's job posting states the work plainly: "Build and tune DSP detection and classification for drone RF links, improve track quality, and validate results against recorded field data."

Ukraine proved why the old playbook fails. By 2025 both sides fielded tens of thousands of electronic-warfare jammers simultaneously. Adversaries responded with FHSS links, fiber-optic control tethers that emit no RF at all, and AI-driven autonomous navigation that drops the operator link entirely. As the Polybolos Institute noted, "as drone autonomy increases, the effectiveness of traditional jamming decreases proportionally." Passive RF sensing remains the only layer that can see the operator — but only if the receiver can pull a frequency-hopping burst from a noise floor raised by thousands of friendly and hostile emitters, then fuse that detection with radar, electro-optical, and acoustic tracks to drive false-alarm rates below one percent. That fusion problem is now a stated Pentagon research priority.

The talent gap follows the physics. Designing a monolithic transceiver that meets SWaP targets while matching super-het linearity requires engineers who understand RFIC layout, DSP algorithm design, and the statistical behavior of FHSS signals in non-Gaussian clutter, all at once. The same skill set builds the cognitive radios that the Yole Group projects will push the military RF device market to nearly $3 billion by 2031 (11% CAGR), with the C-UAS RF segment alone exceeding $700 million. Defense-tech startups nearly doubled funding to $49.1 billion in 2025; Anduril added 1,000 employees in nine months. Engineers are leaving FAANG for 40–100% pay premiums to work on problems where the test range is a combat zone and the spec sheet is written by the adversary.

From Campus Hackathon to Operational Baseline

Three Georgetown physics majors, John Andrzejewski, Eli Kerstein, and Lucas Raskin, were studying signal processing when classmates Rasmus Dey Meyer and Nathaniel Salander invited them to a defense technology hackathon in El Segundo, California. In a 24-hour sprint, they coded proprietary signal-processing software to detect drones and their operators. They won. The project became Guardian RF.

Y Combinator accepted them into the Summer 2024 batch but told them they needed hardware to commercialize. During finals week, the trio pulled three all-nighters in empty Walsh Hall classrooms, crowding around blackboards to build their first sensors. Kerstein now serves as CTO leading product development; Raskin as CEO handling sales and business development; Andrzejewski as COO and head of customer success overseeing operations and sensor deployments.

The company raised $2.3 million from General Catalyst and Space Capital. By early 2026, Forbes had named the founders to its 30 Under 30 list for Transportation & Aerospace.

The first major defense contract came through AFWERX: a Direct-to-Phase II SBIR worth $1.195 million to address the Department of the Air Force's need to secure low-altitude airspace at operationally sensitive sites. The award placed Guardian RF's Scout sensors at Vandenberg Space Force Base, integrated with Picogrid's Expeditionary Command and Control Nodes. The deployment fuses data from multiple points across the base into a single operational picture and automatically cues existing camera systems to visually confirm potential threats.

Simultaneously, the sensors were being combat-tested in Ukraine by the Ukrainian military. Raskin has said the company sees its biggest opportunity in the U.S. public safety sector, but the Ukraine deployments provided something no lab could: validation against adversarial drones in an active electronic warfare environment.

Public safety traction followed. Guardian RF sensors now operate at 33 airports and World Cup venues across Florida, Massachusetts, and Texas, plus critical infrastructure sites including electric utility, LNG, and gas facilities. Law enforcement deployments span the Elk Grove Village, Palm Beach Gardens, and Greenwich police departments. Campus installations include Georgetown University, University of Miami, Baylor University, Iowa State University, and Horace Mann School. The Port of Detroit and MassDOT also run the system.

In December 2025, Guardian RF placed among the top four performers in the Defense Innovation Unit's Counter-sUAS Low-Cost Sensing challenge at USNORTHCOM's Falcon Peak 25.2 exercise — ten finalists culled from 115 submissions. The company received a $100,000 award alongside Hidden Level and Teledyne FLIR Defense. DIU's evaluation found the competing systems showed potential cost savings of 50–80 percent in total cost of ownership while meeting coverage and performance requirements.

