Rising volume, faster delivery promises and thin margins are pushing large retailers toward a new warehouse and checkout model built on automated upstream nodes and store-based micro-fulfilment.
In Brief
- Automation-heavy regional facilities now feed stores that double as high-throughput picking hubs, turning the store network into a distributed warehouse layer.
- Companies are rebalancing centralised fulfilment centres, stores and third-party capacity, while layering AI over labour, inventory and checkout to raise throughput without matching capex.
- This pattern tightens the margin for error on density, mix and working capital: where automation is placed, how hard stores are pushed, and how much inventory is held at each node become existential choices.
The Underlying Pattern and Stakes
Across food, home improvement and electronics, large operators are quietly redefining what counts as a warehouse and what counts as checkout. Central facilities are being automated rapidly, but they are no longer the sole engine of e-commerce growth. Stores, clubs and smaller formats are increasingly engineered to behave like mini warehouses and high-speed handoff points.
Walmart, Kroger, Lowe’s and Best Buy are converging on a model built on three shifts: automated upstream distribution to reduce handling cost, stores acting as fulfilment nodes for delivery and pickup, and redesigned front ends where checkout no longer throttles flow. Rather than a generic digital transformation, this is a structural change in how physical assets move units at scale while keeping capital intensity under control.
How Companies Are Converging
The most visible convergence is the decision to put stores at the centre of the fulfilment architecture. Kroger now describes its stores as the primary way to reach customers quickly, with delivery from 97 percent of locations in under two hours and delivery volumes surpassing pickup for the first time. Walmart reaches 95 percent of the population within three hours, with 35 percent of store-fulfilled orders arriving in less than that window and a growing share of Sam’s Club members able to get three-hour delivery. Best Buy fulfils 70 percent of online orders within two days and is opening smaller formats designed for both selling and local fulfilment.
Automation upstream is the second shared mechanism. Walmart now routes freight to roughly 60 percent of U.S. stores from automated distribution centres, while about half of its e-commerce fulfilment volume runs through automation. Kroger is reviewing its automated fulfilment network site by site, recognising that economics depend heavily on local demand density. Lowe’s is redesigning freight flows into stores while integrating acquisitions like Foundation Building Materials, and Best Buy is targeting supply chain and reverse logistics as efficiency pools to fund digital initiatives.
A third area of convergence is the use of AI and robotics to make existing stores behave more like engineered warehouses. Kroger uses AI to cut shrink and improve sell-through, supporting two-hour pickup with more accurate inventory. Best Buy narrows assortments and configures constrained categories to keep core SKUs in stock while trimming the long tail. Walmart equips over a million associates with handhelds and computer vision to locate inventory in stores, while agentic tools like Sparky aim to increase basket size and drive more fulfilment volume through the network.
Finally, all four are rethinking the checkout and front-of-house experience because it now sits inside the same flow constraint as picking and staging. Lowe’s has completed a front-end transformation across its fleet, expanding buy-online-pickup areas and freeing hours from tills into aisle service and fulfilment. Sam’s Club uses Scan & Go and exit arches to reduce friction at the door, while Best Buy leans on vendor-funded expert labour and immersive zones that pull complex transactions off the main checkout spine.
Operating Model Mechanics
At the heart of this model is a redefined division of labour between central nodes and stores. Automated distribution centres at Walmart and Kroger push palletised or case-ready loads into stores, reducing backroom handling. Stores then act as the last planning layer: they hold short-cycle inventory, host picking for delivery and pickup, and, in some cases, stage job-site or installation orders, as Lowe’s does for Pro customers.
Inventory strategy is changing with that role shift. Walmart grew inventory about half as fast as sales while e‑commerce approached a quarter of revenue, signalling tighter min–max rules and more reliance on marketplace and fulfilment services to carry long-tail assortment risk. Kroger keeps underlying gross margin roughly flat despite broad price investment by using AI to improve fresh ordering, reduce throwaways and tighten seasonal buys. Best Buy’s response to memory-component shortages is to pre-buy constrained stock where it matters, narrow computing assortments to configurations that hit budget points and become a preferred partner for vendors when allocation is tight.
Routing logic is becoming more nuanced. Orders are steered not just by geography but by service promise and node capability. Walmart’s sub-hour express deliveries depend on allocating orders to stores with real-time inventory confidence and the right labour pattern on shift. Kroger’s two-hour promise is delivered through a mix of in-house fleets and partners like Instacart, with the store’s backroom acting as the staging dock. Lowe’s stages Pro orders in waves, using its Extended Aisle interface into supplier catalogues and new features that split job-site deliveries into immediate and scheduled drops.
Checkout mechanics are equally operational. As BOPIS and curbside grow, front ends become staging areas rather than pure transaction lanes. Lowe’s expanded pickup zones and Freight Flow 3.0 sequencing mean overnight teams hit the highest-priority SKUs first, leaving morning staff free for early Pro traffic and pickups. Best Buy’s remerchandising of computing to the centre of the store in dozens of sites and the insertion of outlet and Yardbird sections into existing boxes reflect a similar logic: concentrate high-touch categories and clearance where fulfilment and reverse logistics can be managed without clogging standard lanes.
