Drone and Robotics OEM FAQ

What does "NDAA compliant" mean for a drone?

“NDAA compliant” usually means the seller says the drone and its key parts avoid the sources that U.S. defense laws bar, mainly China, Russia, Iran and North Korea. The main laws are FY2020 NDAA §848 and the American Security Drone Act in the FY2024 NDAA. The phrase is a sourcing claim, not a license or a safety rating. This page is general information checked in September 2026, not legal advice.

Is "NDAA compliant" an official government certification?

No. There is no government “NDAA compliant” certificate or stamp. The term is almost always a maker’s own statement about where its drone and components come from. The closest official check is the Blue UAS list, now run by the Defense Contract Management Agency (DCMA), which vets specific models and parts. Treat any NDAA claim as something to verify with documents, not as a third-party approval.

What is NDAA Section 848?

Section 848 of the FY2020 National Defense Authorization Act bars the Department of Defense from buying or operating drones made in a covered foreign country or by a company domiciled there. It also reaches drones that use certain parts, software or data services from those countries. It was amended by FY2023 NDAA §817 and sits as a note to 10 U.S.C. 4871. It binds DoD and its contracts, not private buyers directly.

Which countries count as covered foreign countries under the drone NDAA rules?

Under Section 848, as amended in 2022, the covered foreign countries are China, Russia, Iran and North Korea. The law also names DJI and its subsidiaries and affiliates as a covered drone company, along with drone makers on the Consolidated Screening List or under unmitigated control by a covered country. The American Security Drone Act uses a different test, based on a federal list of covered foreign entities.

Which drone components does Section 848 cover?

Section 848 covers flight controllers, radios, data transmission devices, cameras and gimbals made in a covered foreign country or by a company domiciled there. It also covers ground control systems or operating software developed there, and network connectivity or data storage located in or run from there. That is why a drone assembled in the U.S. can still fail the test if its camera or radio comes from China. See our drone supply chain partners guide.

Can a Department of Defense contractor use DJI drones?

Generally no, for defense work. Since October 1, 2024, Section 848 bars DoD from entering, extending or renewing a contract with a company that operates equipment from a covered drone company, such as DJI, while performing a DoD contract. Narrow exemptions exist for counter-drone testing and intelligence work, and the Secretary of Defense can issue written waivers. Contractors should check their contract clauses with counsel, since this is not legal advice.

What does an "NDAA compliant" label not tell you?

It does not tell you the drone has FCC authorization, is on the Blue UAS list, has passed a cybersecurity review, or meets Buy American or TAA rules. It also says nothing about flight safety or build quality. NDAA sourcing, Blue UAS status, FCC status and trade rules are separate checks, and a drone can pass one while failing another. Ask for each one in writing. The NDAA and Blue UAS requirements guide covers them side by side.

How can I check a supplier's NDAA compliance claim?

Ask for a bill of materials that lists the maker and country of origin for each covered part: flight controller, radios and data links, cameras, gimbals, ground station and software. Ask where firmware is written and where any cloud data is stored. An ODM such as Aerora, which describes its products as NDAA-compliant, should be able to show that documentation for the specific configuration you are buying, as any supplier should.

What is the American Security Drone Act?

The American Security Drone Act of 2023 (ASDA) is Subtitle B of Title XVIII of the FY2024 NDAA, sections 1821 through 1833. It extends drone restrictions beyond the Pentagon to every federal agency. It bars agencies from buying or operating drones made or assembled by covered foreign entities, and bars federal funds from paying for them. Sections 1823 through 1826 carry the main prohibitions. The law was enacted on December 22, 2023.

What is a covered foreign entity under the American Security Drone Act?

Under ASDA Section 1822, a covered foreign entity is one on a list kept by the Federal Acquisition Security Council and published in SAM. The list can include firms on the Consolidated Screening List, firms under foreign extrajudicial direction, firms DHS finds pose a national security risk, firms domiciled in or influenced by China, and their subsidiaries and affiliates. Unlike Section 848, the test turns on a published entity list rather than a fixed country list.

What does Section 1823 of the FY2024 NDAA prohibit?

Section 1823 bars federal executive agencies from buying any drone manufactured or assembled by a covered foreign entity, including its communication links and control components. It took effect when the law was enacted on December 22, 2023. Section 1826, also effective immediately, bars using government purchase cards to buy such drones. Limited exemptions exist for DHS, DoD, State, the Attorney General, DOT, NTSB and NOAA, and agency heads can seek case-by-case waivers.

When did the federal ban on operating Chinese drones take effect?

Section 1824 of the FY2024 NDAA took effect on December 22, 2025, two years after enactment. From that date no federal department or agency may operate a drone manufactured or assembled by a covered foreign entity. The ban also covers drones used through contracted drone services, so a contractor flying a covered drone for an agency is caught too. National-interest exemptions and case-by-case waivers are narrow.

Can state and local agencies use federal grant money to buy DJI drones?

Not if the drone comes from a covered foreign entity. Section 1825 of the FY2024 NDAA, effective December 22, 2025, bars federal funds awarded through a contract, grant or cooperative agreement from paying for, or supporting the operation of, such drones. That reaches police, fire and utility programs funded with federal money. Agencies should confirm with the grantor how a specific program applies the rule. This is general information, not legal advice.

Does the American Security Drone Act expire?

Partly. Section 1833 says sections 1823, 1824 and 1825 (the purchase, operation and federal-funds bans) stop having effect five years after enactment, which works out to December 22, 2028. The purchase-card ban in Section 1826 has no sunset in that section. Congress could extend or change these rules before then, so check the current text before relying on the date.

What is FAR clause 52.240-1?

FAR 52.240-1 is the federal contract clause that carries the American Security Drone Act into contracts. Its title is “Prohibition on Unmanned Aircraft Systems Manufactured or Assembled by American Security Drone Act-Covered Foreign Entities.” It first took effect through an interim rule on November 12, 2024. It covers buying, operating and using federal funds for covered drones during contract performance. Check which version of the clause your contract includes.

What's the difference between NDAA Section 848 and the American Security Drone Act?

Section 848 applies to the Department of Defense and its contractors and looks at named countries plus specific parts, software and data services. The American Security Drone Act applies to all federal agencies and federal funds, and looks at whether a covered foreign entity made or assembled the drone. A product needs to be checked against both if it will be sold into both defense and civilian federal markets.

What is Section 1709 of the FY2025 NDAA?

Section 1709 of the FY2025 NDAA gave a national security agency one year to decide whether communications and video surveillance equipment from DJI and Autel Robotics, and their affiliates and partners, posed an unacceptable risk. If no agency acted, the FCC had to add that equipment to its Covered List. The FCC added it on December 22, 2025, the same day it added foreign-produced drones in general.

What does Section 162 of the FY2025 NDAA require?

Section 162 directs the Department of Defense to map risks in the small drone supply chain and build resilience with domestic and allied parts. DoD must fully take apart a DJI drone, or a similar drone from a covered country, and repeat that at least every three years until 2034 to build a component risk taxonomy. It also calls for a risk framework and a sourcing strategy. It bans no product by itself. Related: biggest drone supply chain risks.

What is the Blue UAS Cleared List?

The Blue UAS Cleared List is the Department of War’s list of complete drone systems that have been vetted for cybersecurity and supply chain risk and approved for defense purchase and operation. Clearance applies to the specific model and configuration listed, not a whole product family. The Defense Innovation Unit built the list, and DCMA now runs it. When DIU announced the handoff, the program listed 39 systems and 165 components from 81 companies.

What's the difference between the Blue UAS Cleared List and the Blue UAS Framework?

The Cleared List names whole drone systems approved for defense use. The Framework lists vetted components and software, such as flight controllers, radios, cameras and gimbals, that builders can use to design new drones. Using Framework parts can simplify a later review, but it does not put your drone on the Cleared List. For its Covered List exemption, the FCC treats both lists together as the “Blue UAS Cleared List.” See drone flight controllers.

