Amazon says its drones will reach 500 cities by December. Permission to fly over most of them doesn't exist yet.

Amazon spent last week making a promise with a very specific number attached to it. By the end of 2026, the company said, Prime Air drone delivery will reach nearly 500 U.S. cities and towns — a roughly sixfold expansion from where it stands now — with some orders arriving in as little as 30 minutes. It is a clean, confident, quotable figure, and it did exactly what it was designed to do: it produced a wave of headlines about the year drone delivery finally goes mainstream.

Read the announcement beside the paperwork that actually governs it, though, and a gap opens up. The 500 is a target. The permission to hit it is not a target; it is a stack of separate federal approvals, most of which do not yet exist, granted under a rulebook that has not yet been finalized. When a company's ambition and its authorizations disagree this plainly, the authorizations are the document under oath.

What the number is measured against

Start with where Prime Air actually is, because the base rate is the part the headline number is designed to make you forget. Today the service operates in a handful of states — Arizona, Florida, Kansas, Louisiana, Michigan, Nebraska and Texas — with launches described as coming soon in a short list of others, including Georgia, Ohio, Illinois, Idaho and New York. Amazon says it has delivered hundreds of thousands of packages by drone this year. Hundreds of thousands sounds large until you set it against a company that ships that many packages by van in a good few minutes. This is a pilot that works, not a network that scales — yet.

The machine doing the flying is the MK30, and its spec sheet quietly narrows the promise further. It carries packages up to five pounds. It flies about 73 miles per hour, operates in light rain, and works within a radius of a few miles around a fulfillment center — a coverage area of roughly 174 square miles per site, deployed from more than 85 same-day centers. Those are real capabilities. They are also hard limits. A drone that can carry five pounds within a few miles of a warehouse, in weather that is not too bad, is a useful complement to the van. It is not a replacement for it, and 500 cities does not change the physics of what the aircraft can lift or how far it can go.

The honest document is the FAA docket

Here is the sentence Amazon's announcement does not contain, and the one that decides whether the number is real: every new Prime Air site still requires its own FAA environmental review and its own beyond-visual-line-of-sight authorization. The approvals do not come in bulk. They come one location at a time, each with its own paperwork, its own review, its own opportunity for delay. A promise of 500 cities is, translated, a promise to clear 500 individual regulatory gates — or close to it — in a matter of months.

The rulebook that is supposed to make bulk approval possible is the FAA's proposed Part 108, the framework that would replace today's waiver-by-waiver system with a standardized pathway for flying drones beyond the operator's line of sight. Part 108 is the thing the entire industry has been waiting on for years. It was sent to the White House regulatory office for review on July 10. That review can run up to 90 days, which puts the earliest plausible completion around October, and most analysts expect the final rule later this year or in early 2027 — followed by an implementation period of six to twelve months before operators can actually rely on it.

Do the arithmetic on those dates and the conclusion is unavoidable: every city Amazon reaches in 2026 will get there under the old, slow, site-by-site system. The rule that would justify a number like 500 will not be in force while the number is being filled. Amazon has announced a destination on the assumption that the road will be finished in time. That is not a lie. It is a forecast dressed as a fact, and the two are different documents.

A promise of 500 cities is, translated, a promise to clear 500 individual regulatory gates in a matter of months. The approvals do not come in bulk. — On the gap between the target and the authorizations

Amazon has made this promise before

The other reason to read the guidance skeptically is that Prime Air has a history, and it is a history of dates that arrived without the service attached. The program was unveiled in 2013 with a claim that drones would be delivering within a few years. The few years came and went, several times. What Amazon is doing now — attaching a hard, near-term number to a program that has repeatedly missed hard, near-term numbers — is a familiar move, and the familiar move deserves the familiar scrutiny. A company that says '500 by December' after a decade of slipped timelines is asking to be believed on exactly the point its own record undermines.

None of this means the expansion is fake. Amazon has clearly improved the aircraft, integrated drone launch into its existing same-day fulfillment footprint, and reached the point where the constraint is regulatory throughput rather than whether the drone flies. That is genuine progress, and it is why the announcement is more credible than the 2013 version. But 'the constraint is now the FAA's approval speed' is a very different statement than '500 cities,' and only the first one is inside Amazon's control.

Read the announcement as a message to three audiences

The most useful way to read a number like this is to ask who it is actually addressed to. Customers are the least of it. There are three audiences that matter more, and the announcement is doing a different job for each.

1. The regulator. Announcing 500 cities is a way of applying public pressure to the FAA's timeline. It converts Amazon's private wish for faster approvals into a stated commitment the agency now has to react to — a bet that a loud target speeds the paperwork.
2. The competition. Walmart has been expanding its own drone service and talking about reaching tens of millions of customers; Alphabet's Wing and Zipline are ahead of Amazon on delivery volume. A 500-city figure is a flag planted against them, whatever this year's actual footprint turns out to be.
3. The market. A concrete, ambitious, near-term number is the kind of thing that reads well in a quarter where every large company is expected to show an AI-and-automation story. The figure is a narrative asset regardless of how many of the 500 come online on schedule.

That is the strategy the press release will not state in plain words: the 500 is aimed less at the person waiting for a package than at the agency that can grant the permissions, the rivals racing for the same airspace, and the investors pricing the automation story. It is a real operational goal and a positioning move at the same time, and the positioning move does not depend on the goal being met.

The number to watch instead

If you want to know whether drone delivery actually arrives at scale in 2026, ignore the 500 and watch two duller figures. The first is how many individual BVLOS authorizations the FAA actually issues between now and December — the real throughput of the site-by-site system Amazon still depends on. The second is where Part 108 lands in the regulatory review, and whether its implementation clock starts this year or next. Those two numbers, not the press release, will decide how many of the promised cities are flying and how many are still filings.

Amazon's forecast may well prove close to right; the company has the capital, the fulfillment network and the engineering to move fast once the rules allow it. But 'once the rules allow it' is carrying the entire sentence. The confident part of the announcement — 500 cities, by December — is the part Amazon does not control. The part it controls is quieter, and it is the honest one: the drone works, the footprint is small, and the only thing standing between the two is a stack of approvals that has to clear one location at a time. When a company tells you the destination and goes quiet about the road, the road is where the story is.