Amazon announced today that it is closing every Amazon Go and Amazon Fresh store it operates. All 72 of them. The cashierless convenience stores that were supposed to revolutionise shopping. The full-service grocery stores that were going to take on Walmart. Gone.

Some will convert to Whole Foods locations. The rest will simply disappear.

And if Amazon, with its vast resources, its army of engineers, and its decade of experimentation, cannot make physical retail work under its own brand, we need to ask a harder question: can anyone?

Today's announcement is not an isolated retreat. It is the final chapter in a systematic failure that stretches back years.

2015: Amazon opens its first physical bookstore. The company that killed bookstores decides to build bookstores. Irony aside, the concept had promise. Use data to curate selection. Show customer ratings on shelf tags. Merge online and offline.

2018: Amazon 4-Star stores launch. A curated selection of highly-rated products from the website. Same data-driven approach. Same attempt to bring the algorithm into the aisle.

2018: Amazon Go debuts. The marquee innovation. No checkouts. No cashiers. Walk in, grab what you want, walk out. The technology was genuinely impressive. Ceiling-mounted cameras and AI that could track every product you touched.

2020: Amazon Fresh grocery stores begin rolling out. Full-service supermarkets with Dash Carts and Just Walk Out technology. The assault on traditional grocery was supposed to begin here.

2022: Amazon closes all 68 bookstores, 4-Star locations, and Pop Up shops. The explanation: focusing on other formats.

2023: Amazon Style, the fashion store experiment, quietly dies.

2024: Amazon removes Just Walk Out technology from its own grocery stores. The flagship innovation, the one that was supposed to change everything, gets ripped out.

2025: Amazon shutters its UK Fresh stores.

Today: The remaining Go and Fresh stores are announced for closure.

Amazon's own announcement contains the admission buried in corporate language. They "haven't yet created a truly distinctive customer experience with the right economic model needed for large-scale expansion."

Translation: the stores did not make money, and there was no clear path to making them profitable at scale.

The analyst community is less diplomatic. Neil Saunders of GlobalData put it plainly: "The main reason behind the decision is that neither Fresh nor Go stores were delivering the sales needed to make them fully economic. Nor were they producing growth trajectories that might convincingly reverse that position."

The numbers support this. Physical stores accounted for just 3% of Amazon's revenue in Q4 2024, and almost all of that came from Whole Foods, which Amazon acquired rather than built. The formats Amazon created from scratch never gained meaningful traction.

Meanwhile, Amazon's same-day delivery of perishable groceries has grown 40x since January 2025. Fresh groceries now make up nine of the top ten most-ordered items through that service. Customers who add fresh groceries to their same-day orders shop twice as often as those who do not.

The data is stark. Customers want Amazon to deliver groceries. They do not want to visit an Amazon grocery store.

The flagship technology deserves its own autopsy.

Just Walk Out was supposed to be the future of retail. Arrays of cameras tracking every product. AI systems identifying what you picked up. Automatic charging as you exit. No queues. No friction. The holy grail of convenience.

The reality was messier.

The technology required significant infrastructure investment. Ceiling cameras. Shelf sensors. Backend processing. The cost per store was substantial, and that cost had to be justified by either higher sales or lower labour expenses.

Neither materialised in the way Amazon hoped.

In larger grocery stores, where customers fill entire trolleys, the technology struggled with accuracy. Amazon eventually admitted that Dash Carts, essentially smart shopping carts, worked better for bigger shops. Just Walk Out was better suited to quick trips with a handful of items.

The telling pivot came in April 2024, when Amazon began removing Just Walk Out from its own stores while simultaneously expanding its licensing to third parties. The technology now operates in over 360 locations globally, but almost all of them are stadiums, airports, hospitals, and university campuses. Places where customers grab a handful of items and speed matters more than price.

The company insists this is a success story. They point to the 112% sales increase at Lumen Field in Seattle, where the technology reduced queue times during NFL games. They highlight the rollout to more than 30 university campuses globally.

But these are niche applications. The original vision, a network of cashierless stores as ubiquitous as Starbucks, never happened. Amazon Go peaked at around 30 locations before retreating. The dream of replacing every corner shop with a frictionless Amazon experience died quietly.

Here is what the Amazon story reveals, and what the data increasingly confirms.

Physical retail foot traffic in the United States fell 12% in the second half of 2024. Some estimates suggest a 17% year-over-year decline in certain categories. The pandemic accelerated a shift that was already underway, and there is no evidence of a meaningful reversal.