Who runs the Blue UAS program now?

The Defense Contract Management Agency (DCMA) runs it. The Defense Innovation Unit announced on December 3, 2025 that the Blue UAS list was moving to DCMA’s Unmanned Systems Experimental Command (US-X) in Palmdale, California. DIU stays involved as a partner on standards and checklists. The current list and submission portal are at bluelist.dcma.mil, and older links to the DIU site point there.

How does a drone or component get on the Blue UAS list?

Published pathways include sponsorship by a Department of War customer, competitive evaluations, and review by a Recognized Assessor, a third party that checks the product against NDAA requirements and reports through the Blue List portal. Submissions typically need company details, a product description, a hardware and software bill of materials, and critical-component documentation. There is no published fixed timeline or fee, so ask DCMA or an assessor for current terms.

What is a Blue UAS Recognized Assessor?

A Recognized Assessor is a third-party organization authorized to evaluate a vendor’s drone or components against Blue UAS and NDAA requirements. The vendor hires the assessor, and the assessor submits its report through the Blue List portal for DCMA review. AUVSI’s Green UAS program was adopted as one such route in 2025. Assessor fees are set by each organization and are not published by DCMA.

If I build a drone with Blue UAS Framework parts, is my drone Blue UAS cleared?

No. Framework parts are vetted building blocks, not a pass for the finished drone. The complete system, including its other components, software and data handling, still needs its own review to appear on the Cleared List. Using Framework parts can shorten that review and help with NDAA sourcing. Keep records of part numbers and firmware versions, since clearance is tied to a specific configuration.

What is the FCC Covered List?

The FCC Covered List names communications equipment and services that national security agencies have found to pose an unacceptable risk to the United States. It is kept under the Secure and Trusted Communications Networks Act of 2019. Equipment on the list cannot receive new FCC equipment authorization, and most electronics need that authorization before they can be imported, marketed or sold in the U.S. The FCC can add items only on a national security agency’s determination.

What did the FCC do to foreign-made drones in December 2025?

On December 22, 2025, the FCC added all drones and drone critical components produced in a foreign country to its Covered List (Public Notice DA 25-1086). It acted on an Executive Branch national security determination. The rule reaches foreign production in general, not just China. New foreign-produced models cannot get FCC authorization unless an exemption applies, while models authorized earlier are not affected by that notice.

What counts as a UAS critical component under the FCC rules?

The national security determination behind the December 2025 listing says UAS critical components include, but are not limited to, data transmission devices, communications systems, flight controllers, ground control stations and controllers, navigation systems, sensors and cameras, batteries and battery management systems, and motors, along with their software. Because the list is open-ended, suppliers of other parts should not assume they fall outside it.

Which drones are exempt from the FCC Covered List?

As of September 2026 there are four exemptions. Items on DCMA’s Blue UAS Cleared List or Framework are exempt until January 1, 2028. Items that qualify as “domestic end products” under the Buy American standard are exempt until January 1, 2028. Devices granted a Conditional Approval by the Department of War or DHS are exempt. Foreign-produced “toy drones,” as defined in the national security determination, were removed in June 2026.

What is FCC Public Notice DA 26-761?

DA 26-761 is the FCC notice of July 21, 2026 that updated the drone exemptions on the Covered List. It extended the Blue UAS and Buy American exemptions from January 1, 2027 to January 1, 2028, based on a new Department of War determination. It also removed the December 31, 2026 end date from Conditional Approvals, so they last as long as the holder keeps to its onshoring plan and passes updated vetting.

What is an FCC Conditional Approval for a drone?

A Conditional Approval is a finding by the Department of War or DHS that a specific foreign-produced drone or component does not pose unacceptable risk, which exempts it from the Covered List. Applicants must commit to an onshoring plan for manufacturing critical components. The approval ends if the holder breaks that plan or made false statements. Early approvals in March 2026 included SiFly’s Q12. For offshore-built products not on the Blue UAS list, Aerora’s included, this is the exemption route to ask about.

Does the Buy American exemption from the FCC Covered List have an end date?

Yes. Under DA 26-761, drones and critical components that qualify as “domestic end products” under the Buy American standard (48 CFR 25.101(a)) are exempt until January 1, 2028. The date comes from a Department of War determination and could be extended, shortened or replaced. Products that rely on this exemption should plan for what happens after that date, such as a Conditional Approval or Blue UAS listing.

Can I still fly my DJI drone in the U.S.?

Yes, for most private and commercial users. The FCC has said its Covered List actions do not restrict continued use of drones people lawfully bought, and even its pending proposals would not affect use of drones already purchased. The main exceptions are federal agencies and federally funded programs under the American Security Drone Act, plus some state and local agencies bound by state rules. FAA rules such as Part 107 and Remote ID still apply.

Can stores still sell existing DJI drone models?

For now, yes. Under the FCC’s Covered List rules, the December 2025 action applies to new device models, so retailers could keep importing and selling models the FCC authorized earlier. That could change: the FCC has used a 2025 rule to seek comment on stopping import and marketing of some previously authorized covered drones. Check the FCC Covered List page before stocking up. This is general information, not legal advice.

What is the FCC's "military-grade" drone proposal?

In Public Notice DA 26-758 (July 21, 2026, PS Docket 26-189), FCC bureaus proposed stopping continued import and marketing of previously authorized foreign-produced drones and parts that are “military-grade.” The categories are drones of 55 pounds or more, spray drones, drones with thermal or lidar sensors, docking stations, drones built around a defense article, and swarming drones, including light-show fleets. Comments closed September 2, 2026, and it was still a proposal in late September. Payload makers, Aerora included (thermal and lidar options), should watch it.

What is FCC Public Notice DA 26-742?

DA 26-742, released July 17, 2026 in PS Docket 26-184, proposes to stop continued import and marketing of previously authorized covered drones and Section 1709 equipment from nine named companies and their affiliates, including XAG. Like DA 26-758, it would exempt federal government use and commercial testing and product development, and it would not stop use of equipment already bought. It was a proposal, not a final order, as of September 2026.

Does the FCC Covered List also apply to robots?

Yes, since July 28, 2026. The FCC added “advanced robotic devices,” defined as mobile robots such as humanoids and quadrupeds, produced in any foreign country, to its Covered List. New models cannot get FCC authorization unless the Department of War grants a Conditional Approval. As with drones, existing devices can still be used, and models authorized earlier can still be sold. Robotics OEMs using an offshore partner, Aerora included, should ask about FCC status early. See robotic autonomy hardware partners.

If a drone is NDAA compliant, does it automatically pass FCC rules?

No. NDAA rules govern federal purchases, operations and funds. The FCC Covered List governs whether a new drone can be authorized for import and sale at all. A drone made in an allied country can meet NDAA sourcing and still be foreign-produced under the FCC rule. A buyer choosing any partner that builds offshore, Aerora included (Vietnam), should ask for the product’s FCC authorization and exemption status.

Does the FCC ban cover drones made in Vietnam, Taiwan or other friendly countries?

Yes. The December 2025 listing covers drones and critical components produced in any foreign country, not just China. A product built in Vietnam, Taiwan, Europe or elsewhere needs an exemption (Blue UAS, a Conditional Approval, or Buy American status) or an earlier FCC authorization. Aerora manufactures in Vietnam, so buyers should ask Aerora, like any offshore builder, for each product’s FCC authorization and exemption status. Compare UAV contract manufacturers.

What is the Buy American Act domestic content threshold for drones?

Under FAR 25.101, a manufactured end product is a “domestic end product” if it is manufactured in the United States and the cost of its domestic components exceeds 65% of the cost of all components for items delivered from 2024 through 2028. The threshold rises to 75% for 2029 and later. The FCC uses this same Buy American standard for its drone exemption, so the math matters beyond federal sales.

