A format that lasted three years
In 2023, Amazon opened its site to a new kind of partnership: local grocery and specialty retailers could run dedicated storefronts on Amazon.com, reaching customers who did not have a Prime membership. Weis Markets offered same-day delivery across parts of New Jersey, Virginia, Pennsylvania, and Maryland through its storefront. Lucky Supermarkets ran same-day, two-hour delivery in select San Francisco Bay Area cities. Winn-Dixie made its full grocery assortment available to shoppers in the Jacksonville and Orlando areas of Florida. Each partner set its own delivery fees, delivery windows, and speed options inside its storefront.
According to Chain Store Age, Amazon is now ending that model, with the current storefronts shutting down across several countries in the coming months. An Amazon spokesperson told the outlet the company is 'simplifying the shopping experience' in response to customer feedback, and that Amazon is 'always evaluating how we can best serve our grocery customers.' Customers using the affected storefronts will get advance notice before any closure. For a three-year-old partnership format, that is a short runway, and it is worth asking what the per-partner delivery variability the spokesperson cited actually cost Amazon in support tickets and customer confusion before the company decided to kill it.
This is the second partner format Amazon has unwound
The grocery storefront shutdown is not Amazon's first retreat from a partner-run retail format. In 2024, Amazon stopped developing Amazon Today, a service that sold partner inventory and fulfilled orders from local stores or malls, citing overlap with its own same-day and one-day delivery offerings. That is now two multi-year partner programs Amazon has built, run for a few years, and then wound down in favor of consolidating control over the shopping experience itself. For partners who joined either program, the lesson is the same: a storefront inside Amazon's marketplace is a tenancy, not a permanent fixture, and the terms of that tenancy can change based on Amazon's own operating priorities rather than the partner's performance.
The pattern also reveals something about how Amazon treats fragmentation at scale. Running dozens of partner storefronts, each with distinct delivery fees and windows, is operationally expensive to maintain and support, even if each individual partnership performs fine on its own terms. Amazon's own grocery business has had years to mature since 2023, including its Whole Foods integration and newer formats, which likely reduced its need to lean on third-party storefronts to fill grocery selection gaps. Retailers partnering with any large platform should assume that once the platform's own capability catches up, the partnership that once filled a gap becomes redundant from the platform's point of view.
Consolidation, not retreat, from grocery
This closure comes alongside other recent moves that show Amazon pulling its grocery strategy toward fewer, more directly controlled formats rather than exiting the category. In January, Amazon said it would close its Amazon Go and Amazon Fresh physical stores and convert various locations into Whole Foods Market stores, while keeping the Amazon Fresh online grocery service running. Amazon also continues to test new physical formats, including Amazon Grocery in Chicago and a 'store within a store' concept at a Whole Foods location in Plymouth Meeting, Pennsylvania, plus a planned mass-format physical store.
Read together, these moves describe a company narrowing its grocery footprint to formats it owns and operates directly, whether online or physical, rather than ones built through third-party partnerships. The third-party storefront model gave Amazon fast geographic coverage in markets where it had no delivery infrastructure of its own, but that advantage erodes once Amazon's own fulfillment network and store formats mature. For a CTO watching Amazon's grocery moves as a signal for where retail infrastructure investment is heading, the direction is toward owned, vertically controlled grocery experiences and away from federated marketplace models that require reconciling multiple partners' delivery promises.
What this means if you are the partner, not the platform
Weis Markets, Lucky Supermarkets, and Winn-Dixie each built operational workflows, likely including inventory feeds, fulfillment integrations, and customer service processes, specifically to support their Amazon storefronts. Those integrations now have to be unwound or repurposed on short notice, and the customer relationships built through those storefronts revert to whatever direct channels each grocer maintains outside Amazon. This is the risk every retailer takes when a platform partnership becomes a meaningful revenue or customer-acquisition channel: the platform's roadmap decisions, not the partner's performance, determine whether the channel continues to exist.
For any CTO currently negotiating a storefront, marketplace, or distribution partnership with a dominant platform, whether that is Amazon, a major marketplace, or an AI shopping agent provider, this closure is a concrete argument for building exit and data-portability clauses into the agreement up front. That means retaining your own customer data and order history independent of the platform's systems, and having a fallback fulfillment and marketing plan ready before the platform decides to simplify its experience at your expense. Three years is not a long partnership lifecycle to plan around, and Amazon has now shown that pattern twice.
The roadmap implication
For retail and grocery CTOs, the actionable takeaway is less about Amazon specifically and more about what this says about platform dependency risk in 2026 and beyond. As agentic commerce pushes more retailers to plug their catalogs into AI shopping agents and third-party marketplaces, the Amazon grocery storefront story is a useful case study in how quickly a platform can decide that federated partner experiences are not worth the operational cost, even when individual partnerships are performing well.
Any roadmap item that proposes deep integration with a single external platform, whether a marketplace storefront, an AI agent's checkout flow, or a retail media network's ad infrastructure, should carry an explicit assumption that the platform's terms will change within a three-to-five-year window, and that the integration needs an exit path. Building that flexibility in up front costs more engineering time now, but it is cheaper than the scramble Amazon's grocery partners are facing on a timeline measured in months rather than years.
