As of November 7, 2026, the public record on Walmart's drone build-out is short and specific. On November 5 the retailer said it had signed a multi-year agreement with Zipline to add 300 drone delivery sites nationwide, on top of a program that has completed more than 2 million deliveries since the two companies started flying together in 2021. For teams that plan store networks, the number that matters is not 300. It is the 90% of Americans Walmart says live within ten miles of one of its stores, because that proximity is the asset the new sites are being hung on.
What Walmart actually committed
The mechanics are worth stating plainly, because they decide what can be sold from a rooftop. Zipline's home delivery aircraft holds position roughly 330 feet above the ground while a pod containing the order descends on a tether into a front yard, backyard or driveway. That is a small-payload, short-hop model, and Walmart describes it as the option measured in minutes, sitting above a ladder that runs from 30-minute-or-less delivery to one-hour express, three-hour on-demand and scheduled windows. Nothing in that ladder is new; what changed is how many addresses can be reached without adding a building.
The agreement was announced in Bentonville on November 5, the same day Zipline announced a $1.5 billion raise at a $21.5 billion post-money valuation, with the capital earmarked for US manufacturing and hiring. Read together, the two releases describe a supply constraint rather than a demand question: aircraft, charging hardware, launch fixtures and flight crews have to exist before a site can be scored, and the retailer has not published which states or metros come first.
Store-based launch avoids new warehouses
Launching from an existing store means the inventory is already there, which removes the warehouse step that makes most last-mile pilots expensive. It also means the serviceable area is fixed by the store's own trade area, not by a logistics planner's ideal polygon. A rooftop site inherits the assortment of the store beneath it, so the eligible item list is a store-level decision before it is a network-level one.
Why rooftop density is a trade-area question
Most coverage models treat a store as a point that serves a ring. A rooftop site breaks the ring in a useful way, because it removes the road network from the constraint: the flight is a straight line and the drop is vertical, so the effective catchment stops being a drive-time shape and becomes a radius. The practical effect is that two stores a few miles apart can end up with very different serviceable populations once the radius is applied to household density rather than to traffic.
Payload and radius cap the catchment
A 30-minute promise is a payload promise as much as a speed promise. The item has to fit the pod and the weight has to stay inside the aircraft's envelope, which pushes the assortment toward small, light, high-urgency goods: the frosting, the creamer, the tennis balls that Walmart itself uses as examples. Categories that anchor a normal basket, such as beverages and bulk household, mostly stay on the ground, so the drone site complements the store rather than replacing a delivery shift.
Wing's 270-store plan sets a second baseline
Walmart runs a second drone partner. Wing, part of Alphabet, said in October that the two companies plan to reach more than 270 Walmart stores and nearly 20 US metro areas, with Denver and Seattle added for 2027, and Wing was reported this week to be committing up to $125 million to build delivery aircraft in Fort Worth. Two partners means two coverage maps drawn on the same store base, which is exactly the comparison a store-selection team needs: same rooftops, different radius, different payload, different launch year.
Three networks, three coverage maths
| Network | Scale announced | Stated coverage logic | Disclosed timing |
|---|---|---|---|
| Walmart + Zipline | 300 additional sites, announced Nov 5, 2026 | 90% of Americans within 10 miles of a store | Multi-year; no site list or schedule |
| Walmart + Wing | More than 270 stores, nearly 20 metros, per October statements | Metro-by-metro, Denver and Seattle added for 2027 | 2027 for the new metros |
| Zipline network, all partners | 1.5 million deliveries implied by DFW alone; 150 million autonomous miles | Metro density; median order 19 minutes after checkout in October | Operating now |
The table is the argument. A retailer choosing where to put a rooftop is not choosing a vendor, it is choosing which of three coverage curves to inherit. Zipline's own disclosure that the median October order arrived 19 minutes after checkout, against 30-plus minutes for traditional delivery, sets the service expectation that any pilot will be judged against, and the Dallas-Fort Worth figure of more than one million deliveries this year is the only public volume benchmark that exists at metro scale.
For a market opportunity model, that changes three inputs. First, the addressable population per site is a radius calculation, not a drive-time one, so household density replaces road speed in the score. Second, the addressable basket is a payload filter, so category mix per store matters more than total store sales. Third, the launch sequence is a hardware question, so the order in which sites go live will track aircraft supply, not shopper demand.
What to instrument before the sites open
The baseline is the part that gets skipped. A rooftop site that opens without a pre-launch measurement window will be judged on novelty traffic, and novelty decays. The useful comparison is the same store's fill-in basket in the four weeks before the site went live: if drone orders are incrementally new trips rather than shifted van trips, the site earns its place; if they are shifts, the retailer has bought speed with margin and learned nothing about demand.
FAQ
Does a 300-site commitment mean 300 new stores?
No. The sites are drone launch points attached to existing stores, and Walmart has not published the list. The store base is the constraint, not the expansion.
Why does a ten-mile radius matter more than drive time?
Because the aircraft flies a straight line and drops vertically, so the catchment is a circle rather than a road-network shape. Household density inside that circle replaces traffic speed in the model.
How should a brand read this without store-level data?
Track eligible item lists rather than site counts. The items that appear in the app for a drone-eligible address show which categories the format can actually carry.
