What is shutting down, and why
Amazon confirmed to Grocery Dive on October 13 that it is closing an unspecified number of retailer-specific grocery storefronts on Amazon.com. It named US partners Winn-Dixie, Metropolitan Market, and Weis Markets, plus Morrisons in Europe, without disclosing which of those partnerships are actually affected by the change. Amazon says it will direct shoppers to other grocery options on its marketplace and will notify customers before the changes take effect. A limited number of jobs are affected, and Amazon says it is helping those employees find other internal roles.
The company's stated rationale is refreshingly candid for a platform announcement: separate storefronts forced shoppers to navigate different fees, delivery windows, and checkout flows depending on which retailer's storefront they happened to land on. That description, coming from Amazon itself rather than a competitor's talking point, amounts to an admission about what its multi-partner grocery setup had become. A spokesperson offered only that the company is 'always evaluating how we can best serve our grocery customers,' which is the kind of line companies use when the honest answer is that an architecture decision did not scale the way it was supposed to.
The fragmentation problem Amazon built for itself
Letting individual grocers run their own storefront inside Amazon's marketplace looked, on paper, like a reasonable way to extend reach without each retailer building its own e-commerce stack. In practice, it meant a shopper buying from Winn-Dixie one week and Weis the next could hit entirely different rules for delivery windows, minimums, and fees, all inside what was nominally one shopping experience. That inconsistency is exactly the kind of friction Amazon has spent two decades training customers not to tolerate.
The deeper issue is that grocery storefronts, unlike a standard third-party marketplace listing, carry operational complexity that does not compress well into a one-size-fits-all template: perishables, delivery windows tied to local fulfillment capacity, and pricing that varies by region and banner. Amazon apparently concluded that supporting that complexity per-partner, rather than standardizing it, was not worth the inconsistent experience it produced for shoppers who expect Amazon-level polish regardless of which banner they are technically buying from.
What Amazon is building instead
Rather than walking away from grocery partnerships, Amazon is repositioning what it offers them. The company says future arrangements will lean on formats like Amazon Lockers, co-locating operations with partner stores or warehouses, and selling AWS services into partner operations. That is a shift from retail real estate, a storefront with Amazon's name on the checkout, to infrastructure services, where Amazon gets paid for compute, logistics, and pickup points rather than for owning the shopper relationship directly.
It is a more defensible business model for Amazon. Infrastructure and logistics services scale without the customer-experience liability of a storefront that reflects on Amazon's own brand when a partner's fulfillment falls short. It is also a more honest one: Amazon was never going to subsidize inconsistent grocery experiences indefinitely just to claim partner logos, and shifting to AWS and logistics revenue is the version of this relationship Amazon actually wants to be in.
The lesson for grocers who rented the shelf
For Winn-Dixie, Weis, Metropolitan Market, and Morrisons, the practical question now is how much of their incremental online grocery volume depended on a storefront that Amazon can retire on its own schedule, with partners finding out roughly when customers do. Amazon says it will communicate before changes land, but the underlying lesson for any retailer using a platform partner's checkout infrastructure is the same: that channel functions as a lease whose terms the landlord can rewrite on its own schedule, and the tenant finds out roughly when its own customers do.
This is a familiar pattern for CTOs who have watched marketplace and payments platforms reshape partner terms before. The specific mechanism here, grocery storefronts, is less important than the structural reminder: any revenue running through infrastructure a company does not own carries platform risk that should be modeled explicitly, not discovered when the platform decides to simplify its own experience. That modeling should include a credible estimate of how long it would take to replace the channel entirely, because the answer for most of these partners is almost certainly longer than the notice period Amazon will give them before the storefronts actually go dark.
Why this matters beyond grocery
Amazon still offers same-day fresh grocery delivery directly in more than 2,300 US cities and towns, plus parts of Tokyo and London, with stated plans to expand further in Japan, the UK, and additional countries. That is the part of this story that should worry grocery partners most: Amazon is actively consolidating control over the grocery experience it offers directly, while repositioning partners as infrastructure customers rather than storefront co-owners.
That sequencing, open the platform to partners, learn where the experience breaks, then pull the customer-facing layer back toward owned channels while monetizing partners through infrastructure instead, is a pattern worth recognizing early. It shows up across retail media networks, marketplace APIs, and now grocery storefronts, and it tends to benefit the platform far more than the partners who built their go-to-market plans on top of it. Any enterprise currently negotiating deep integration with a dominant platform, in grocery, in agentic commerce, or anywhere else, should assume the same arc is coming and price it into the deal rather than treating it as a surprise two years in.
What to do with this before the next partnership deal
Any CTO or CIO negotiating a platform partnership, whether with Amazon, a retail media network, or an agentic commerce protocol provider, should treat storefront ownership and infrastructure ownership as separate negotiating lines, not a bundled package that gets signed off in one meeting. A partner that can simplify its own experience by discontinuing a storefront format is a partner whose roadmap decisions can show up as a line item on next quarter's revenue, with limited notice and even less input from the retailer on the receiving end.
The practical move is to track what percentage of digital grocery or e-commerce revenue flows through infrastructure a company does not control, and to have a credible fallback, an owned app, a direct delivery fleet, or a different aggregator, ready before that number gets too large to unwind quickly. Amazon just gave every grocery partner a real-world deadline for doing that math, and the retailers who had already started that work before this announcement landed are the ones who will treat it as a minor operational update rather than a scramble.
