All posts
Case Studies

Computing in the Store: What Chick-fil-A, Walmart and McDonald's Built at the Edge

By

October 25, 2026

7 min read

Chick-fil-A runs a Kubernetes cluster in every restaurant, managed by GitOps. Walmart runs regional clouds plus edge clouds in stores and clubs. McDonald's is putting Google Distributed Cloud into thousands of restaurants. Why retailers put compute in stores, how they manage thousands of tiny sites, and the AWS equivalents.

Case Study
Edge Computing
Retail
Kubernetes
GitOps
EKS Hybrid Nodes

A restaurant or store has to keep taking orders when its internet link drops, and kitchen equipment, sensors and point-of-sale systems need answers in milliseconds. That is why the largest operators run computing on site, and why managing thousands of small sites has become its own engineering discipline.

What three companies built

Company

What they run

How they manage it

Chick-fil-A

A small bare-metal Kubernetes cluster in each restaurant, connecting the restaurant's devices; the fleet has since grown to roughly 2,800 clusters

GitOps: each cluster pulls its desired state from Git in the cloud, so thousands of sites update without anyone logging in

Walmart

A "triplet" of its own cloud plus two public clouds in US East, West and Central regions, plus edge clouds in stores and clubs

A common platform layer so developers deploy the same way to any of them

McDonald's

Google Distributed Cloud hardware and software in thousands of restaurants, announced December 2023

Cloud-managed edge, running cloud and McDonald's own applications and AI locally

Want a second opinion on this for your system?

Book a free 30-minute call. You'll leave knowing what we'd do first and roughly what it costs, whether or not you hire us.

The pattern underneath

  • Tens of containers on thousands of clusters, the reverse of most Kubernetes estates. Chick-fil-A's engineers noted that this scale breaks a lot of standard tooling.

  • Pull, not push. Sites fetch their configuration when they can reach the cloud, so a site that was offline catches up on its own.

  • Local first. The store works without the cloud; the cloud gets data when the link is up.

  • Identical sites. Every location runs the same image and configuration, so a fix is one commit, not 2,000 tickets.

Building this on AWS

Need

AWS option

Kubernetes on store hardware, managed from the cloud

EKS Hybrid Nodes (on-site machines join an EKS cluster) or EKS Anywhere (clusters that run entirely on site)

Device software and messaging at the edge

AWS IoT Greengrass

AWS-managed hardware on site

AWS Outposts servers

Fleet configuration

GitOps with Argo CD or Flux, including through EKS Capabilities

Data back to the cloud

Kinesis, IoT Core or S3 uploads when connected

hoolco Monthly Review

Once a month: what changed in tech that month and what it means for you, from cloud and AI to security and the tools you build with. Every item links to its source. No sales pitch, unsubscribe in one click.

As of October 25, 2026. Provided as is, for general information, and accurate to the best of our knowledge on the date shown. Not legal, compliance, tax, financial or other professional advice. Prices, limits and vendor services change; check current vendor documentation before relying on anything here. Terms of use.