On August 19, Amazon named a year-end number for Prime Air: nearly 500 U.S. cities and towns by the close of 2026, a sixfold increase on the current footprint.
The base it is expanding from is 11 sites across seven states — Arizona, Florida, Kansas, Louisiana, Michigan, Nebraska and Texas. Each site covers roughly 175 square miles. Chicago, Cleveland, Atlanta, Syracuse and Boise are queued next.
David Carbon, the Amazon vice president who runs Prime Air, put a volume on the year so far: hundreds of thousands of packages already delivered by drone.
The load-bearing number is not 500
500 is footprint. What tells you what this does to delivery work is a different sentence in the same post.
Nearly all items 5 lb or less that fit in a large shoebox can go by drone, and Amazon says that band covers more than 60% of the items customers most frequently purchase.
That reframes the question. It is no longer “can a drone deliver this,” it is “what is still in the van.”
Three supporting conditions arrived with it. Speed: as fast as 30 minutes, most orders around 60. Price: free for Prime members on orders of $50 or more, $2.99 for Prime under $50, $4.99 without Prime. Regulatory standing: FAA Part 135 certification, the same framework commercial air carriers operate under and the highest tier of oversight available to a drone delivery operator.
Put those three together and the pilot phase is over. The pricing sits inside the normal shipping table. The certification sits inside the normal aviation regime. The catalog sits inside the normal checkout — same app, same returns, same A-to-z Guarantee.
Where 500 sits against the rest of the network
500 towns is a large number in drone terms and a small one in Amazon terms, and holding both at once is how you read it correctly.
Amazon’s own breakdown puts Prime Air alongside three other speed tiers. Amazon Now runs 30-minute delivery in dozens of cities from small dedicated sites in denser areas. One-hour and three-hour delivery covers more than 90,000 items through the Same-Day network. Same-Day Delivery itself reaches more than 10,000 cities and towns.
So drone coverage at year-end will be roughly 5% of the Same-Day footprint. This is not a replacement network. It is a wedge driven into one specific band of the parcel mix, in one specific geography type, at the speed tier where the margin pressure is highest.
Amazon also opened its first non-U.S. Prime Air site in Darlington, England, serving customers within a 7.5-mile radius of a fulfilment centre. The 7.5-mile figure is the more informative one for anyone modelling this: the economics are built around a tight radius from an existing building, which is why suburban sites work and dense city centres do not.
How a route pays
The economics of delivery work run on density.
A DSP contractor’s driver earns against how many stops a route holds and how far apart they sit. Five houses in a row, one small parcel each, and a single park handles three deliveries. Thin that out and the mileage stays, the hours stay, and the stop count drops.
Now pull 60% of the most commonly ordered items out of the van.
The route survives. The houses survive. Bulk, heavy and multi-item orders still need a driver. What changes is the stop count per mile, which is the thing the route was priced on. Drones do not remove the line. They remove the part of the line that made it worth running.
Amazon also committed $1.9B to the DSP program this year for safety initiatives and AI tools. One company funding the human delivery network while building a separate channel that lifts the small-parcel band out of it is not a contradiction. It is both ends of a transition described at once.
Three companies, one cut
Set this week’s shipments side by side and the shape resolves.
Serve Robotics cut full-year revenue guidance to $9M on August 17 and left the 2,000-robot fleet target standing. Sidewalk robots were never blocked on manufacturing. They are blocked on unit economics, and holding fleet while cutting revenue says the company is buying density rather than this year’s revenue.
JD ran unmanned delivery vans across 20 provinces with headcount flat on August 13. That is the ground-fleet route, which sidesteps airspace regulation entirely.
Amazon took the air route. The cost is Part 135. The return is that traffic stops mattering.
Three different technical bets landing in the same place: the most standardized, lightest, highest-frequency tier of last-mile work leaves human hands first.
Which hours go first
DSP drivers lead, and it lands as hours, not headcount. Routes will not be cancelled. They will thin. Same suburb, fewer stops, lower output per hour worked, and a contractor who can carry fewer drivers. That process generates no layoff notice. It generates shifts that stop appearing on the schedule and a peak season that stops adding people.
Delivery-station sortation is second. Small parcels get sorted and loaded by hand at the station. The drone share launches from the site directly and skips that step.
Seasonal hiring is the first readable signal. Amazon’s Q4 surge hiring is a line item in U.S. retail employment. The nearly-500-town target is set for the end of 2026, which puts it on top of peak. The size of this year’s seasonal announcement is the first quantifiable read on any of this.
New roles exist on the other side, at a different order of magnitude. Site ground operations, drone maintenance technicians, Part 135 operational compliance. Real jobs, and new ones. They ask for a certificate and a training program rather than a license and stamina. Moving from the driver’s seat into one of them is not a shift change.
Timing
The company set the pace itself: nearly 500 towns by year-end, 175 square miles per site, and Amazon states the sites are primarily suburban. Exposure lands in suburbs and small towns first, not in dense urban cores.
A driver working DSP routes inside a covered suburb should feel it in Q4 2026 scheduling. Station sortation trails that. Urban delivery work is largely untouched in the near term.
One follow-on metric is worth watching: cost-to-serve in Amazon’s next quarterly disclosure, alongside the headcount in its seasonal hiring announcement. The first tells you whether a drone is actually cheaper than a person. The second tells you whether Amazon believes the answer.
External sources: Amazon News, Prime Air expanding to nearly 500 U.S. cities and towns; Amazon News, $1.9B Delivery Service Partner investment