The FCC Wants to Pull "Military-Grade" Drones Off the Market. 661 Filings Later, Almost Nobody Agrees on What That Means.

The final report from docket-lens, generated 2026-09-05 and mirrored unedited from the repository's reports/report_2026.09.05_2.md. Its direct quotes were checked against the filings they cite before the report was written out.

Sometime last summer, a search-and-rescue captain in Utah, a Reuters editor in New York, a Massachusetts police drone unit, a Swiss inspection-robot maker and a Nebraska farmer all sat down to write the Federal Communications Commission about the same thing: a laser rangefinder, a heat-sensing camera, and what those parts say about whether a drone belongs in a war.

The occasion was PS Docket No. 26-189, a proposal from the FCC’s Public Safety and Homeland Security Bureau and its Office of Engineering and Technology to prohibit the continued importation and marketing of foreign-made drones that the agency calls “military-grade.” The proposal would reach equipment the FCC has already authorized for sale. It defines “military-grade” through seven capabilities: a takeoff weight of 55 pounds or more, aerosol dispensing, thermal imaging, LiDAR (a laser sensor that measures distance and can build 3D maps), docking stations, integration of a defense article, and swarming.

Because Chinese manufacturers such as DJI and Autel dominate the American commercial drone market, the practical effect would be to close off much of that market to new sales. The notice does carry an important limit, stated in the Commission’s own words: “while importation and marketing would be prohibited, this prohibition would not affect continued use or operation of already-purchased UAS or UAS critical components.” The listing rests on a determination that the equipment can “pose an unacceptable risk to the national security of the United States or the security and safety of United States persons.”

Comments were due September 2, 2026. The record that arrived is one of the most lopsided a technology docket has produced in years.

Who showed up

The docket contains 661 unique analyzed filings, and 680 filings counting near-identical copies. Of the unique filings, 484 oppose part of the proposal, 128 oppose all of it, 16 are mixed, 13 unclear, 12 purely procedural, and 8 are classified as supporting it. That last figure includes the Commission’s own public notice; seven outside filers argued for the prohibition.

The filers skew heavily toward people who fly for a living: 260 operators and pilots, 158 individuals, 71 dealers and service firms, 59 public-safety agencies, 38 mapping or survey companies, 21 manufacturers, 21 trade associations, plus academics, law firms and others. Form-letter volume is negligible by federal-docket standards. Only 19 filings were duplicates; the largest campaign, organized around an agricultural comment from Ezra Alan Woodliff of Maple Park Agridrones, drew 12 copies. Because that form was fillable, each signer wrote their own answers: Rich Uhrenholdt described cedar and brush overtaking rangeland that airplanes cannot safely reach, and Jay Gubert of Premier Outdoor Drone Services in Texas said his four aircraft cover more than 25,000 acres a season and support five employees.

Across the whole record, 568 filings offered civil-use evidence, 473 argued overbreadth, 304 raised safety dependence, 303 flagged definitional ambiguity, and 276 supplied prices, market figures or study data.

The argument almost everyone made

The Drone Service Providers Alliance, which says it represents more than 100,000 remote-pilot businesses, put the common objection in four words: “Dual-use does not mean military-grade.” The Law Enforcement Drone Association, a 5,000-member nonprofit, framed it the same way: “A technology does not become military-grade merely because it may have military utility.” Autel Robotics, one of the two Chinese manufacturers most directly affected, wrote: “Potential military utility is not military classification.”

DJI, represented by Cooley LLP, told the Commission the label itself is wrong, arguing the seven categories “describe ubiquitous commercial capabilities found in thousands of products sold for routine agricultural, public-safety, and recreational purposes.”

LiDAR: the category the record fought hardest over

More filings addressed LiDAR than any other criterion: 433, or 66 percent of the record. The reason is that the same three letters now cover two very different things. One is a survey instrument that builds centimeter-accurate point clouds. The other is a cheap fixed sensor that stops a small drone from flying into a power line in the dark, and it has spread across the consumer market in about eighteen months.

The Drone Service Providers Alliance listed the commercial uses first: surveying, construction, topographic mapping, forestry, mining, utility corridors, stockpile calculations, engineering models. “The fact that military organizations can use LiDAR for terrain mapping or autonomous navigation does not transform a surveying UAS into a military-grade aircraft.” It also noted the category would capture “a DJI Neo 2, which weighs a measly 151 gram, yet uses front facing LiDAR for obstacle avoidance for low-light conditions” , and asked that LiDAR be dropped as a stand-alone trigger.

The most detailed technical filing came from an unexpected quarter. Reuters News & Media, which flies small camera drones with staff journalists holding Part 107 certificates, filed on this item and no other. Its central observation: “Item (4) is the only category in the proposed definition that has no threshold. Every other category turns on weight, capability, purpose or design intent.” The forward LiDAR in its DJI Air 3S measures from 0.5 to 25 meters, returns a single pulse, cannot be aimed, produces nothing an operator can see, record or export, and does not work properly in bright light. “It cannot map, cannot be aimed and cannot be repurposed.” Reuters set that against a survey unit from the same manufacturer, the Zenmuse L3: 950 meters of range, up to 16 returns per pulse, 240,000 points per second, a three-axis gimbal, and a $17,400 price tag on a $10,450 airframe. “It is a safety interlock with an infrared emitter, in the same family of parts as a reversing sensor on a car.”