The Scout sensor (portable, sub-10W power draw, LTE or LEO backhaul, 60-second deployment) covers 40 MHz to 6 GHz with 360-degree passive coverage and a 4–5 km fixed-site detection range in dense urban terrain. A fixed-site variant, Scout-X, adds weatherproofing, wired power, integrated mesh networking, and months of unattended operation. A Full Spectrum wideband SDR variant targets improvised and home-built FPV systems with signal fingerprinting across telemetry, control, and video links.

Guardian RF's platform converts passive spectrum observations into a queryable intelligence product that persists, correlates, and resolves across time and geography. The platform produces durable, queryable records of unmanned activity that support analysis, reporting, and inter-organizational use. Discrete sightings become patterns. Short-duration incidents become documented events. Local observations feed regional and national airspace awareness without centralizing sensor ownership.

The architecture reflects a structural mismatch in low-altitude security. Drone activity occurs locally (at a campus, a utility substation, a port terminal) while compliance, oversight, and enforcement responsibilities sit at higher levels of government and enterprise. Guardian RF's platform addresses this with a common data model that supports site-level operations, regional fusion, and enterprise-level analysis from the same underlying data. Local entities focus on detection and documentation. Higher-level organizations manage aggregation, compliance, and long-term intelligence functions.

At Vandenberg, this plays out in real time. Scout sensors feed the Expeditionary Command and Control Nodes; their infrastructure supplies modular power, secure networking, and edge computing, merging inputs from across the installation into a unified operational picture. The system can automatically cue those cameras to visually confirm potential threats, strengthening perimeter security for launch infrastructure and personnel.

Airport deployments follow a similar logic but with different stakeholders. The FAA's UAS Detection, Mitigation, and Response guidance requires airport operators to coordinate with Air Traffic Control, Airport Operations, TSA, and law enforcement when unauthorized drone activity occurs. Guardian RF's API-first design and standardized outputs integrate with existing security, aviation, and reporting workflows — enabling upward data flow without operational disruption at the local level. The platform's role-based access control across organizations means airport operations, TSA, and local police can each see what they're authorized to see from the same sensor network.

Port authorities face comparable coordination challenges. The Port of Detroit appears on Guardian RF's customer list alongside MassDOT, suggesting transportation agencies are adopting the platform for critical infrastructure monitoring. Ports combine expansive perimeters, dense RF environments, and jurisdictional complexity: federal, state, and local agencies all have equities. The platform's multi-site visibility and longitudinal intelligence retention let port security establish baselines, identify repetition, and support post-event analysis across weeks or months of operations.

Law enforcement deployments reveal the attribution challenge. Elk Grove Village Police Department, Palm Beach Gardens Police Department, and Greenwich Police Department all use Guardian RF. The platform's drone pilot attribution capability (geolocation and signal fingerprinting) supports investigative follow-up and accountability where real-time intervention is constrained. Universities including Georgetown, University of Miami, Baylor, Iowa State, and Horace Mann School round out the customer base, each treating low-altitude airspace as a routine security gap that mature ground security doesn't cover.

Integration remains the hard part. GAO reports on federal efforts to address unauthorized drone flights at airports highlight three persistent friction points: federal and local roles for responding to incidents, legal authorities for using detection and counter-drone technology, and FAA actions to plan for technology effects on drone integration. Guardian RF's approach (passive sensing with no emissions, cloud-native architecture with on-premise deployment options for classified environments, open integration via standardized outputs) attempts to navigate these constraints by staying in the detection-and-documentation lane while feeding data to authorities who hold mitigation authority.

The SAFER SKIES Act, which Guardian RF tracks on its policy page, would expand counter-UAS authorities for state and local law enforcement. Until legislation catches up, the intelligence layer is the product: persistent visibility, attributed detections, pattern recognition across sites, and a data model that scales from a single police department to a regional fusion center without re-architecting the sensor network.

