Ahold Delhaize is hard-wiring a store-first, AI-enabled fulfilment model into its network, reshaping how inventory, capacity and digital demand are orchestrated across its U.S. and European banners.
In Brief
- The group has pivoted online grocery to a store-first model, using more than 2,000 supermarkets as integrated fulfilment nodes while keeping margins close to 4 percent.
- Omnichannel remodels and a common digital platform are standardising how stores run e-commerce, self-checkout and in-store service, turning format upgrades into infrastructure for same-day delivery at scale.
- AI tools and assortment harmonisation are pushing more decisions upstream, tightening delivery slotting, routing and own-brand complexity so that rapid growth in online and price investment can be absorbed in the supply chain, not just the P&L.
The Strategic Break: Stores Become The Default Fulfilment Engine
Ahold Delhaize has made a clear structural choice in the U.S.: online grocery will be fulfilled from stores, not built on a new lattice of dedicated e-grocery warehouses. That pivot, made in 2023 and now fully embedded, underpins six consecutive quarters of double-digit online grocery growth while keeping the U.S. underlying operating margin at 4.6 percent, 20 basis points ahead of last year.
The company now leans on a network of more than 2,000 supermarkets to serve as local inventory pools and pick sites. Over the past three years, the share of orders delivered same day has risen from 65 percent to nearly 90 percent. That is not a marketing claim; it is an operating constraint that forces changes in routing, labour scheduling and backroom design.
In parallel, all five U.S. brands now sit on PRISM, Ahold Delhaize’s proprietary omnichannel platform. The rollout was completed in 2025 with Food Lion and Hannaford. PRISM standardises order capture, slotting and customer-facing options, giving the group a single backbone for new fulfilment logic rather than five parallel development streams.
The same logic appears in Europe, albeit through different tools. At Albert Heijn, double-digit online growth is being supported by an AI-driven delivery slot system that offers personalised windows at checkout based on location and order history. The system dynamically recalculates routes and slots as orders come in, designed to minimise emissions and maximise van utilisation without loosening service promises.
How The Store-first Model Works Operationally
In operational terms, moving to a store-first omnichannel model changes several core mechanics of the grocery supply chain:
- Inventory policy shifts from clear separation between store and e-commerce stock to a pooled approach. Safety stock and presentation minimums must now serve walk-in customers and same-day pickers, raising the importance of accurate demand sensing and perpetual inventory.
- Labour planning moves from fixed store staffing to a blended model where picking, packing and staging are integrated into front-of-house routines. Same-day commitments at 90 percent of orders push picking into narrow time bands that need to be reflected in scheduling and task management.
- Backroom and front-of-store layout must support order assembly and hand-off. Omnichannel remodels at Food Lion in Raleigh, Wilmington and now 153 stores in Charlotte include explicit provision for self-checkout, in-store pickup under the Food Lion To-Go brand and easy meal solutions. Stores are being rebuilt as micro-fulfilment nodes, not only as selling floors.
- Transport and routing are constrained by tighter cut-off times and denser order profiles. AI-led routing at Albert Heijn is one visible response; it allows the network to absorb rising order counts while keeping drops per route and cost per order within thresholds.
At network level, this model requires consistent system architecture. PRISM in the U.S. and AI slotting tools in Europe play that role. They centralise master data on assortments, service thresholds, order-cut off times and capacity limits, and they allow changes in one brand to be propagated as configuration rather than bespoke projects.
The decision to anchor online on stores also reduces the need for a parallel CFC estate, which several peers have built. Instead, Ahold Delhaize is committing big-ticket capital to mechanised regional distribution hubs, such as the newly announced Burlington, North Carolina facility. With more than one million square feet and proven mechanisation technology, the site, due online in 2029, is designed to feed a high-growth corridor around Food Lion with case- and piece-handling efficiency, not to pick customer orders directly.
Omnichannel Remodels as Operating Model
The Food Lion omnichannel remodel programme illustrates how the operating model is being standardised in hardware as well as software. Charlotte’s 153-store wave is the third market after Raleigh and Wilmington. In those earlier markets, average weekly sales in remodelled stores are already outpacing non-remodels, which provides a hard financial signal that the format is not merely cosmetic.
Remodels bring together three elements:
- A reworked assortment, with easy meal solutions and, progressively, more own-brand penetration.
- A front end geared for mixed journeys, with self-checkout for speed and staffed lanes where needed.
- Full e-commerce enablement via Food Lion To-Go for pickup and local delivery, integrated into day-to-day store operations rather than bolted on.
Across the Growing Together strategy period, management has committed to remodel 1,000 U.S. stores. That effectively means one in two stores will be re-platformed for omnichannel execution, locking the store-first model into the physical network.