Can a drone built outside the U.S. qualify as a Buy American domestic end product?

Generally no. The FAR test requires the end product to be manufactured in the United States as well as meeting the domestic component threshold. A drone assembled in Vietnam, Mexico or Europe would not qualify, even with mostly U.S. parts. An ODM such as Aerora, which manufactures in Vietnam, is a case where buyers should ask which FCC exemption, if any, a given product relies on. See Aerora’s ODM services.

What is TAA compliance for drones?

TAA compliance means a product meets the Trade Agreements Act country-of-origin rules for federal contracts covered by trade agreements, including GSA Schedules. The product must be made in, or substantially transformed in, the United States or a “designated country” listed in FAR 25.003. China and Vietnam are not on that list, while countries such as Japan, Taiwan, South Korea and the United Kingdom are. The TAA applies to supply contracts at or above set dollar thresholds.

What's the difference between Buy American and the Trade Agreements Act?

The Buy American statute favors domestic end products and applies to many federal purchases below trade-agreement thresholds. The Trade Agreements Act applies to larger contracts, $174,000 or more for WTO GPA supply contracts under current FAR 25.402, and lets agencies buy from designated partner countries on equal terms. Buy American asks where components come from. TAA asks where the product was made or substantially transformed.

Are drones controlled under ITAR or EAR for export?

It depends on design and capability. Drones specially designed for military use and described on the U.S. Munitions List, Category VIII, fall under ITAR, run by the State Department. Most commercial and dual-use drones fall under the EAR, run by the Commerce Department’s Bureau of Industry and Security, often under ECCN 9A012 or 9A991, while some military-modified drones fall under ECCN 9A610. A commodity jurisdiction request can settle an unclear case.

What changed in U.S. drone export controls in 2026?

A Bureau of Industry and Security rule effective August 13, 2026 eased EAR controls on many drones. It raised the endurance threshold for national security controls from 30 minutes to 3 hours, dropped wind-gust tolerance as a control factor, and removed national security controls on certain parts for lower-endurance drones. Military end-use and end-user controls still apply. The rule did not change ITAR controls on drones listed on the U.S. Munitions List.

Do drone components need an export license?

Some do. Parts “specially designed” for an ITAR-controlled drone are generally ITAR-controlled too, and parts for EAR-controlled drones take their own classification, which may be 600-series, 9A012-related or lower. Sensors, thermal cameras and navigation parts often carry separate controls. Classify each part before shipping it abroad or sharing technical data with foreign nationals, which can be a “deemed export.” A U.S.-led ODM with offshore production, such as Aerora, has to manage this too. This is not legal advice.

What do I need to fly a drone commercially in the U.S.?

Under FAA Part 107, drones under 55 pounds can fly for business if the pilot holds a Remote Pilot Certificate, earned by passing the FAA knowledge test, and completes free online recurrent training every 24 calendar months. Each drone must be registered, at $5 for three years, and meet Remote ID. Standard limits include flying at or below 400 feet and within visual line of sight unless the FAA grants a waiver.

What is Remote ID and does my drone need it?

Remote ID is the FAA rule that makes drones broadcast identification and location data in flight, like a digital license plate. Registered drones need it, and FAA enforcement began March 16, 2024. You can comply by flying a Standard Remote ID drone, attaching a Remote ID broadcast module, or flying without Remote ID only inside an FAA-Recognized Identification Area. Drones flown with a broadcast module must stay in visual line of sight.

Do drones need FAA type certification?

Most do not. Routine Part 107 flights under 55 pounds need no type certificate. Operators planning larger drones, package delivery or other advanced operations may need one, which the FAA issues under its “special class” process in 14 CFR 21.17(b), using durability and reliability testing. Matternet’s M2 became the first drone type certificated this way in 2022. The FAA’s Part 108 BVLOS rule, proposed in August 2025, was not final as of September 2026.

Which states ban Chinese drones for government agencies?

Several states limit what their own agencies may buy or fly. Florida requires agencies to use drones from a state-approved manufacturer list. Texas lists DJI and Autel on its prohibited technologies list for state agencies. Nevada’s SB 11, effective January 1, 2025, limits public safety agencies, and Arkansas passed a 2023 law phasing out Chinese drones for law enforcement. Rules vary by state, so check current state law.

Does Florida ban DJI drones?

For government agencies, effectively yes; for private users, no. Florida Statute 934.50, added by SB 44 in 2021, required the Department of Management Services to publish a list of approved drone manufacturers by January 1, 2022, and required all state, county and local agencies to stop using drones from other makers by January 1, 2023. Private and commercial pilots in Florida are not covered by that rule.

Does Texas ban Chinese drones?

Not by statute, as of September 2026. Texas House Bill 41, which would have barred state and local agencies from buying drones from companies tied to China and other flagged countries, died in the 2025 session. Separately, Governor Abbott’s prohibited technologies list, maintained by the Department of Information Resources, names SZ DJI Technology and Autel Robotics, which restricts their use by state agencies and on state-owned devices and networks.

• 12 Key NDAA and Blue UAS Requirements Every Drone Buyer Should Know
• The Biggest Drone Supply Chain Risks for U.S. OEMs

What is edge AI on a drone?

Edge AI on a drone means running machine learning models on a computer carried by the aircraft instead of sending data to a remote server. The onboard module processes camera, thermal or lidar data in real time to detect objects, track targets, navigate or avoid obstacles. Because nothing has to travel over a radio link first, the drone can react within milliseconds and keep working when the connection drops.

Why do drones need onboard AI instead of just a flight controller?

A flight controller keeps the aircraft stable and follows waypoints, but it does not have the compute to understand images. Onboard AI adds perception: recognizing people, vehicles or defects, estimating position from cameras, and choosing what to do next. Most designs pair a small real-time flight controller with a separate companion computer that runs the AI models and sends high-level commands to the autopilot.

What is a companion computer on a drone?

A companion computer is a second onboard processor that handles heavy tasks such as computer vision, mapping and mission logic, while the flight controller handles stabilization. The two usually talk over MAVLink or a ROS 2 bridge. Common companion platforms include NVIDIA Jetson modules, Qualcomm-based system-on-modules and boards with Hailo accelerators. An ODM such as Aerora designs electronics around Qualcomm, NVIDIA, Rockchip, Intel, TI, ST and Ambarella platforms.

What does TOPS mean for AI chips?

TOPS stands for tera operations per second, meaning trillions of math operations per second, and it is the headline number for AI accelerator speed. It is usually quoted at INT8 precision, and sometimes with sparsity or INT4, which inflates the figure. Two chips with the same TOPS can perform very differently on a real model because of memory bandwidth, supported layers and software. Treat TOPS as a rough filter, then benchmark your actual network.

How many TOPS does a drone need for AI?

There is no single number, because it depends on the models, camera resolution and frame rate you need. A single lightweight object detector can run on accelerators in the low tens of TOPS, such as a Hailo-8 at 26 TOPS. Running detection, tracking, visual navigation and mapping together pushes teams toward larger modules like the Jetson Orin NX. Profile your full pipeline at target frame rate before choosing hardware. See our guide.

What NVIDIA Jetson modules are used in drones?

The Jetson Orin family is the most common NVIDIA choice for drones and robots today. NVIDIA rates the Orin Nano Super at up to 67 TOPS (sparse INT8) with 7, 15 and 25 W modes, the Orin NX 16GB at up to 157 TOPS in a 10 to 40 W range, and AGX Orin at up to 275 TOPS at 15 to 60 W. Smaller aircraft tend to use Nano or NX; larger ground robots can carry AGX Orin.

Is Jetson Thor suitable for drones?

Jetson AGX Thor is aimed mainly at humanoids and larger robots rather than small drones. NVIDIA rates it at up to 2070 FP4 TFLOPS with 128 GB of memory and a configurable power range of 40 to 130 W. That power draw, plus the cooling it needs, is hard to fit on most small multirotors. It makes more sense on large UAVs, ground vehicles and robots with bigger batteries.

