Mashable reported that Amazon is extending drone delivery to hundreds more cities. Drone delivery has been announced, paused and re-announced for over a decade, so it is worth understanding what the actual constraints are and why they have taken so long to shift.
The flight is not the hard part
Flying a package a few miles and lowering it into a garden is, technically, straightforward. Consumer drones have done the flying part reliably for years.
The hard parts are regulatory and logistical, and they are what has kept the technology in trials.
Beyond visual line of sight
The single biggest constraint is a rule, not a limitation of hardware. In most countries a drone must remain within the operator’s visual line of sight.
That rule makes delivery pointless, because the operator would have to stand where the drone is going. Every commercial delivery programme therefore depends on an exemption, usually called BVLOS, granted for a defined area after the operator demonstrates the aircraft can detect and avoid other traffic without a human watching.
This is why expansions are announced city by city rather than nationally. Each area is a separate regulatory approval, and that approvals process, not the engineering, sets the pace.
The physics that shapes everything
Battery energy density is the constraint behind every design decision.
A multirotor spends most of its energy simply staying up. Payload and range trade directly against each other, which is why delivery drones carry small parcels, typically a few kilograms, over short ranges of a few miles, and why they return to base rather than making multiple stops.
It is also why most designs lower the package on a tether or drop it from a low hover rather than landing. Landing and taking off again is expensive in energy and risky in an unpredictable garden.
The economics follow from the physics. Drone delivery suits small, urgent, light items, which is why pharmacies and convenience goods appear in every trial and furniture does not.
Detect and avoid
An autonomous aircraft sharing airspace has to notice other aircraft, and the sensing is genuinely difficult. Radar is heavy, cameras struggle in poor visibility, and the objects that matter, such as a light aircraft or another drone, are small and fast.
Most systems combine several approaches: broadcast position data from cooperative aircraft, onboard cameras or radar for the uncooperative ones, and strict geofencing to keep away from airports and restricted areas.
The failure case regulators care about is not a lost parcel. It is a collision with a crewed aircraft, or an uncontrolled descent onto a person, and the certification burden reflects that.
The problems nobody solved yet
- Weather. Wind, rain and icing ground the fleet, and a delivery method that stops on a bad day needs a conventional fallback anyway.
- Noise. Multirotors are loud in a pitch people find particularly irritating, and this has generated real opposition in trial areas.
- Somewhere to put it. Flats, terraced housing without gardens and shared entrances are most of the urban housing stock and are poorly served.
- Theft and interference. A package left in an open garden, and drones being targeted, are both practical problems.
- Cost per delivery. Still high once approvals, ground staff and low utilisation are counted.
What to expect realistically
The plausible near-term shape is not drones replacing vans. It is a narrow, high-value slice: urgent pharmacy items, small medical deliveries, and low-density areas where a van journey is disproportionately expensive.
Medical logistics is where the technology has been most convincing, because the value of speed is high and the payload is genuinely light.
If your area is added to a programme, the useful questions are which items qualify, what the weather cancellation policy is, and whether you have a delivery spot that works. Those determine whether it is useful to you far more than the technology does.
Join the discussion