In Europe, similar integration is occurring through network reshaping rather than a single banner remodel. Delhaize Belgium is expanding via an affiliate model for new supermarkets and closing an acquisition in the Delfood/Louis Delhaize convenience business. In Romania, Profi has added more than 200 stores over three years, with plans to ramp roll-out again from 2026 once commitments to the competition authority are fulfilled. These moves change the mix of formats, with implications for replenishment frequency, last-mile reach and franchise support, but they are being executed under a common sourcing and own-brand playbook.
AI and Assortment Harmonisation Push Control Upstream
The omnichannel network sits on top of a less visible but equally material shift in how Ahold Delhaize controls its assortments and costs. Own-brand penetration has grown across all banners, and in Europe it now sits around 50 percent of sales. Each European brand carries at least 900 ‘Price Favorite’ SKUs as a standardised value tier. That requires harmonised specifications, consistent packaging and stable supplier relationships across countries.
In the U.S., management has reviewed roughly 90 percent of categories to harmonise assortments, align product specifications and reduce supply complexity. That is a SKU policy decision at scale. Fewer, better-standardised SKUs simplify forecasting, lower working capital and make store picking more predictable in an omnichannel context. It also strengthens the commercial position with suppliers at a time when the group has committed a cumulative EUR 1 billion to price investments in the U.S. over four years.
AI is being applied primarily where variability and cost are highest: last-mile routing and delivery slotting in the Netherlands, and, by implication, replenishment and promotion timing across banners. The group is explicit that it is building foundational AI platforms, not only deploying isolated pilots. In operational terms, that means common data models, standardised events (such as promotions and weather shocks) and shared optimisation engines, which can then be tuned per banner.
Retail media is another digital layer with operational implications. The Edge platform, to be scaled across U.S. brands, will coordinate on-site display, sponsored search and in-store digital screens. While media revenue sits above the supply chain, the placement of digital prompts changes traffic patterns in stores and order mix online, which must be reflected in merchandising, planograms and replenishment logic.
Trade-offs: Margin Pressure and Regulatory Limits
The store-first, AI-enabled network is being built under explicit financial and regulatory constraints. The group continues to target an average 4 percent operating margin and 4 percent sales CAGR over the Growing Together plan. Q3 2025 delivered a 4.1 percent underlying operating margin at group level, with the U.S. at 4.6 percent and Europe at around 3.8–3.9 percent.
Those numbers sit against several headwinds:
- In Serbia, a government decree on grocery pricing will cap shelf prices until at least February 2026, compressing unit margins.
- In Romania, a 2 percent VAT increase and a shift in food voucher timing are depressing reported sales and complicating demand planning.
- Profi’s first year of consolidation is dilutive to European margins, even though synergies are reported as ahead of plan and the banner is expected to reach average European margin levels within two to three years.
- In the U.S., mix is tilting toward online and pharmacy, both structurally higher cost-to-serve. Management acknowledges that this is dilutive, and is relying on ‘careful timing of promotions’ and leverage from volume growth to offset it.
External benchmarks show similar tensions elsewhere. Best Buy, for instance, has recently retooled its control towers into execution layers and introduced automated guided vehicles in warehouses to defend operating margins as e-commerce exceeds 30 percent of domestic revenue. Woolworths in Australia is spending heavily on automated ambient and chilled DCs to support sub-24-hour delivery while explicitly flagging several years of dual-running costs before net benefits appear. The pattern is consistent: omnichannel networks can be margin-accretive, but only if automation, SKU policy and labour models are treated as part of the same design problem.
For Ahold Delhaize, the decision to keep average margin stable while stepping up price investments and absorbing regulatory shocks implies that productivity gains from assortment harmonisation, mechanised DCs and AI tools are not optional; they are embedded in the maths of the strategy.
What The Operating Model Now Enables and Constrains
Ahold Delhaize’s store-first omnichannel network and AI-enabled logistics stack now give it the ability to offer near-universal same-day fulfilment across 2,000-plus U.S. stores and dense European markets without a wholesale rebuild of its physical estate. The combination of PRISM, omnichannel remodels, AI routing and a harmonised own-brand architecture means that new service propositions and price moves can be rolled out as configuration changes rather than bespoke projects, at least within each geography.
The same design, however, locks in a dependence on store execution quality and on the stability of large-scale mechanised DC investments such as Burlington. It constrains the group to keep investing in labour models, inventory accuracy and backroom processes at a time when regulatory interventions and price investments are squeezing unit economics. The strategic bet is that a standardised, AI-enabled, store-centric network can carry those pressures while still delivering the 4 percent margin and 4 percent growth profile that management has reaffirmed.
