As e-commerce volumes climb and last-mile costs bite, large retailers are rebuilding around store-first, AI-enabled fulfilment networks that turn existing footprints into active digital infrastructure.
In Brief
- A clear pattern is emerging: dense store estates are being recast as primary e-commerce and same-day fulfilment nodes, with specialist hubs added only where density breaks.
- Companies are reconfiguring layouts, inventory flows, labour models and AI control layers so that stores, DCs and automated CFCs operate as a coordinated, digital network.
- This pattern tightens capital and operating constraints: breakeven economics depend on ruthless productivity, selective node specialisation and constant trade-offs between speed, cost and in-store experience.
The Underlying Pattern and Stakes
The common problem is simple to state and hard to solve: digital and rapid-delivery demand are growing faster than the cost base can tolerate. Pure-play warehouse networks struggle with freshness and density, while store estates built for walk-in traffic were never designed to handle thousands of picked orders and gig deliveries a day.
Across Target, Albertsons, Lowe’s and Woolworths, a distinct operating pattern is taking shape. Rather than building stand-alone e-commerce infrastructures at scale, they are pushing a store-first configuration, then bolting on automated hubs and specialist nodes only where volumes and density justify it. AI and data platforms sit over the top, orchestrating what is picked where, routed how, and at what service level.
This is not generic omnichannel talk. It is a structural repositioning of stores as active infrastructure in the fulfilment network, with capital, labour and systems being redirected accordingly.
How Companies Are Converging
The anchor move is to treat stores as the primary last-mile network. Target is explicit that more than 97 percent of its sales are fulfilled by stores and is putting most of its 5 billion dollar capital budget and 1 billion dollar operating reinvestment into those sites. Albertsons leans on roughly 2,270 neighbourhood stores, describing them as having effectively solved the last mile, with 30-minute flash delivery and a store-based model that is now near breakeven for e-commerce. Woolworths’ supermarkets carry the bulk of picking for a digital grocery business growing in the mid-teens, with pickup alone reaching 42 percent of online orders.
Alongside that store-first stance, all three grocers are adding targeted automated capacity where store picking starts to degrade economics or experience. Woolworths has opened the Auburn customer fulfilment centre in Sydney with capacity for 60,000 orders per week and is building a mirror site in Melbourne North, relieving high-density catchments where store aisles were becoming de facto warehouses. Its Moorebank NDC and RDC and a forthcoming semi-automated chilled DC in Sydney complete a modernised backbone, while stores continue to handle most pickup and sub-60-minute missions. Albertsons, by contrast, is pushing remodels and retrofits to add cold and hot holding in stores rather than large dedicated e-commerce sheds, keeping automation tighter and closer to demand.
General merchandise players are following a similar pattern with different volumes. Target’s Chicago test deliberately designates certain stores as fulfilment specialists for next-day parcels while allowing others to focus purely on in-store trade. That node specialisation is now being expanded to more markets, allowing a few stores in each area to accumulate pick density and shipping volume while the rest operate as lighter hubs for same-day pickup. Lowe’s is doing something comparable in its own way: front-end transformations have enlarged BOPIS areas across the fleet, while freight-flow redesign and early-morning stocking are tuned so there is capacity to serve Pro customers and handle online orders at open.
AI is the cross-cutting enabler. Albertsons is using OpenAI-based tools to optimise store pick paths, apply AI to fresh order writing and control shrink via Vision AI at self-checkouts. Woolworths reports a 15 percent reduction in picker walking distance from algorithms deployed across stores. Lowe’s is rolling out AI-driven full-shelf replenishment and Freight Flow 3.0, which use real-time data to prioritise what comes off the truck and what must hit the shelf before doors open, while Target’s AI-powered personalisation and Circle data are used both to shape demand and to determine how, and in which node, that demand is served.
Finally, all four are leaning on loyalty and digital programmes as operational sensors, not just marketing tools. Albertsons’ 48 million-plus loyalty members, Woolworths Everyday Rewards penetration above 70 percent of sales, Target Circle and Lowe’s Pro programmes all provide item-level visibility into planned and executed baskets, enabling more precise forecasting, promotion targeting and capacity planning for same-day slots and pickup windows.
Operating Model Mechanics
Under a store-first, AI-enabled configuration, the roles of network nodes are being redrawn quite precisely.
Stores carry dual roles as both showroom and micro-fulfilment point. That forces new layout and back-of-house designs: Target is touching parts of store floor pads that have not changed in over a decade, rebuilding back rooms and pickup zones so that BOPIS orders can be staged without blocking aisles. Albertsons is using remodel cycles to carve out e-commerce operational space and to add refrigeration and hot holding so that 30-minute flash delivery can be supported without service breakdowns. Woolworths has added more than 200 Direct to Boot Now sites on top of 750 standard drive-up locations, effectively formalising parking-lot real estate as a throughput-constrained fulfilment area.
Above stores, DCs and CFCs absorb the heaviest digital flows where it makes sense. Woolworths’ Auburn CFC and Moorebank complex handle a slice of ambient and online demand for Sydney, freeing local stores to focus on higher-margin missions and fast-turn items. In Melbourne, a new CFC will replicate Auburn’s automation. These hubs are engineered for high-throughput, routinised picking and packing; store stock is then reserved for same-day pickup, rapid delivery and top-up missions. Target’s parcel network uses stores as the shipping node, but next-day coverage depends on aggregating volume into a subset of stores with packing capacity and direct linehaul access.
Inventory positioning follows these role definitions. In practice, this kind of configuration typically requires:
- A minimum set of fast movers and fresh SKUs held in each store to support both walk-in and same-day orders.
- Bulkier, slower-moving and long-tail items shifted either into automated CFCs, drop-ship arrangements or marketplace models.