From a hackathon win to Vandenberg's launch pads, Ukrainian frontlines, and 33 U.S. airports in under two years — the trajectory illustrates how fast passive RF sensing has moved from campus project to operational baseline.

The DC Talent Crucible: How the Defense Boom Is Rewiring the Local Market

The Washington, D.C. region (Northern Virginia, Maryland, and the District itself) has long been the gravitational center of U.S. defense contracting. What changed is the velocity and composition of the money flowing through it. Venture capital investment in defense technology reached $49.1 billion in 2025, nearly doubling from the year before. Equity funding alone more than doubled to $17.9 billion. In the first half of 2026, defense tech startups raised somewhere between $12.3 billion and $35.6 billion depending on which provider's sector definition you use — a nearly threefold spread that Axis Intelligence Research tracks as the Defense Funding Reconciliation Index.

That capital is converting directly into headcount. Saronic, valued at $9.25 billion after a $1.75 billion Series D, expanded its Austin headquarters beyond 500,000 square feet and opened new hubs in San Diego and Washington, D.C., pushing total headcount past 1,300. Guardian RF, the Y Combinator S24 startup building passive RF drone detection, sits in the District with 15 employees and two open engineering roles — a microcosm of the broader scramble. LinkedIn lists more than 11,000 defense contractor jobs in the DC-Baltimore area today; Indeed shows 883 postings specifically for RF frequency roles. The Pentagon requested a $961.6 billion topline for fiscal 2026, with $848.3 billion in discretionary authority. Of that, $295.3 billion targets weapon-system accounts, including $13.4 billion earmarked for autonomous and remotely operated systems across air, land, and sea.

The talent market has tightened accordingly. Engineers are leaving FAANG roles for 40–100% pay premiums, a shift Jobs by Culture calls "the most aggressive hiring shift in a generation." RF engineering (always a specialized corner of electrical engineering) now commands a clearance premium on top of already elevated base bands. RF-careers.com's 2026 salary guide breaks compensation by experience level, specialty, security clearance tier, and location; irecruit.co's analysis overlays federal wage series with employer-published bands from defense-tech boards and ClearanceJobs data. The bidding wars for DSP talent with experience in software-defined radio and spectrum sensing are fierce.

Proximity to the Pentagon remains a structural advantage no other hub replicates. Georgetown University now requires a physics course co-designed by Guardian RF co-founder Lucas Raskin and student Rowan O'Sullivan (C'25), aimed at funneling talent directly into defense-tech entrepreneurship. But the market has constraints. Telework has normalized across the contractor community post-pandemic, yet classified programs often require in-person presence in SCIFs, which limits how much of the DC defense RF market is genuinely flexible on location. That geographic tether keeps the talent pool local and the competition zero-sum.

The macro backdrop reinforces the trend. World military expenditure hit $2.887 trillion in 2025, up 2.9% in real terms, with European NATO members spending a combined $559 billion — Germany's expenditure grew 24% year-over-year to $114 billion. European venture capital follows European budgets with roughly a one-year lag. Meanwhile, the NATO Innovation Fund reports VC investment in deep-tech defense, security, and resilience startups grew from $159 million in 2014 to $5.2 billion in 2024, a 32-fold increase in a decade. Manufacturing-focused defense investment rose to $4.7 billion across 39 deals in 2025, nearly doubling year over year. Venture exits in the sector surged to a record $54.4 billion in 2025, driven largely by acquisitions including Nvidia's €20 billion purchase of Groq.

For these specialists in the DC corridor, the signal is clear: the passive RF sensing layer that Guardian RF and its peers are scaling (from Vandenberg Space Force Base to 33 U.S. airports to combat zones in Ukraine) has become a procurement priority. The money, the contracts, and the hiring plans are all aligned. The only bottleneck left is people who can build it.

How Drone Operators Are Evading Passive RF Detection

Passive RF detection works because most drones talk. They maintain a radio link to a pilot, and that link betrays both aircraft and operator. But the assumption that every drone emits is collapsing. In Ukraine, the shift began with a technology older than the operators using it: fiber-optic guidance. The Soviet Union fielded wire-guided Sagger and Swatter antitank missiles in the 1960s. The principle is identical — spool a physical cable behind the vehicle, carry commands and video through glass instead of air. No radio emissions means no RF signature to detect, no signal to jam. As of early 2026, more than 35 Ukrainian manufacturers produce fiber-optic FPV drones, and Russian front-line units have adopted them at 30 to 50 percent rates. Chinese shipments of optical fiber to Russia jumped from roughly 119,000 miles in May 2025 to 328,000 miles by August, the material feedstock for an estimated 10,000 Russian FPV launches per day. Ukrainian medevac teams reported during the Kursk offensive that fiber-optic drones were monitoring all routes, collapsing logistics despite RF sensors declaring corridors clear. In October 2025 a Russian fiber-optic drone struck Kramatorsk, 19 kilometers behind the front line. Ukraine's Fold company has pushed fiber range to 100 kilometers using larger airframes, though standard FPV-class drones under 10 kilograms remain limited to 10 to 20 kilometers by cable weight.

Autonomy closes the gap a different way. Russian drones now sever their command link roughly five kilometers from target and fly the final approach on pre-programmed waypoints or onboard vision. The RF trail simply ends. Operators who understand detection theory exploit it deliberately: minimize transmit power, point directional antennas away from likely sensor positions, hop to unusual bands, or switch to RF-silent modes entirely. DJI's OcuSync 4.0, standard on the Air 3, Mini 4 Pro, and Matrice 4 series, already hops dynamically between 2.4 GHz and 5.8 GHz based on interference. In July 2026 Ukrainian radio specialist Serhii Flash documented the Russian Boomerang FPV moving its video channel above 6 GHz, carrying an antenna operating in the 6 to 7.2 GHz range, outside many legacy detector bands. DIY flight controllers running ArduPilot or PX4, LoRaWAN links, 900 MHz mesh networks: each produces signatures that commercial signature libraries have never cataloged. Modified Shahed variants now carry Starlink terminals with phased-array antennas pointed skyward; ground-based jamming arrives from below, outside the main lobe. The control link never enters the spectrum where RF sensors watch.

The operational consequence is blunt. Facilities that rely exclusively on electronic defenses believe they are protected until a fiber-optic drone passes through undetected and unaffected. Zero alarm warnings — because there is no transmission for an RF scanner to detect, there is frequently no warning at all. NATO recognized the gap in April 2025 with an urgent Innovation Challenge seeking detection solutions for fiber-optic FPV drones, specifying a required detection range of only 300 to 500 meters — a tacit admission that current systems see nothing at tactically useful distances. The counter-UAS market, projected at $9 to 20 billion by 2030-2034, still treats RF detection as a core component, but the component alone is no longer sufficient.

The response is forced fusion. Radar catches the airframe regardless of emissions. Acoustic sensors hear motor and propeller noise. Electro-optical and infrared cameras confirm visually and identify payloads. Sensor fusion platforms (DedroneTracker.AI, DroneShield DroneSentry, Sentrycs Horizon Engine, DroneShield RFAI, Ukraine's Kara Dag handheld) combine these inputs into unified pictures. The SAPIENT protocol has emerged as an interoperability standard; DroneShield implements it across all sensors for NATO compatibility. Kinetic layers enter the stack: Drone Round's specialized counter-drone ammunition, Epirus Leonidas high-power microwave (the Marine Corps ExDECS variant delivered April 2025, explicitly documented to defeat "autonomous and fiber-optic guided systems"), the Army's DE M-SHORAD achieving 100 percent success against 15 target drones in June 2024 testing. At a U.S. government test range in January 2026, Epirus disabled a fiber-optic guided FPV drone with the Leonidas VehicleKit — the first known instance of weaponized electromagnetic interference defeating a fiber-optic drone. D-Fend EnforceAir performs RF cyber takeover instead of jamming: identify the protocol, seize the link, land the drone safely. Contra Drone's D-JACK attacks the protocol directly. The Army's $49 million other transaction agreement with ThinKom for Alecto, a mobile high-power microwave effector on a VICTS antenna, and the Marine Corps' $11 million HAVOC contract to Epirus for an autonomy-enabled successor to ExDECS (both awarded in August 2026) signal where procurement is moving.