Where details are not explicit, this pattern typically requires a few non-negotiables:
- A single order management and inventory visibility spine that can decide, in real time, whether a central facility or a store should serve a given order at a given promise.
- Store labour models that separate picking, staging and checkout tasks, with AI or rules engines prioritising what gets done first by hour of day.
- Clear rules on which SKUs must always be present at local nodes versus which can be virtualised through marketplace or supplier-direct extended aisle.
Commercial structures are adapting as well. Walmart’s fulfilment services product, already used by just over half its marketplace sellers, centralises 3P inventory in its network in exchange for fees, increasing carton density and utilisation. Lowe’s Pro Extended Aisle and contractor partnerships, including a tie-up with the home builders’ association, introduce contract-based flows that sit on top of its retail network, while FBM’s commercial sales into data centres keep volumes steady when residential cycles soften. Best Buy’s Best Buy Business and experiential services move large-ticket orders and technicians through the same network that carries consumer laptops and televisions.
Risk, constraints and trade-offs
The model is not free. It compresses margin for error around where automation is located and how hard stores are driven as warehouses. Kroger’s review of its automated fulfilment sites is an admission that capital-deep nodes only pay if fed with dense, repeatable demand; underperforming markets can turn automation into stranded cost. Costco is not in this sample, but the same density logic now applies across rivals building big-box automation: wrong geography or wrong mix leaves expensive square footage competing with a store network already being used as micro-fulfilment.
Store-based fulfilment itself carries labour and complexity trade-offs. Walmart’s promise of sub-three-hour delivery to almost the entire U.S., and Kroger’s two-hour coverage from nearly all stores, raise pick density and staging requirements inside locations designed for shopper flow. Without sufficient automation upstream and AI-enabled tasking in the aisles, fatigue and error rates can climb, eroding both customer experience and shrink.
Inventory is another tension point. Best Buy’s decision to bring in as much constrained computing inventory as it can, while narrowing the range and tuning configurations, ties up working capital precisely as memory costs climb. At the low end of its guidance, management is explicit that constrained supply, not just demand, would cap sales. Lowe’s holds inventory flat even after adding about half a billion dollars of stock from acquisitions and experiencing higher tariffs, thanks to SKU rationalisation and AI tools, but that demands ruthless discipline on what remains in the assortment.
Margin mix is shifting too. Pharmacy and e‑commerce grow faster than the rest of Kroger’s business and carry lower gross margins; FBM and Artisan Design Group bring wholesale distribution into Lowe’s with structurally thinner unit economics. All four companies are leaning on alternative profit pools – advertising, membership fees, retail media and fulfilment services – to rebalance the P&L. For Walmart, nearly one third of operating income in the most recent quarter came from advertising and membership; Best Buy expects around 30 basis points of gross margin improvement next year largely from Ads and Marketplace, even as product margins stay flat under promotional pressure.
Capital allocation sequence is the final friction. Walmart approaches peak supply chain automation and remodel spend at around three and a half percent of sales, expecting SG&A leverage from automation to show through while e‑commerce continues to grow near 25 percent a year. Kroger’s strategic e‑commerce review is not yet in guidance, meaning the financial upside of any network reconfiguration lies beyond the near term. Lowe’s accepts about 50 basis points of annualised operating margin dilution from integrating FBM and ADG in exchange for scale and cross-selling, betting that a billion dollars of productivity gains per year can more than cover the gap.
Operational Self-check
A few simple checks expose whether this warehouse-and-checkout model is actually present or only aspirational:
- The proportion of digital orders consistently fulfilled from stores or other frontline nodes, and whether those locations have clear inventory and labour rules distinct from pure retail.
- The share of volume flowing through automated facilities and whether those facilities sit in markets with enough density to avoid stranded fixed cost.
- The extent to which high-margin, asset-light profit streams already offset the incremental cost of faster fulfilment and front-end reconfiguration.
What This Pattern Signals
Taken together, these moves signal a shift away from viewing central warehouses as the sole fulcrum of growth. Automation upstream reduces cost and error, but the real growth engine is a dense store and club network acting as a distributed warehouse layer. Checkout itself is being redesigned as part of the same operational flow that includes picking, staging, scanning and rapid delivery.
If this configuration persists, network design will increasingly start from customer promise and work backwards: which nodes can support sub-two-hour delivery, which assortments must sit where, and which flows can be virtualised to marketplaces and supplier-direct catalogues. Capital will favour automation that feeds those flows and remodels that make stores more pick-friendly over standalone fulfilment build-outs without clear density.
This is unlikely to be a temporary response. The economics of rapid delivery, the scale of e‑commerce relative to total sales, and the rise of advertising and membership income all point to a long-term model where high-volume growth relies on turning every viable box into part of the warehouse, and every front end into a flexible handoff point for both people and parcels.
This article is based on recent earnings reports and public disclosures from the companies referenced.