E-commerce now accounts for 21% of all retail transactions globally. That figure is projected to reach 23.7% by 2030. Some analysts project that by 2040, online purchases could represent 95% of all retail, with physical stores accounting for just 5%.

The growth rate of e-commerce is more than double that of physical retail. Every year, the gap widens.

This does not mean physical stores will disappear entirely. But it does mean the economics are shifting in ways that make Amazon's retreat look less like failure and more like recognition of reality.

Consider what Amazon is doing instead:

Same-day grocery delivery is now available in over 5,000 US cities and towns. The service added perishables in 2025 and has seen explosive growth. Amazon is testing "Amazon Now" for 30-minute delivery of essentials.

Whole Foods is expanding. Over 100 new stores planned in the coming years, plus smaller "Daily Shop" locations. But critically, Whole Foods was already a successful retailer before Amazon bought it. Amazon is not building a grocery brand. It is leveraging one it acquired.

The Chicago supercenter is still in the works. Amazon received approval for a Walmart-style big-box store in an Illinois suburb. But even here, the strategy appears to be learning from others rather than inventing something new.

The pattern is clear. Amazon excels at logistics, delivery, and digital commerce. When it tries to build physical retail experiences from scratch, it fails.

This brings us to the uncomfortable intersection of Amazon's retreat and the broader trajectory of retail automation.

The Bureau of Labor Statistics projects that retail trade will lose 181,900 jobs by 2034. That is the largest numerical decline of any industry sector in the American economy. Automation, consolidation, and e-commerce are cited as the primary drivers.

In the retail sector specifically, 65% of cashier and checkout jobs are expected to face automation pressure. Self-checkout expansion at major retailers could replace tens of thousands of positions. Sam's Club's AI verification rollout alone is projected to eliminate 12,000 cashier jobs.

Amazon's retreat from physical stores might seem to contradict this narrative. If even Amazon cannot make cashierless stores work, perhaps the automation threat is overstated?

Not quite.

Amazon's technology did not fail because it could not replace workers. It failed because it could not attract enough customers to justify the investment. The economics of physical retail are broken not because automation is too expensive, but because foot traffic is declining regardless of how sophisticated your checkout technology is.

The stores that remain will automate. They have to. With fewer customers walking through the door, labour costs become an even larger percentage of operating expenses. The pressure to cut staffing will intensify, not diminish.

Amazon's Just Walk Out technology is finding success in exactly the environments you would expect: high-traffic venues where speed matters and customers have limited alternatives. Sports stadiums. Airport terminals. Hospital cafeterias.

The technology works. The customer base for standalone retail stores does not.

For communities that hosted Amazon Go or Fresh stores, today's news means job losses. Amazon says it will help affected employees find roles elsewhere in the company, including at Whole Foods or in its logistics network. Those who choose to leave will receive severance.

But the broader story is what Amazon's retreat signals about the viability of neighbourhood retail in an increasingly digital economy.

If the world's largest retailer, with unlimited capital and the best logistics infrastructure on the planet, cannot make a go of convenience stores and grocery shops, what chance does the independent operator have?

The answer, counterintuitively, may be "better than you think."

Amazon's stores failed in part because they tried to apply e-commerce logic to physical spaces. Data-driven curation. Algorithmic inventory. Technology-forward checkout. These approaches assumed that customers visited stores for efficiency. But efficiency is what delivery provides. Stores need to offer something else.

The retailers thriving in 2025 and 2026 are those offering experiences that cannot be replicated online. Foot Locker redesigning stores around try-on experiences and sneaker culture. Sephora investing in beauty studios and personalised consultations. Lego creating interactive play spaces.

Experiential retail is not dead. Transactional retail, the kind Amazon tried to build, increasingly is.

Amazon will continue to dominate grocery through delivery. The company claims to be a top-three US grocer with over $150 billion in gross sales. Whole Foods will expand. Same-day delivery will reach more communities.

The Just Walk Out technology will find its niche in venues and campuses, licensing to third parties rather than powering Amazon's own stores. It may eventually become profitable as a B2B product, even if it never achieved its original consumer vision.

For the rest of the retail industry, the message is harder to ignore. Physical stores face structural headwinds that technology alone cannot overcome. The shift to digital is not a temporary pandemic effect. It is a permanent reallocation of how consumers spend their money and their time.

The stores that survive will be the ones that give people a reason to show up. That reason cannot be convenience. Amazon proved that.

What is your take? Is Amazon's retreat a unique failure or an early signal of what is coming for all of physical retail? We would like to hear from those of you building in this space.

Related reading from Major Matters: