As of November 7, 2026, Walmart has committed to 300 additional Zipline drone delivery sites under a new multi-year agreement announced on November 5, with the company saying the number has room to grow. Walmart describes the programme as the largest retail drone delivery operation in the country and says it has completed more than 2 million deliveries since work with Zipline began in 2021. Deliveries arrive in as fast as 30 minutes where the service is available, drawn from tens of thousands of eligible items.
Three hundred sites, two million drops
The number that matters for availability teams is not 300. It is tens of thousands of eligible items sitting inside a promise measured in minutes. A drone launch pad does not create shelf depth; it exposes whatever depth the host store already has, to a customer who has been told the order will arrive before dinner. That is a different operational test from a next-day parcel, where a short pick can be absorbed by the network.
The same day, Zipline announced a $1.5 billion raise at a $21.5 billion post-money valuation, led by Paradigm, WCM Investment Management and Layer Global. Zipline says it has flown 150 million commercial autonomous miles without a major safety incident, and that its median Dallas-Fort Worth order in October arrived 19 minutes after checkout against 30 minutes or more for conventional delivery.
A minutes promise is an availability promise
The eligible assortment is deliberately narrow
Drone delivery works because the basket is small. Walmart positions it as the fastest tier for a handful of items, sitting above 30-minute-or-less, express at one hour or less, on-demand at three hours and scheduled windows. Each tier carries a different shelf requirement: a scheduled order can be sourced from a wider radius, while a drone order has to be picked from the store underneath the flight path.
Store density sets the radius
Walmart's advantage is proximity rather than warehouse automation. With 90% of Americans living within 10 miles of a store, a drone pad turns existing inventory into minute-scale fulfilment without building a dark store. But that also means on-shelf availability at the host store becomes the ceiling on the promise: a line that is short at 4pm cannot be rescued by a second site 40 miles away.
Five delivery tiers, five shelf requirements
Treating the tiers as one service is the most common planning error. They draw on different inventory positions, they tolerate different substitution rates, and they fail in different ways. A store that looks healthy on a blended availability number can still be failing badly on the fastest tier, because that tier only reads the lines it advertises.
| Tier | Promise | Where stock comes from | Shelf requirement |
|---|---|---|---|
| Drone delivery | Minutes | Host store under the flight path | Depth on a narrow, high-intent list |
| 30 minutes or less | Half an hour | One store or micro-site | Pick accuracy on fast movers |
| Express | One hour or less | Store plus nearby replenishment | Substitution rules that hold |
| On-demand | As soon as three hours | Regional mix | Buffer stock across stores |
| Scheduled | Chosen window | Wider network | Forecast accuracy over days |
An availability checklist for a minutes promise
When the promise shrinks to minutes, the measurement unit has to shrink with it. National availability hides exactly the variance that decides whether a customer trusts the service next week, because two households a few miles apart can be served by different stores with different gaps. The reporting unit should be the store, and the review unit should be the line.
| Field | Example | Keep it because |
|---|---|---|
| Store id | 4521 | Availability is a store property, not a national one |
| Eligible line list | 412 SKUs | Defines what the promise actually covers |
| Short picks per shift | 23 | Catches gaps before customers do |
| Cancel rate by reason | Out of stock 61% | Separates shelf failure from payment failure |
| Days of supply | 1.8 days | Flags lines that cannot survive a spike |
Why proximity beats building more warehouses
The economics of a store-based drone network rest on avoiding capital. Zipline's decision to put new money into US manufacturing suggests aircraft supply, not shopper demand, is the binding constraint for the next two years, and that is a signal worth watching if you plan coverage by metro. Airspace approval is the other gate: the FAA's proposed Part 108 rule would set a framework for routine flights beyond visual line of sight.
For brands selling through Walmart, the practical implication is unglamorous. A drone site does not need a new listing; it needs the existing listing to be accurate at line level, with pack sizes and images that survive a narrow eligible assortment. Availability depth on the fastest-moving lines is the asset that a minutes promise monetises, and it is the one most retailers still measure with a national average.
FAQ
Does drone delivery need separate inventory?
No. Orders are picked from the store that hosts the launch site, so the eligible assortment is limited by what that store actually has on the shelf at the time of the order.
Why is 30 minutes harder than next-day delivery?
A next-day network can absorb a short pick from another node. A 30-minute promise has no second node, so a missing line becomes a cancellation in front of the customer.
What should availability reporting use as its unit?
The store, reviewed at line level. National averages hide the variance between two catchments served by different stores on the same evening.
Where does the 300-site rollout stand?
Walmart has not named states, metros or a schedule. The agreement is multi-year with potential to expand, and Wing remains a second drone partner.
Sources
Facts and figures in this article come from the public reports listed below. Definitions and sample sizes should be checked against the original documents. Any calculation shown here is an illustration of method rather than advice on pricing, promotion or investment, and every claim can be traced back to the linked source.