What Qualcomm chips are used in drones?

Qualcomm offers several robotics processors that appear in drones. The QRB5165, used in the Qualcomm Flight RB5 5G reference design, is rated at 15 TOPS and supports multiple cameras, Linux, ROS 2 and PX4. The newer Dragonwing QCS8550 is rated at 48 INT8 TOPS and 12 FP16 TOPS and is sold on system-on-modules such as the Lantronix Open-Q 8550CS. Qualcomm chips are popular where low power and many camera inputs matter.

What is Hailo and how is it used in drones?

Hailo is an Israeli chipmaker whose AI accelerators add neural network processing to a host processor at low power. Hailo rates the Hailo-8 at 26 TOPS with about 2.5 W typical power, the Hailo-8L at 13 TOPS, and the Hailo-10H at 40 TOPS INT4 (20 TOPS INT8). Its Hailo-15 vision processors, rated from 7 to 20 TOPS, are built for smart cameras. Drone builders often pair a Hailo module with a separate host CPU.

What is Ambarella used for in drones?

Ambarella makes vision system-on-chips that combine image processing, video encoding and AI in one chip, which suits camera payloads. Its CVflow architecture runs neural networks, and the 5 nm CV72S supports transformer networks and camera-radar fusion, according to Ambarella. Drone makers value these chips for image quality and low power. Aerora lists Ambarella among the platforms it designs electronics on for camera and vision products.

Jetson vs Qualcomm vs Hailo: which is best for a drone?

Choose NVIDIA Jetson for the broadest software ecosystem and GPU flexibility, Qualcomm for strong performance per watt and many camera inputs, and Hailo for adding efficient AI acceleration to an existing host. Jetson is often easiest for prototyping because CUDA, TensorRT and ROS 2 support are mature. Qualcomm and Hailo can save weight and power in production. Many programs prototype on one platform and then optimize for another.

What's the difference between onboard and cloud processing for drones?

Onboard processing runs AI on the drone itself, so decisions happen in milliseconds and keep working without a link. Cloud processing sends data to remote servers, which allows larger models and fleet-wide analytics but adds latency and depends on bandwidth. Flight-critical tasks such as obstacle avoidance and tracking belong onboard. Post-flight work such as photogrammetry, model training and long-term trend analysis usually runs in the cloud or on a ground station.

Can a drone run AI without an internet connection?

Yes, a drone with an onboard AI module can run detection, tracking and navigation with no internet connection at all. The trained model is stored on the aircraft and runs locally on the companion computer or AI camera. This is one of the main reasons teams choose edge AI: it works in remote areas, under jamming and when cellular coverage is poor. You only need connectivity to update models or upload results.

When should drone AI run on the ground station instead of the aircraft?

Run AI on the ground station when models are too large for the onboard module and a delay of a few hundred milliseconds is acceptable. Examples include rich scene analysis for an operator, multi-drone coordination and live mapping. The tradeoff is dependence on the video link, which adds compression artifacts and can drop out. Safety-related functions such as collision avoidance should stay onboard.

How does object detection work on a drone?

Object detection on a drone uses a neural network, often a YOLO-style model, to draw boxes around people, vehicles or other classes in each camera frame. The model runs on the onboard AI chip, usually after being compressed and compiled for that hardware. Aerial views are harder than ground-level images because objects are small and seen from above, so teams usually fine-tune models on aerial datasets captured at their real flight altitudes.

What is the difference between object detection and object tracking?

Detection finds objects in a single frame, while tracking follows the same object across many frames and gives it a persistent ID. Trackers combine detections with motion prediction, often a Kalman filter, so the drone can keep a lock when the target is briefly hidden. On a gimbal payload, the tracker also drives the gimbal and zoom to keep the target centered. Aerora’s gimbal camera payloads include AI object tracking.

How do drones detect small objects from high altitude?

Drones detect small objects from altitude by combining higher-resolution sensors, optical zoom and models trained on small aerial targets. Common techniques include tiling a large image into smaller crops, running detection at higher input resolution and using thermal cameras where heat contrast helps. Each technique costs compute, so teams balance altitude, lens choice and frame rate against the AI module’s budget. Zoom gimbals let the aircraft confirm a detection at closer effective range.

What is visual navigation for drones?

Visual navigation means a drone estimates its position and motion from cameras instead of relying only on GPS. It tracks features in the image between frames, combines them with inertial data and builds an estimate of where it is and how it is moving. It lets drones fly indoors, under bridges and in GPS-jammed areas. Lantronix announced a collaboration with Aerora in June 2025 on edge-AI visual navigation using Qualcomm-based Open-Q modules.

What is VIO (visual-inertial odometry)?

Visual-inertial odometry (VIO) estimates a drone’s position and orientation by fusing camera images with IMU data from accelerometers and gyroscopes. The camera tracks visual features, while the IMU fills in fast motion between frames and gives scale for a single camera. VIO is accurate over short distances but slowly drifts over time. It struggles in darkness, fog, featureless surfaces such as water or snow, and very fast motion.

What is SLAM and how is it different from VIO?

SLAM (simultaneous localization and mapping) builds a map of the surroundings while locating the robot within it. VIO only tracks motion from moment to moment, so its error keeps growing. SLAM adds a persistent map and loop closure: when the robot recognizes a place it has seen before, it corrects accumulated drift. SLAM can use cameras, lidar or both. It needs more compute and memory than VIO alone.

Visual SLAM vs lidar SLAM: which is better for drones?

Visual SLAM is lighter and cheaper because it uses cameras, but it depends on lighting and texture. Lidar SLAM measures distance directly, works in the dark and produces accurate 3D maps, but the sensors add weight, power draw and cost. Small drones often use visual or visual-inertial methods. Larger drones and ground robots mapping tunnels, mines or forests often use lidar, sometimes fused with cameras. See our guide.

How do drones navigate without GPS?

Drones navigate without GPS by combining other position sources: visual-inertial odometry, SLAM, optical flow sensors, lidar, radar altimeters and matching camera views against stored maps or satellite imagery. A sensor fusion filter blends these into one position estimate. No single method works everywhere, so GPS-denied designs usually layer several and fall back gracefully. Jamming and spoofing have made this a priority for defense and critical-infrastructure programs.

What is GPS spoofing and how can drones detect it?

GPS spoofing is when an attacker broadcasts fake satellite signals so a receiver computes a false position. Drones can detect it by checking GPS against independent sources such as visual odometry, inertial data and barometric altitude, and flagging large disagreements. Multi-constellation, multi-band receivers and antennas that reject interference also help. When spoofing is suspected, the autopilot can switch to vision-based navigation or trigger a safe failsafe.

What is terrain-relative or map-based visual navigation?

Map-based visual navigation estimates a drone’s absolute position by matching what its downward camera sees against a stored map or satellite image. Unlike VIO, which drifts, map matching can correct position over long distances without GPS. It works best over textured ground with a recent reference map and struggles over water, snow, or areas that have changed. Thermal cameras can extend it to night flights when the reference data suits them.

What is sensor fusion on a drone?

Sensor fusion combines data from several sensors, such as IMU, GPS, barometer, magnetometer, cameras, lidar and radar, into one best estimate of the drone’s state. Each sensor has different strengths and failure modes, so fusing them gives better accuracy and resilience than any single one. At the flight controller level this is usually an extended Kalman filter. At the perception level it may merge camera and lidar detections of the same object.

What is an EKF in drone autopilots?

An EKF (extended Kalman filter) is the estimator that most autopilots use to work out position, velocity and attitude from noisy sensors. PX4 uses an estimator called EKF2, and ArduPilot uses EKF3. The filter predicts motion from the IMU and corrects it with GPS, barometer, magnetometer, optical flow or external vision data. Feeding VIO output into the EKF is a common way to add GPS-denied navigation.