Reuters also traced how far down-market the part has traveled: obstacle-avoidance LiDAR appeared on a consumer camera drone in October 2024 and by April 2026 was fitted to aircraft weighing 151 and 249 grams and selling for a few hundred dollars. Because the Department of War has separately exempted drones of 150 grams or less as posing no unacceptable risk, Reuters observed of the resulting line: “One gram now separates the two conclusions.” Its proposed fix is carve-out language keyed to maximum range, fixed orientation, absence of point-cloud output and integration for obstacle avoidance. Without it, the agency warned, “the United States would become the only country in the world in which Reuters cannot fly the drones it uses everywhere else.”

Other filers proposed versions of the same line. Voliro AG, which builds industrial inspection drones with a rigidly mounted sensor under 300 feet of range, offered an eight-factor safe harbor based on weight, integration, range, angular sampling, marketing purpose and data control. Airborne Innovations urged the Commission to separate navigation and obstacle-avoidance LiDAR from targeting and terrain-denial units, and argued that stripping collision sensors out of small aircraft would make them less safe, including the risk of ruptured lithium cells after a strike. Skysense said the definition as written sweeps in ordinary laser rangefinder and autofocus modules and asked that it be limited to scanning sensors that produce point clouds. Flyability asked for a capability-and-purpose test rather than a presence test. Elevation UAV and DroneXL both told the Commission it was conflating short-range avoidance sensors with mapping payloads.

Public-safety and industrial users defended the mapping side of the technology. LEDA wrote that “public-safety agencies use LiDAR and other three-dimensional mapping technologies for crash reconstruction, crime-scene documentation, disaster assessment, structural analysis, and the preservation of evidence” , and noted those systems shorten road closures and reduce time spent in hazardous locations. The New York State Police made the same point about crash and crime scenes. Associated Builders and Contractors said that “Lidar components support topographic surveying, volumetric measurement, progress verification, clash detection, digital twins and inspection of structures and infrastructure.” Propeller Aerobotics described using it to see ground through vegetation on construction sites and likened it to a survey rover. The Marcellus Shale Coalition and the American Gas Association listed corridor mapping, pipeline inspection, geohazard identification and vegetation management. The Agricultural Retailers Association noted the same sensor is used both for crop canopy modeling and simply to keep aircraft from hitting things at low altitude.

The price evidence points one way. Minister & Glaeser Surveying, a small Washington state firm flying a DJI Matrice 400 with an L3 payload, said a compliant equivalent can cost up to ten times as much. Elevate UAS reported paying about $50,000 for a professional mapping aircraft against roughly $120,000 for a U.S.-built configuration that still underperformed it.

Thermal imaging, and the agencies that live on it

Thermal imaging appears in 410 filings, 62 percent of the record, and is where the public-safety bloc concentrated. Jon Beal of LEDA listed what his members do with it: locate missing, endangered or vulnerable persons, support nighttime searches, identify fire hotspots and monitor fire spread, assess hazardous-material incidents, and keep responders out of dangerous places.

The CEMLEC Drone Unit, a regional Massachusetts council drawing operators from more than 40 agencies, gave the argument a street-level form. It flies more than 60 aircraft, responds to over 500 deployments a year, and says roughly 35 percent involve missing persons, including autistic children and people with dementia. It also supplied one of the record’s sharpest price comparisons: a DJI M30T bought for roughly $9,600 against a quote of about $56,000 for a comparable American-made Skydio platform. “A five- or six-fold increase in acquisition cost is not simply a purchasing inconvenience, it can determine whether a small municipality is able to maintain a drone program at all.”

Similar filings arrived from Tualatin Valley Fire & Rescue, the Gresham police UAS team, the Lewistown and Blair police departments, and South Hadley police chief Jennifer Gundersen, who asked for a 24 to 36 month transition and federal funding if agencies are forced to replace fleets early. LEDA’s structural complaint applies to all of them: “A permission to operate an aircraft that can no longer be repaired or supported is not meaningful grandfathering.”

Commercial filers attacked the category on technical grounds. Airborne Innovations said 640x512 long-wave infrared microbolometers are now commodity parts made in several countries. DSLRPros argued that commercial imagers fall below the government’s own military sensor threshold in Category XII of the U.S. Munitions List. Flyability noted that Switzerland’s export-control authority determined its Elios 3 is not a dual-use item. Zeitview pointed out that thermal inspection of solar plants is effectively required by an international standard, IEC TS 62446-3. Unmanned Vehicle Technologies, which says it has equipped more than 2,400 public-safety agencies, called this classification the proposal’s most consequential error.

The counterweight came from Obsidian Sensors, a San Diego manufacturer building uncooled thermal cores on flat-panel glass and shipping cameras made in California. Its filing argues the dependency is exactly the problem.

Weight, spraying, docks and swarms

Category 2 appears in 231 filings and the 55-pound threshold in 157; the two travel together, because an aircraft carrying enough liquid to treat a field cannot weigh less than 55 pounds. The Agricultural Retailers Association put it plainly: “the fact that a drone can dispense an agricultural product does not by itself demonstrate that the drone possesses military capabilities or presents a unique national security threat.” The American Spray Drone Coalition reported more than 15,000 such aircraft in service. The American Mosquito Control Association added a wrinkle nobody else raised: “one of the two manufacturers that actively served this niche market announced its withdrawal from the mosquito control market in July 2026.” On weight alone, the Academy of Model Aeronautics objected that “A large model aircraft with a conventional radio receiver, servos, electronic speed controller, battery, propulsion system, and manually controlled flight system does not become military-grade merely because it reaches a particular weight.”

Docking stations, in 187 filings, drew the most incredulous responses. The FPV Freedom Coalition asked simply: “How on earth is a docking station a national security risk?” Zeitview estimated more than half its U.S. fleet falls into one or more categories and put its transition cost at roughly $1.5 million.

Category 7, in 181 filings, expressly names multi-UAS light shows. Creative Skies argued its aircraft are pre-programmed and do not sense or talk to each other: “Functionally, a drone light show fleet is a synchronized playback device, closer to a theatrical lighting console or a pixel display than to any autonomous swarm.”

Category 6, defense-article integration, is the outlier: only 92 filings addressed it, and many endorsed it. Ohio LEDA, the Blair Police Department, Raptor Maps and Legacy Technology Lab all said weapons integration, not sensors, is what the definition should target.

The case for the rule

The Foundation for Defense of Democracies argued that “While not every heavy-lift drone poses a threat, platforms that combine military-grade capabilities and foreign ties deserve stricter scrutiny due to their capacity to inflict significant harm on the United States.” It cited Ukrainian battlefield use of the Mavic 3 Thermal and Chinese export controls treating long-range radio links as security-sensitive.

Three domestic agricultural manufacturers holding Department of War Conditional Approvals, Ceres Air, Elevon Aerial and Exedy Drones, filed nearly identical comments backing the prohibition for heavy, dispensing and LiDAR-equipped foreign aircraft. The Agricultural Drone Initiative supported the ban while insisting that “each agricultural drone they legally purchased represents $50,000.00 or more in spent capital” and that farmers should not bear the cost. Obsidian Sensors went furthest: “we see no reason to allow any exceptions to the proposed rule as such exceptions will allow China to maintain a dangerous foothold on a key technology and capability that is vital to U.S. national security and economic interests” .

The most candid supporter was Swarm Defense Technologies, a Michigan manufacturer: “SDT discloses plainly that it competes with foreign producers of coordinated multi-UAS systems and would benefit commercially from the proposed prohibition.”

The numbers, and the law

Pilot Institute, a training company, contributed the largest dataset: a survey of 8,056 operators, 83 percent commercial. It reported 96.7 percent flying DJI aircraft, 70 percent flying all-DJI fleets, 97 percent of responding public-safety agencies flying DJI, and 76.6 percent expecting a significant to business-ending impact. On the FCC’s citation of new domestic manufacturing investment, it answered: “However, capital raised is not product shipped.” These are the filer’s own survey results, from respondents recruited through its student base.

The American Drone Data Protection Coalition estimated that replacing the 83,972 DJI drones registered with the FAA in 2023 alone would cost buyers more than $1.1 billion. Edison Electric Institute, the American Petroleum Institute, AFPM and four other energy trade groups jointly asked for a definition limited to systems specially designed for military purposes.

DJI’s filing is the record’s main legal challenge. It argues the December 2025 Covered List addition was itself unlawful, that the seven categories appear nowhere in the underlying national security determination, that the notice fails the Commission’s own requirement to identify affected devices and assess impact, and that “An agency cannot outsource its analytical obligation to the public it proposes to regulate.” DJI estimates its own “losses DJI estimates will exceed $1.5 billion” from the listing and set-asides. Autel adds that Section 302a of the Communications Act reaches radiofrequency interference, not weight or sensors.

What happens next

The notice proposes that importation and marketing cease 180 days after Federal Register publication. Almost no one accepts that timeline. Requests ranged from the Associated General Contractors, which asked for a 90-day grace period so construction firms could procure compliant aircraft without stalling projects, to Autel, DJI and Zeitview asking for 18 to 36 months, with several asking that any deadline be tied to demonstrated availability of equivalent alternatives.

Three organizations, AFPM, DRONERESPONDERS and the Commercial Drone Alliance, moved to extend the comment deadline. The Alliance noted that “Affording sufficient time to participate is all the more critical as the Public Notice does not provide an opportunity to submit reply comments.” The Bureaus denied all three requests, citing national security urgency, and the September 2 deadline held.

That leaves the FCC with a record in which 484 filers want the rule narrowed rather than killed. The single most common ask, from police departments, fire districts, surveyors, farmers and a wire service alike, is not that the prohibition disappear. It is that the Commission explain what a commodity heat sensor, or a 25-meter rangefinder on a 151-gram drone, has to do with a war.

Sources

Quoted filings in FCC PS Docket No. 26-189 (ECFS):