- Dynamic rules in order management systems that route each order to the node that can serve it at the lowest cost for the promised speed.
Albertsons and Woolworths are applying classic warehouse optimisation techniques to stores: multi-order batching, pick-path sequencing and real-time prioritisation of high-value or time-critical items. Both are explicit that in-store picking productivity is the main lever moving e-commerce towards breakeven. Lowe’s Freight Flow 3.0 goes further upstream, altering which pallets are broken down first overnight so that by the time Pro customers arrive early in the morning, the items they care about most are already on the shelf.
Contracts and commercial structures complement the physical design. Woolworths and Albertsons use service fees and tiered convenience pricing to differentiate cost-to-serve by speed and mode: sub-60-minute and flash delivery carry higher charges, while pickup nudges customers into the lowest-cost path. Target’s Circle 360 subscription bundles unlimited same-day delivery with membership fees and higher wallet share, effectively amortising delivery costs across a broader relationship. Lowe’s uses its NAHB partnership and Pro Extended Aisle to push more volume into job-site deliveries and supplier-direct shipments, reducing reliance on stores for heavy fulfilment while still orchestrating the order.
Data and orchestration layers sit across all this. Each company is converging on cloud-native platforms where transaction, inventory and capacity data are centralised. That allows AI tools to prioritise restocking (Lowe’s full-shelf initiative), direct store teams to the bays with the highest lost-sales risk (Target’s store operating reset), or generate fresh orders in produce with multidimensional inputs on seasonality and waste (Albertsons’ AI tools co-designed with store managers). Woolworths’ WooliesX and Cartology, Target’s Roundel and Lowe’s media network all monetise the same data streams while feeding back into forecasting and promotional plans that the supply chain must support.
Risk, Constraints and Trade-offs
The economics of a store-first e-commerce network remain finely balanced. All four companies report margin pressure from mix shifts into digital and rapid channels. Albertsons’ gross margin fell by 63 basis points year-on-year once fuel and LIFO are stripped out, largely due to digital and pharmacy growth. Woolworths’ combined food retail EBIT declined by more than 13 percent even as WooliesX digital margins edged up to 4.5 percent, a clear sign that the store side is carrying the heavier cost load. Target and Lowe’s are blunt that the only way to fund price investments, higher wages and technology is through billion-dollar productivity agendas that attack cost-to-serve in detail.
Dual-running and commissioning costs add a second layer of constraint. Woolworths expects around 110 million Australian dollars of supply chain commissioning and transition costs in both the current and following financial year, and similar levels again as its chilled DC and second CFC come online. Only from the late-2020s does it expect net benefits from the automated network. For now, this is a drag that has to be absorbed while still delivering price cuts, wage increases and digital capacity.
Node specialisation and marketplace usage manage some of this tension but introduce others. Target’s move to concentrate parcel fulfilment in a subset of stores lowers per-order costs and unlocks next-day coverage, yet it also creates a tiered network where some stores carry heavier operational complexity than others. Keeping labour, space and service standards balanced across those tiers is not trivial. Target’s marketplace growth above 30 percent in categories like furniture and rugs reduces its balance-sheet exposure and store congestion, but raises questions about how third-party fulfilment performance meshes with same-day and next-day promises.
Security and shrink are another non-trivial trade-off. Woolworths has seen stock loss accelerate to levels not seen for some time and is spending on exit gates, trolley locks and other controls. Albertsons and Target have addressed shrink with AI cameras and store design changes. These interventions protect inventory and margin but can add friction and capital burden, and they need to be harmonised with the flow of online pickers and gig drivers moving through stores.
Finally, there is the tension between working capital and resilience. Woolworths deliberately increased inventory days by 1.6 to protect availability during industrial action and to pull in seasonal stock earlier; that tied up cash but prevented deeper service failures. Albertsons uses its largely domestic, 90 percent local sourcing base as a hedge against tariff volatility, while Target’s experience of half a billion dollars in one-off tariff and inventory costs shows how sharp the adjustment can be when sourcing and pricing decisions lag policy shifts.
Operational Self-check
Seen together, these moves expose a set of uncomfortable but necessary diagnostics for any store-led network that is now carrying e-commerce volume at scale.
- Are store roles explicitly tiered in systems and labour models so that some can act as high-volume shipping nodes without overwhelming in-aisle service in the rest of the fleet?
- Do order management rules genuinely minimise end-to-end cost for a promised speed, or are legacy priorities (e.g. nearest store first) quietly eroding productivity as volumes rise?
- Where automation and CFC capacity have been added, is there a clear path and timeline for winding down dual-running costs and rebalancing inventory between hubs and stores, rather than allowing parallel networks to persist by default?
What This Pattern Signals
The structural pattern across these companies is a decisive move away from separate e-commerce and store networks towards a single, store-anchored system augmented by targeted automation. Stores are being re-engineered as infrastructure, not just as retail destinations, with AI-enabled control layers deciding what they pick, what they stage and what they ship.
If this configuration persists, network and sourcing design will continue to favour dense, owned or long-leased urban and suburban locations that can serve both walk-in and last-mile roles. CFCs and automated DCs will be deployed sparingly, where volumes are dense enough to amortise automation and to relieve stores without fragmenting inventory excessively. Marketplace and drop-ship models will expand in categories where physical storage and in-store handling costs are hardest to justify.
This looks less like a temporary response to pandemic-era spikes and more like a durable reset of how retail supply chains are wired. The capital commitments are multi-year, the AI and data platforms are being built as permanent control layers, and the commercial models around subscriptions, fees and media are being tuned to sustain economics over time. The hard work now is not proving the concept, but executing it at scale without losing either cost discipline or the in-store experience that made these networks valuable in the first place.
This article is based on recent earnings reports and public disclosures from the companies referenced.