How does camera and radar or thermal fusion help drones?

Fusing an RGB camera with thermal or radar gives a drone perception that works in more conditions. Thermal detects people and hot equipment at night or through light smoke, while visible cameras give detail and color. Radar measures range and speed in fog or rain where cameras struggle. Aerora’s gimbal payload options include RGB plus thermal, and one Aerora case study integrated a Teledyne FLIR Hadron dual EO/IR module.

PX4 vs ArduPilot: what's the difference?

PX4 and ArduPilot are the two main open-source autopilots, and both support multirotors, fixed-wing, VTOL, rovers and boats. The biggest practical difference is licensing: PX4 uses the permissive BSD 3-clause license, while ArduPilot uses GPLv3, which requires sharing modifications to the autopilot code you distribute. ArduPilot has a very large feature set and long history. PX4 has a modular architecture and close ROS 2 integration. See our guide.

Which autopilot is better for a commercial drone product, PX4 or ArduPilot?

Many commercial teams choose PX4 because its BSD license lets them keep changes to the flight code proprietary. ArduPilot is also used commercially under GPLv3, and mission logic and AI on a separate companion computer are usually unaffected by the autopilot’s license. The better choice depends on your vehicle type, the features you need, your team’s experience and your IP strategy. Ask a lawyer to review license obligations early.

What is MAVLink?

MAVLink is a lightweight messaging protocol that lets a drone’s autopilot, companion computer, ground control station and payloads exchange telemetry and commands. Both PX4 and ArduPilot use it. Messages cover things like position, battery state, mission items and camera control. MAVLink 2 is the current version and adds features such as extension fields and message signing. Most ground control software, including QGroundControl and Mission Planner, speaks MAVLink.

Is MAVLink secure?

MAVLink is not encrypted by default, so anyone with access to the radio link can read the traffic. MAVLink 2 adds optional message signing: each message carries a 48-bit signature derived from SHA-256 and a 32-byte shared secret key, which helps block spoofed or replayed commands. Signing proves authenticity but does not hide content. For confidentiality, teams add encryption at the radio link, such as AES-256 on data links like Aerora’s D64TR.

What is ROS 2 and why do drone developers use it?

ROS 2 (Robot Operating System 2) is an open-source framework of libraries and tools for building robot software from modular nodes that exchange messages. Drone developers use it on companion computers for perception, mapping, planning and simulation. It runs on DDS middleware, supporting real-time and multi-robot communication. The current long-term support release is Lyrical Luth, released May 2026 and supported to May 2031. See our guide.

How do PX4 and ROS 2 communicate?

PX4 talks to ROS 2 through the uXRCE-DDS bridge. A uXRCE-DDS client runs on the flight controller, and a Micro XRCE-DDS agent runs on the companion computer, exposing PX4’s internal uORB topics as ROS 2 topics. Communication works both ways, so ROS 2 nodes can read vehicle state and send setpoints or commands. ArduPilot also supports DDS for ROS 2, and MAVROS remains an option for MAVLink-based setups.

What are the levels of drone autonomy?

Drone autonomy is commonly described on a scale from Level 0 (fully manual) to Level 5 (fully autonomous), a framework popularized by the analyst firm Drone Industry Insights. Lower levels add pilot assistance such as stabilization and obstacle alerts. Middle levels let the drone handle tasks or whole missions with a human supervising. Level 5 would plan and adapt with no human input. Most commercial drones sit around Levels 2 and 3.

What's the difference between automated and autonomous drones?

An automated drone follows a preplanned script, such as flying a set grid of waypoints, and does not change its plan when conditions change. An autonomous drone perceives its surroundings and makes its own decisions, such as rerouting around a new obstacle or following a moving target. Autonomy needs onboard sensing and compute. In practice, most products mix both approaches, with autonomy handling exceptions inside an automated mission.

What is detect and avoid (DAA) for drones?

Detect and avoid is a drone’s ability to sense other aircraft or obstacles and maneuver to stay safely clear, which it needs to fly beyond visual line of sight without a human watching. DAA systems use sources such as ADS-B receivers, radar, cameras and acoustic sensors, sometimes combined with ground-based surveillance. ASTM F3442 sets performance requirements for DAA systems on smaller uncrewed aircraft, and ACAS sXu provides collision avoidance logic.

What is ACAS sXu?

ACAS sXu is a collision avoidance system designed for small uncrewed aircraft, published by RTCA as DO-396. It adapts the ACAS X family used on crewed aircraft to drones under 55 lb, taking in surveillance data and issuing avoidance guidance. ASTM F3442, the standard for DAA system performance, references ACAS sXu as a possible reference implementation. Developers typically integrate it with their own sensors and autopilot.

Can cameras alone detect other aircraft for BVLOS flights?

Cameras can detect other aircraft, but relying on cameras alone is hard for BVLOS because small aircraft appear as a few pixels at useful ranges, and haze, glare and cluttered backgrounds cause misses and false alarms. Many DAA designs therefore combine cameras with ADS-B, radar or ground-based surveillance. Cameras add value for aircraft that do not broadcast their position. Any system must be tested against the performance its operation requires.

What is drone swarming?

Drone swarming is when multiple drones coordinate as a group to share a task, such as searching an area, mapping or putting on a light show. Each drone runs its own autonomy and shares position and intent with the others over a mesh network or central controller. Swarms need reliable communication, collision avoidance between members and a way to reassign tasks when a drone fails.

What is the difference between centralized and decentralized swarm control?

In centralized control, a ground station or leader drone plans every aircraft’s path and sends commands. It is simpler but fails if that node or its link goes down. In decentralized control, each drone decides for itself from local sensing and messages from neighbors, so the swarm keeps working if some members drop out. Most real systems combine both, with central tasking and local collision avoidance.

What is an AI camera payload for a drone?

An AI camera payload is a gimbal camera with its own onboard processor that runs detection, tracking or classification before sending video down. It offloads work from the companion computer and can drive the gimbal and zoom automatically to follow a target. Aerora’s gimbal payloads run on a 7 nm Qualcomm processor with up to 200 MP image processing and AI object tracking. See our guide.

What should I look for in an AI gimbal camera?

Look at sensor types (RGB, thermal, zoom, laser rangefinder), stabilization accuracy, the onboard AI processor and which models it can run, and how it integrates with your autopilot, usually through MAVLink or a vendor SDK. Check weight, power draw, and whether you can deploy your own trained models or only the vendor’s. For defense or government buyers, component sourcing and FCC authorization status also matter.

How much power does AI compute use on a drone?

Onboard AI modules typically draw from a few watts to several tens of watts. Hailo rates the Hailo-8 at about 2.5 W typical, while NVIDIA’s Jetson Orin NX runs in a 10 to 40 W range and AGX Orin in a 15 to 60 W range. On a multirotor, that draw competes directly with the motors, so heavier compute cuts flight time twice: through power use and added mass, including cooling.

How do you cool an AI module on a drone?

Drone AI modules are usually cooled with a heat sink and the propeller downwash or forward airflow, sometimes with a small fan. Designers place the module where air moves, use aluminum or magnesium enclosures as heat spreaders and connect hot chips with thermal pads. The hardest cases are hovering in hot sun and ground testing without airflow. Thermal testing across the full temperature range belongs in EVT and DVT.

What happens when an AI chip overheats in flight?

When an AI chip gets too hot, it throttles, lowering its clock speed to protect itself, and in extreme cases shuts down. For a drone, throttling means dropped frames, slower detections and lagging tracking, which can quietly degrade autonomy mid-mission. Good designs monitor temperature, pick a power mode with margin and tell the autopilot when perception is degraded so it can slow down, hold position or return.

What is model quantization for edge AI?