Software-defined electronic warfare with AI adaptation compresses defender innovation cycles from "months to years" to "weeks to months." Polarimetric thermal systems identify the physical materials of a drone (carbon fiber, polymers, metal alloys) rather than relying on emissions, detecting targets up to 50 percent more effectively than standard cooled mid-wave infrared sensors. The layered system-of-systems approach the Air Force now mandates (ground and airborne sensors, RF, radar, acoustic, EO/IR, electronic warfare, directed energy, guns, missiles, low-cost interceptors) is not doctrinal preference. It is the only architecture that survives contact with an adversary who has already stopped transmitting.

A $3.3 Billion Horizon: Market Projections and the Future of Multi-Sensor Counter-UAS

The global drone detection market was valued at $694.6 million in 2024 and is on track to hit $2.8 billion by 2030, a 28.9 percent compound annual growth rate. GMI Research projects $8.4 billion by 2034 at the same CAGR. The numbers differ in absolute terms but agree on the trajectory: detection is the fastest-growing segment of a counter-UAS market pegged at $6.6 billion in 2025 and forecast to reach $20.3 billion by 2030.

North America held the largest share in 2024 at $246.9 million, driven by defense spending and FAA-driven airport deployments. Asia-Pacific is accelerating faster, with a 31.1 percent CAGR through 2034, led by China, India, and Japan. Europe follows at 27.1 percent, fueled by NATO procurement programs — Germany's TYTAN ramp alone targets 3,000 interceptors per month. The U.S. counter-UAS market specifically is forecast to grow from $433 million in 2025 to $1.1 billion by 2035 at a more modest 9.7 percent CAGR, though the FY27 DAWG budget of $54.6 billion and the July 2026 state-and-local defeat procurement rule could revise that upward.

Passive RF sits at the center of this expansion because it solves the first-order problem: finding the operator. But the research is unanimous that RF alone cannot sustain the growth curve. The DIU's Low-Cost Sensing challenge tested finalists across radio frequency passive detection, active radar, acoustic sensing, optical and infrared systems, and hybrid approaches. MatrixSpace won the $500,000 top prize; Guardian RF, Hidden Level, and Teledyne FLIR each took $100,000 as runners-up. The winning solutions were not single-modality. They fused them.

GMI Research identifies the hybrid multi-sensor segment as the fastest-growing, at 30.9 percent CAGR through 2034. Fixed integrated systems and portable systems each exceed 30 percent. The logic is operational: fiber-optic control links emit no RF. Autonomous GPS-waypoint flights emit no RF. Frequency-hopping spread spectrum and directional satellite links bury the signal in noise. Acoustic arrays catch the propeller noise that RF misses. Radar catches the airframe that acoustics miss at range. EO/IR cameras confirm what the others only infer. AI-driven classification at the edge (pushed by the same DIU challenge requirements) turns raw sensor streams into a single track file an operator can act on.

The FAA's Remote ID mandate, effective March 2024, adds a cooperative layer: compliant drones broadcast ID and location. AUVSI estimates that alone will lift RF-based detection adoption by 35 percent. But non-compliant actors — the ones that matter for defense — don't broadcast. That gap is why the market is moving toward modular, AI-driven sensor fusion platforms that incorporate radar, RF, EO/IR, and acoustic seamlessly, with edge computing moving processing from cloud to sensor node. Next-generation specs already call for autonomous neutralization, 5G/IoT-enabled distributed networks, and cloud-based command platforms.

The $2.8 billion detection figure is a floor. The real market is the integration layer that makes passive RF useful when the spectrum goes dark. The foundational layer is deployed, the procurement priority is funded, and the sole remaining bottleneck is the expertise to build it.

Working in frontier tech? Zero G Talent tracks the openings: see every open ASML role, browse frontier tech jobs, openings at Stripe, and the people building the field.