What drones are actually used for

What works. Drone delivery stopped being a demo some time ago. In a growing set of US metros you can have groceries, pharmacy items, household essentials and takeaway flown to a yard or a driveway, usually inside half an hour. The unit economics work best on exactly what you would guess: small, light, urgent, and awkward to drive. A forgotten prescription beats a case of bottled water every time.

What is overstated. Coverage maps. A metro being "served" often means a handful of suburbs inside a tight radius of one launch site, chosen for favourable airspace and low tree density. National coverage claims are aspirations with a launch schedule attached.

What gates it. Line of sight. Every observer a network needs along a route is a salary, and the whole model only pencils out when one operator can supervise many aircraft over a wide area. That is precisely what Part 108's area approvals are designed to allow. Publication is the starting gun for the buildout everyone has planned.

What works. This is the fastest-growing commercial category, and it splits in two. Heavy-lift spray platforms apply crop protection and fertiliser directly, and on small, irregular or steeply sloped fields they increasingly beat a manned agricultural aircraft on cost — no airstrip, no ferry time, no minimum acreage to justify the call-out. Separately, multispectral and NDVI scouting produces variable-rate prescription maps that cut input spend by finding the parts of a field that actually need the product.

What is overstated. Replacing the ag plane outright. On a large, flat, contiguous section, a manned aircraft still covers ground far faster per hour. Spray drones win on the awkward acres, not on the easy ones. Battery swaps and mixing time also eat into field-day throughput more than spec sheets imply.

What gates it. Weight. Loaded spray platforms sit well above the thresholds most operators are certificated for, which pulls in exemptions and additional requirements. Applicator licensing is a separate state-by-state layer that has nothing to do with the FAA.

What works. The safety argument here is the strongest in the industry and it does not require any economic modelling: nobody has to climb the tower. Cell towers, bridges, roofs, wind turbine blades, flare stacks, transmission lines and dam faces are all now routinely inspected from the air. High-resolution imagery plus thermal picks up delamination, hot joints, moisture ingress and corrosion that a visual climb inspection can easily miss, and the imagery is a dated record you can compare against next year's.

What is overstated. Full automation. Software will flag candidate defects, but a qualified inspector still makes the call, and the deliverable is an engineering judgement rather than a folder of photos. Selling the flight without the analysis is how a lot of small operators discovered the margin is in the report.

What gates it. Linear assets. Inspecting forty miles of transmission line under visual line of sight means driving the corridor and relaunching repeatedly. Corridor operations are one of the clearest wins in the Part 108 proposal.

What works. Drone as first responder programmes launch an aircraft from a rooftop on the same dispatch that sends a patrol car, and it frequently arrives first. Officers get eyes on the scene before they get there, which resolves a meaningful share of calls without an in-person response at all. Search and rescue with thermal in open terrain has an obvious value case, and fire departments use aerial thermal for hotspot mapping and to plan attack lines.

What is overstated. How settled this is. DFR is the most contested application in the entire field, and the objections are serious rather than reflexive: persistent overhead camera presence, retention policies for footage of people who were never suspected of anything, and the question of whether a routine aerial response changes what a neighbourhood feels like to live in. Departments that have handled this well published their policies before they bought the aircraft.

What gates it. Line of sight again, plus community consent. Several programmes have been paused by councils rather than by regulators.

What works. Photogrammetry and drone lidar collapsed the cost of a site survey by an order of magnitude. Construction volumetrics — how much material is in that stockpile, how much has moved since last Tuesday — went from a specialist engagement to a routine weekly flight. Add ground control points and the accuracy is good enough for a large share of the survey work that used to require a crew on foot for days.

What is overstated. Survey-grade accuracy without ground control. Marketing quotes centimetre figures that assume careful GCP placement and good conditions; skip those and the numbers drift. Vegetation is also unforgiving — photogrammetry maps the canopy, and only lidar gets you meaningful ground returns under tree cover.

What gates it. Area size and, in many states, professional surveyor licensing for anything that becomes a legal document.

What works. The original commercial use and still the largest by number of operators. A shot that once needed a helicopter, a crane or both now needs one person and a case. Real estate listing video is the volume business underneath it. And FPV one-shots — a single continuous take flying through a whole building, out a window and away — did to interior cinematography roughly what the Steadicam did to walking shots.

What is overstated. That it is a business rather than a skill. The barrier to entry collapsed, which means the barrier to earning collapsed too. The operators making a living are the ones who can also light, edit, colour and deliver, not the ones who can only fly.

What gates it. Part 107 for anything commercial, airspace authorisation near airports, and operations over people rules for anything shot at an event.