Quantization shrinks a neural network by storing its weights and activations at lower precision, such as INT8 or INT4 instead of 32-bit floating point. This cuts memory use, speeds up inference and reduces power, which matters on drones. The tradeoff is a small accuracy loss that must be measured. Post-training quantization converts a finished model using calibration data, while quantization-aware training simulates low precision during training and usually keeps more accuracy.

What is TensorRT?

TensorRT is NVIDIA’s software for optimizing trained neural networks to run fast on NVIDIA GPUs, including Jetson modules. It takes a model, often exported in ONNX format, fuses layers, selects the fastest kernels and can run at FP16 or INT8 precision with calibration. The result is an engine file tuned to one specific GPU and TensorRT version. Qualcomm, Hailo and Ambarella provide their own compilers that do similar jobs.

How do I deploy a YOLO model on a Jetson or other edge chip?

Train or fine-tune the model on data that matches your camera and altitude, then export it, usually to ONNX. Compile it with the target’s toolchain, such as TensorRT for Jetson or the vendor compiler for Qualcomm or Hailo, and quantize it with a representative calibration set. Then check accuracy against the original model and measure end-to-end latency on the real hardware, including camera capture and preprocessing, not just inference.

Does quantizing a model reduce accuracy?

Quantization usually causes a small accuracy drop, and for many vision models INT8 post-training quantization lands close to floating-point accuracy. Results vary by model and data: small objects, unusual layers and poor calibration sets make losses worse. If the drop is too large, use quantization-aware training or keep sensitive layers at higher precision. Always validate on your own flight data rather than trusting benchmark results.

How is ground robot autonomy different from drone autonomy?

Ground robots share most of the drone autonomy stack, including SLAM, sensor fusion and object detection, but face different problems. They must judge whether terrain can be driven, handle slopes, mud and stairs, and avoid people at close range. They can usually carry heavier compute, batteries and lidar than small drones. ROS 2 and its Nav2 navigation stack are common starting points. An ODM such as Aerora works across aerial, ground and maritime systems.

How do autonomous boats and maritime drones navigate?

Autonomous surface vessels navigate with GNSS, inertial sensors, marine radar, AIS receivers, cameras and sometimes lidar, fused into a picture of nearby traffic and obstacles. Their collision avoidance logic generally has to follow the COLREGs, the international rules for preventing collisions at sea. Water brings its own problems: glare, waves, spray and few visual features for camera navigation. Underwater vehicles lose GNSS entirely and rely on inertial and acoustic navigation.

Can one ODM build edge AI hardware for aerial, ground and maritime robots?

Yes, many of the building blocks carry across: compute modules, camera payloads, data links, motor control and sensor integration. The vehicle-specific work is in mechanical design, environmental sealing, power systems and control tuning. Aerora describes itself as an ODM for aerial, ground, maritime and robotic systems, with U.S. engineering and manufacturing in Vietnam. For program-specific scope, pricing and IP terms, contact the team.

• Best Edge AI Chips for Drones
• Best Edge AI Integration Partners for Autonomous Drones (2026)
• Best Autonomy Software Stacks for Drones
• Best Sensors for Autonomous Navigation
• Best Robotic Autonomy Hardware Partners (2026)

How big is the drone market?

Industry estimates put the commercial drone market in the tens of billions of dollars and growing steadily. Drone Industry Insights expects the commercial market to reach $54.6 billion by 2030, growing at a 7.7% CAGR. MarketsandMarkets, which measures the broader UAV market, projects growth from $26.12 billion in 2025 to $40.56 billion in 2030. The figures differ because each firm defines the market differently.

Why do drone market size estimates vary so much?

Drone market estimates vary because research firms count different things. Some include only commercial hardware, others add software, services and data processing, and some fold in military and consumer drones. Regional scope, base year and currency assumptions also shift the totals. When you quote a market figure, name the firm and the segment it measures, and compare growth rates within one report rather than mixing numbers across reports.

How much money is being invested in commercial drone companies?

Investment in commercial drones hit a record in 2025. Drone Industry Insights reports that commercial drone companies raised $3.86 billion in 2025, the highest figure in its investment database, which goes back to 2015. Large rounds went to delivery, defense-adjacent and autonomy companies. For hardware startups, that capital often funds the move from prototype to production, which is where an engineering and manufacturing partner usually enters the picture.

Which industries use commercial drones the most?

Agriculture, infrastructure inspection, construction and surveying, energy, public safety and delivery are among the largest commercial drone markets. Mapping and inspection work drives much of the fleet today because the return is easy to measure: fewer people at height, faster surveys and better records. Agriculture spraying and drone delivery are the fastest-moving newer segments in the U.S., while public safety grew sharply after the FAA simplified waivers for drone-as-first-responder programs.

How are drones used in agriculture?

Farmers use drones for three main jobs: spraying crops, scouting fields and capturing multispectral imagery. Spray drones apply fungicides, herbicides, fertilizer and seed, especially on wet, hilly or small fields where ground rigs struggle. Scouting drones give a quick overhead view of stand counts, weeds, flooding and storm damage. Multispectral cameras reveal crop stress before it is visible to the eye. See our agriculture use cases guide.

How many acres do spray drones treat in the U.S.?

Spray drones treated more than 16.4 million acres in the U.S. in 2025, according to the American Spray Drone Coalition’s industry survey. That was a 58.7% increase over the prior year. The survey links the growth mainly to more operators: the FAA count of approved Part 137 drone operators rose 58.3% to 1,710. Average acres treated per operator stayed roughly flat.

How much does drone spraying cost per acre?

Drone spraying averaged about $13 per acre in 2025, down from $21 the year before, according to the American Spray Drone Coalition’s 2025 survey. The survey attributes the 38% drop mostly to more competition. Actual prices depend on the product being applied, field size and shape, travel distance, water volume per acre and whether the operator supplies the chemical. Treat any single figure as a regional average, not a quote.

Do I need permission to spray crops with a drone?

Yes, in the U.S. spraying crops with a drone generally requires an FAA Part 137 agricultural aircraft operator certificate, and drones over 55 pounds also need a Section 44807 exemption. Pesticide applicator licensing is handled separately by state agencies. The rules change, so check the FAA’s current guidance on dispensing chemicals with drones before you buy a spray drone or start offering the service.

What is crop scouting with drones?

Crop scouting with drones means flying a field to spot problems faster than walking it. A farmer or agronomist flies an RGB or multispectral camera over the field, then checks the areas the imagery flags: thin stands, weed patches, nutrient stress, pest damage, drainage issues or hail damage. Scouting drones are smaller and cheaper than spray drones, and many growers use them to decide exactly where a spray drone or ground rig should go.

What does a multispectral drone camera show on crops?

A multispectral camera measures reflected light in several narrow bands, including red edge and near-infrared, that show plant health. Software turns those bands into indices such as NDVI, which highlight stressed, diseased or nutrient-deficient areas before they are visible in a normal photo. Growers use the maps for variable-rate application and yield planning. An ODM such as Aerora offers multispectral gimbal payload options for OEMs building agriculture drones.

How are drones used for power line inspection?

Utilities fly drones along transmission and distribution lines to inspect insulators, conductors, hardware and vegetation encroachment without climbing towers or rolling bucket trucks. Zoom cameras capture close detail from a safe standoff distance, thermal cameras find hot connections, and lidar measures vegetation clearance. After storms, drones speed up damage assessment. Long corridors push utilities toward beyond-visual-line-of-sight operations and docked drones. See our guide to drones for infrastructure inspection.

Can drones inspect bridges?

Yes, drones are widely used to inspect bridges, especially decks, bearings, girders and hard-to-reach undersides. They capture high-resolution imagery for crack and corrosion detection and can build 3D models for tracking change over time. Collision-tolerant or caged drones can work under decks where GPS is weak. Drones usually support, rather than fully replace, hands-on inspections, which state transportation departments still require for some elements.

How are drones used to inspect wind turbines?

