Amazon's next delivery contest with Walmart will be decided partly before a driver leaves a building. Stocking the right goods close to households can shorten the journey, but it also commits inventory and operating capacity before the next order arrives. The question for investors is whether more local availability creates enough useful demand to pay for that commitment.
Reports this week put Amazon's expansion plans back in focus. PYMNTS, summarising reporting that traces to Business Insider, described a proposed major increase in same-day hubs. It also relayed Amazon's warning that internal projections were preliminary and could change. Those plans should not be read as completed facilities or a confirmed profit forecast. The company's public descriptions of its existing delivery system provide a firmer basis for assessing the economics.
The delivery promise begins on a shelf
Amazon explains that forecasting helps position popular products near customers, with some sites combining stages of the fulfilment journey. It also describes converting existing rural delivery stations into hybrid facilities that hold inventory. This is more than making the final vehicle travel faster: it changes where an order can begin.
Consider an illustrative household purchase that includes detergent and a fresh ingredient for dinner. If both products are available locally, one dispatch may satisfy the basket. If one is held far away, the retailer must offer a different arrival time, split the shipment or lose part of the sale. Proximity creates options, while the actual assortment determines which options are usable.
Local stock has a cost, however. Spreading goods across more locations can leave the wrong product in the wrong catchment area. Replenishment, storage and forecasting become central to the delivery proposition. Perishables add a time constraint: a unit that expires before sale cannot recover its cost merely because the warehouse is near the customer. These are operational trade-offs, not evidence that Amazon currently faces a particular loss rate.
A shorter route can still be an expensive route
Distance matters, but so does the number of orders that can share labour and transport capacity. A nearby hub serving scattered requests may save miles while carrying substantial idle time. A busier location may spread fixed costs across more completed orders. Neither outcome can be inferred from the number of buildings announced.
Speed also limits flexibility. A retailer promising a narrow delivery window has less time to combine compatible orders than one offering a broad window. That creates a tension between convenience and efficient dispatch. The relevant comparison is the contribution from the basket after the extra picking, handling and delivery effort, not simply whether a customer receives it earlier.
Demand frequency is part of the attraction. In a company grocery update revised in May, Amazon said customers adding fresh groceries to same-day orders shopped about twice as often as those who did not. This is a company-reported association, not proof that adding groceries doubles a customer's spending or profit. More frequent shoppers may already differ from other customers. The observation supports investigating repeat use, while leaving the causal and financial questions open.
Walmart already owns a different starting point
Walmart's May announcement of delivery in 30 minutes or less across 33 US markets described a service built around its store network. Eligibility depends on factors including basket size, driver availability and distance. That qualification matters: a nationwide physical presence does not mean every product reaches every address at the same speed.
Stores can bring inventory close to households while serving shoppers who arrive in person. That creates an existing base of activity from which to offer delivery. It also creates competing demands on shelf availability, employee time and space. A dedicated fulfilment site may be designed around order processing, whereas a store has to support several shopping journeys. Neither design is automatically cheaper for every basket or neighbourhood.
The competition is therefore not adequately described as a race to match a facility total. An Amazon hub and a Walmart store are different operating units. Comparing them requires attention to selection, throughput, coverage and the work each building performs. An address on a map is not a standardized measure of delivery capacity.
Count productive orders, not buildings
As expansion proceeds, the informative questions will concern utilisation and incremental demand. Are customers placing additional profitable orders, or moving an existing purchase into a more expensive delivery channel? Does better local availability reduce split shipments? Can sites sustain service quality without accumulating slow-moving stock? These are analytical tests, not claims that the companies publish every answer today.
The counterargument to caution is credible: reliable convenience can deepen a shopping habit, and a denser network can improve over time as demand develops. Early spare capacity may be intentional. But that possibility still needs to become observable performance. Amazon's reported ambitions signal a strategic commitment to local fulfilment; the return on that commitment will emerge from baskets, inventory turns and productive routes, rather than from a preliminary hub count alone.