Drones inspect wind turbine blades for leading-edge erosion, cracks, lightning damage and delamination, usually while the turbine is stopped for a short window. Automated flight paths capture every blade surface in a repeatable pattern, and software stitches and tags the defects. That replaces much of the rope-access and crane work, cuts downtime and gives operators a consistent record to compare year over year, both onshore and offshore.

How do drones inspect solar farms?

Drones inspect solar farms with thermal cameras that spot hot cells, failed bypass diodes, string outages and soiling across thousands of panels in one flight. An RGB camera flown alongside confirms physical damage. Software geotags each anomaly so a technician can go straight to the right panel. For large utility-scale sites, a drone survey can cover in hours what a manual walk-down covers in days.

What payloads do inspection drones carry?

Inspection drones typically carry a zoom RGB camera, a thermal camera, or both on one gimbal, plus lidar for 3D measurement and gas sensors for leak detection. The right mix depends on the asset: thermal for electrical and solar work, zoom for bolts and insulators, lidar for vegetation and structure modeling. Aerora offers gimbal payloads with RGB plus thermal, lidar and laser range meter options. See our inspection payloads guide.

Is drone inspection safer than rope access or bucket trucks?

Drone inspection generally reduces the time people spend working at height or near energized equipment, which is the main safety argument for it. Crews stay on the ground while the drone collects the imagery. Drones bring their own risks, including flyaways, battery failures and airspace conflicts, so programs still need trained pilots, maintenance procedures and site risk assessments. Most operators use drones to target hands-on work, not to eliminate it.

What is OGI (optical gas imaging)?

Optical gas imaging (OGI) uses a specially filtered infrared camera to make hydrocarbon gases like methane visible as a plume on screen. Leaks that are invisible to the eye show up as moving smoke-like clouds, so an operator can find the exact source. OGI is a standard leak detection and repair (LDAR) tool in oil and gas. Handheld OGI cameras are common, and more of them are now mounted on drones.

Can drones detect methane leaks?

Yes, drones can detect methane leaks using OGI cameras or laser-based methane sensors. OGI drones show the gas plume so crews can pinpoint the source. Laser sensors measure methane concentration along the beam and suit pipeline and landfill surveys. Drones reach flare stacks, tank tops and remote well pads without scaffolding. Aerora’s EO/IR integration approach, which it says extends to OGI sensors, is described on our technology page.

Has the EPA approved drones for methane leak inspections?

Yes, the EPA has approved specific drone-based OGI systems as alternative test methods under its oil and gas rules (Subparts OOOOa and OOOOb). ChampionX’s aerial OGI platform received approval in early 2025, and Percepto announced approval for its autonomous OGI drone system on 1 October 2025. Approval covers a specific system and method, so operators must follow the approved protocol, not just fly any thermal drone.

What is the EPA's OOOOb methane rule, and where does it stand?

OOOOb and OOOOc are the EPA’s 2024 methane and VOC standards for new and existing oil and gas sources, including leak detection requirements. In 2025 the EPA extended many compliance deadlines, first through a July 2025 interim final rule and then a November 2025 final rule. Further reconsideration has followed in 2026. Operators should check the EPA’s current actions page, because deadlines and requirements have moved several times.

What's the difference between OGI cameras and laser methane sensors on drones?

An OGI camera produces an image of the gas plume, which is best for finding exactly where a leak is on equipment. A laser methane sensor, often a tunable diode laser, measures methane concentration along a line and is better for fast surveys of pipelines, landfills and large sites. Many operators use laser sensors to screen and OGI to confirm and locate. Both can be drone mounted. See our guide to drone sensor integration companies.

What is drone as first responder (DFR)?

Drone as first responder (DFR) is a public safety model where a drone launches from a dock or rooftop to a 911 call, often reaching the scene before officers. The live video helps dispatchers decide what resources to send. The Chula Vista Police Department in California launched the first DFR program in 2018 under the FAA’s Integration Pilot Program. Chula Vista has reported drone arrival times under 2.5 minutes.

How many police and fire agencies have drone-as-first-responder waivers?

DFR adoption jumped after the FAA introduced a simpler waiver process in May 2025. Vertical Magazine reports that about 50 DFR waivers had been approved by 2024, and more than 1,000 were approved in the six months after the new process launched. Many of those agencies are still building out docks, staffing and policies, so waivers issued do not equal fully operating programs.

How do firefighters and search-and-rescue teams use drones?

Firefighters use drones with thermal cameras to find hot spots through smoke, map wildfire perimeters and check roofs before crews go in. Search-and-rescue teams use thermal to find people at night or under light cover, and loudspeakers or drop mechanisms to reach them. Hazmat teams use drones with gas sensors to assess a scene from a distance. Thermal and zoom are the core payloads for all of these missions.

Is drone delivery actually happening at scale?

Yes, drone delivery is now running at meaningful scale in parts of the U.S. Walmart announced on 29 May 2026 that it had passed one million drone deliveries, working with Wing and Zipline from 66 stores in four states, with an average delivery time of 23 minutes. In January 2026, Wing and Walmart announced an expansion to another 150 stores. Coverage is still limited to specific metro areas.

How many deliveries has Zipline made?

Zipline passed 2 million commercial deliveries in January 2026, when it also announced a $600 million funding round and expansion to Houston and Phoenix. Zipline started with medical supply delivery in Rwanda and Ghana and now runs home delivery in the U.S. with retailers and health systems. Its count covers its full global operation, not only U.S. home delivery.

What limits drone delivery today?

Drone delivery is limited mainly by beyond-visual-line-of-sight approvals, payload weight, weather, noise acceptance and the cost of running each hub. Most services carry a few pounds over short distances, so they suit small, urgent orders like groceries, meals and pharmacy items. Unit economics depend on flying many deliveries per drone per day. Landing zones, community approval and integration with store operations also slow expansion.

What is drone mapping and surveying?

Drone mapping means flying a planned grid, capturing overlapping images or lidar data, and processing them into orthomosaic maps, 3D models, elevation surfaces and volume measurements. Surveyors, construction firms, mines and engineers use these outputs for site planning, progress tracking and earthwork calculations. A drone can map a site in a fraction of the time a ground crew needs. See our drone mapping software guide.

What's the difference between photogrammetry and lidar for drone mapping?

Photogrammetry builds maps and 3D models from overlapping photos, which is affordable and produces realistic color outputs. Lidar measures distances with laser pulses, so it can capture ground under vegetation and thin structures like power lines more reliably. Photogrammetry suits open sites, stockpiles and construction. Lidar suits forested terrain, corridors and utility work. Many survey teams own both and choose per job.

How accurate is drone mapping?

Drone mapping accuracy depends on the positioning method, sensor quality, flight altitude and ground control. Drones with RTK or PPK GNSS correction, checked against surveyed ground control points, can produce survey-grade results suitable for many engineering tasks. Without correction, maps look good but can be offset by meters. Always verify with independent checkpoints, and let a licensed surveyor sign off where the law requires one.

How are drones used in mining?

Mines use drones to measure stockpile volumes, map pits and haul roads, monitor highwall and tailings dam stability, and plan blasts. Regular surveys give engineers up-to-date topography without putting surveyors near active equipment or unstable slopes. Some sites fly docked drones on a schedule for daily volume reports. Thermal and gas sensors support additional safety checks around equipment and ventilation.

Can drones fly underground in mines?

Yes, drones can fly underground, but they need different technology because there is no GPS. Underground drones typically use lidar-based SLAM for navigation and mapping, carry their own lighting, and often have protective cages for collisions in tight spaces. Mines use them to inspect stopes, ore passes and areas unsafe for people. See our guide to sensors for autonomous navigation.

How does the oil and gas industry use drones?

Oil and gas companies use drones to inspect flare stacks, tanks, pipelines, well pads and offshore platforms, and to find methane leaks with OGI. Flare stack inspections can often be done while the flare is running, which avoids a shutdown. Pipeline operators fly corridors to watch for encroachment, erosion and leaks. Docked drones at large facilities can run routine patrols without sending crews out each time.

Can drones inspect inside tanks and confined spaces?

Yes, caged, collision-tolerant drones are used to inspect the inside of tanks, boilers, vessels and chimneys. They carry their own lights and sometimes lidar, so they work without GPS or daylight. Inspecting from outside the space reduces confined-space entries, scaffolding and downtime. Ultrasonic thickness gauges on contact drones can add wall-thickness measurements, although many results still need follow-up by a certified inspector.

Can drones be used for perimeter security?

Yes, drones are used to patrol perimeters at data centers, utilities, ports, solar farms, prisons and industrial sites. A docked drone can launch automatically on an alarm from a fence sensor or camera and send live video to a security operator within minutes. Thermal cameras make night patrols practical. See our edge AI integration guide. Aerora integrates AI cameras with object tracking, which suit this kind of monitoring, into OEM drone platforms.

What is a drone-in-a-box or drone dock?

A drone-in-a-box, or dock, is a weatherproof station where a drone lands, charges and launches automatically. Remote operators schedule missions or trigger them from alarms, so no pilot needs to be on site. Docks are used for security patrols, DFR, mining surveys and industrial inspection. Buyers should ask about FCC status, since the FCC has proposed further restrictions on foreign-made docks.

How are drones used in maritime operations?

Maritime users fly drones to inspect ship hulls, cargo holds, cranes and offshore structures, to support search and rescue, and to monitor ports and coastlines. Tethered drones on vessels provide a long-endurance overhead view. Drones can also carry small parcels from shore to ships, which is cheaper than a launch boat. Salt, wind and moving decks mean maritime drones need corrosion protection and good landing control.

What are underwater drones used for?

Underwater drones, meaning remotely operated vehicles (ROVs) and autonomous underwater vehicles (AUVs), inspect ship hulls, dams, pipelines, subsea cables, offshore wind foundations and aquaculture pens. They also map the seafloor and support search and recovery. ROVs are tethered and piloted live, while AUVs follow preprogrammed routes. Radio does not work well underwater, so they rely on tethers, acoustic links or onboard autonomy.

What is an uncrewed surface vessel (USV)?

An uncrewed surface vessel (USV) is a boat that operates without a crew, either remotely piloted or autonomous. USVs survey harbors and rivers with sonar, monitor water quality, patrol ports and support offshore energy work. They can stay out for long periods at lower cost than crewed boats. Aerora’s ODM model covers maritime systems alongside aerial and ground platforms, as listed on its services page.

How are robots used in warehouses?

Warehouses use mobile robots to move shelves and totes, sort parcels, pick items and haul pallets, while people handle the tasks robots still find hard. Amazon said in July 2025 that it had deployed its one-millionth robot, after starting with its 2012 acquisition of Kiva Systems. Autonomous mobile robots, robotic arms and automated forklifts are the main categories.

What do agricultural ground robots do?

Agricultural ground robots weed, scout, thin, spray and harvest, usually in specialty crops like vegetables, fruit and vineyards where labor is scarce. Some use cameras and machine learning to tell crops from weeds and remove weeds mechanically or with lasers. Others carry sensors between rows for close-up plant data. Autonomous tractors and carts that follow workers are also moving into row crops and orchards.

When should I use a ground robot instead of a drone?

Use a ground robot when the job needs long runtime, heavy payloads, physical contact or close work at ground level, such as weeding, hauling or indoor patrols. Use a drone when you need speed, height or coverage across a large or hard-to-reach area. Many operations combine them: a drone finds the problem and a ground robot fixes it. See our guide to robotic autonomy hardware partners.

What does "dual-use" mean for drones?

Dual-use means a technology has both civilian and military applications. Many drone components and systems, such as thermal cameras, data links, autonomy software and long-range airframes, serve commercial inspection and public safety as well as defense. Dual-use products can fall under export controls, and defense customers often add sourcing requirements. Companies selling into both markets usually plan for those requirements early in design.

How is selling drones to defense different from selling to commercial customers?

Defense buyers usually require stricter supply chain rules, security features, documentation and testing than commercial buyers, and their procurement cycles are longer. U.S. programs typically require NDAA sourcing, and some require products on the DCMA-managed Blue UAS list. Commercial buyers focus more on price, ease of use and return on investment. See our overview of NDAA and Blue UAS requirements.

Which industries use thermal cameras on drones?

Thermal drone cameras are used across utilities, solar, oil and gas, public safety, firefighting, search and rescue, security, building inspection, agriculture and wildlife surveys. Thermal finds heat differences, so it reveals hot electrical connections, failing solar cells, people at night, fire hot spots, roof moisture and livestock. It is one of the most widely used drone payloads after standard RGB cameras. Aerora’s gimbal payloads include RGB plus thermal options.

Do inspection drones need both thermal and zoom cameras?

Many inspection drones carry both, because thermal finds the problem and zoom shows what it is. A dual EO/IR payload combines a thermal sensor with a high-resolution optical camera on one gimbal. In one case study, Aerora integrated a Teledyne FLIR Hadron (640×512 thermal plus 64 MP optical) with Qualcomm processing on a 3-axis gimbal for an OEM, as described in the Lantronix collaboration post.

What is the ROI of a commercial drone program?

Drone program ROI usually comes from four places: fewer labor hours per inspection or survey, less equipment rental such as bucket trucks and scaffolding, reduced downtime, and fewer safety incidents. Better data can also catch defects earlier, which avoids larger repairs. Returns vary widely by industry and how often the drone flies. Programs that fly frequently on repeatable missions tend to see the clearest payback.

How long does it take for a drone program to pay for itself?

Payback time depends on how many flights replace costly manual work, so there is no single answer. A mapping team replacing ground surveys on busy sites may recover costs quickly, while an occasional-use program may take much longer. Build the case from your own numbers: current cost per task, drone cost per task, flights per year and data processing time. Pilot projects give the most reliable estimates.

What costs do people forget when calculating drone ROI?

Common hidden drone costs include pilot training and certification, batteries replaced over time, insurance, software subscriptions, data storage and processing time, maintenance, waivers and program management. Docked systems add installation, connectivity and site permits. Fleet replacement also matters, especially if regulatory or supply changes make a model hard to service. Count these before comparing a drone program against the manual process it replaces. Our guide to drone supply chain risks covers the sourcing side.

Does drone ROI differ for OEMs building drones versus companies using them?

Yes. For operators, ROI comes from cheaper, safer, faster work. For OEMs building drones or robots, ROI comes from unit margins, time to market and how many units a design sells across markets. OEMs gain most from reusing a platform across several applications, such as one airframe with swappable inspection, mapping and security payloads. Engineering and tooling costs are spread over more units that way, which is how an ODM such as Aerora approaches platform programs.

What should I build for each drone market?

Match the product to the market’s core job. Agriculture needs heavy-lift spray drones and multispectral scouting. Inspection needs stable gimbals with zoom, thermal and lidar. Oil and gas needs OGI. Public safety and security need docked drones with thermal and live video. Delivery needs efficient airframes and reliable BVLOS systems. An ODM such as Aerora can adapt a platform to one of these through its ODM services.

How do I choose which market to target first with a new drone or robot?

Start with the market where your payload, range or autonomy solves a costly problem that buyers already pay to fix, and where regulatory approval is achievable. Talk to operators before designing, validate with pilots, and check sourcing requirements early if public safety or government buyers are on the list. Aerora, with locations in Santa Clara and Portland, discusses program specifics through its contact page.

• 10 Best Drone Use Cases in Agriculture
• 9 Best Drones for Infrastructure Inspection
• 9 Best Drone Mapping Software Platforms
• 9 Best Drone Industry Conferences for 2026 and 2027