Amazon’s physical grocery push shows why AI matters for omnichannel retail. Learn practical AI steps for demand, inventory, and personalisation in Singapore.
Amazon is doing something it used to laugh at: building the kind of massive physical store people associate with Walmart.
Reuters reported that Amazon is planning a 225,000-square-foot “mega-store” outside Chicago, designed to sell groceries and general merchandise, while doubling as a distribution point for same-day delivery. Analysts expect Amazon’s Q4 physical store revenue to hit US$5.9B (+5.4% YoY), even as AWS remains 18% of total revenue. Meanwhile, Walmart keeps stacking advantages from scale: 4,600 stores, 90% of Americans within 10 miles, and 26.5M Walmart+ members (as of 2025, Morgan Stanley).
This isn’t just a US retail rivalry. It’s a practical case study for our “AI dalam Peruncitan dan E-Dagang” series: the winners are building omnichannel systems where stores, delivery, data, and loyalty work as one. For Singapore retailers—where rent is high, customer expectations are higher, and delivery is table stakes—AI tools aren’t a “nice-to-have.” They’re the difference between running tight operations and bleeding margin.
One-liner worth remembering: Omnichannel isn’t about being everywhere. It’s about making every channel share the same brain.
Why Amazon is betting on physical stores again
Amazon’s move is a strategy correction, not a nostalgia trip.
Amazon has experimented with physical retail for a decade: Amazon Go (2016), Whole Foods acquisition (US$13.7B, 2017), Amazon Fresh launch (2020). The Reuters piece notes Amazon decided to close all Amazon Fresh and Amazon Go locations and convert some into Whole Foods Market stores—a blunt signal that “interesting tech” doesn’t automatically create a scalable store model.
The mega-store concept is about one thing: economics of grocery and last-mile delivery.
Grocery isn’t a category—it's a customer lifetime value engine
Amazon seller consultant Martin Heubel (quoted in the article) makes a point many businesses underestimate: grocery and FMCG buyers tend to have higher customer lifetime value because they purchase frequently. If you win the weekly basket, you get repeat transactions, richer data, and more cross-sell opportunities.
For Singapore businesses, this maps cleanly to:
- subscription-style replenishment (staples, household essentials)
- loyalty perks that actually change behaviour
- basket-building (bundles) that reduce delivery cost per order
Physical stores are also delivery infrastructure
Amazon wants what Walmart already has: a dense network that makes delivery cheaper.
Walmart claims 90% of the US population lives within 10 miles of a store, and those stores function as fulfillment points for pickup and same-day delivery. That’s not just convenience—it’s a structural advantage in last-mile costs.
Amazon said it delivered 4 billion grocery and everyday items in same/next day in 2025. The missing ingredient is proximity at scale. A mega-store that also acts as a distribution centre is basically a hybrid of:
- store
- micro-fulfilment node
- returns point
- customer acquisition channel
The real battle: last-mile economics and data, not “online vs offline”
Retailers often frame omnichannel as a branding story (“we’re available everywhere”). The Amazon vs Walmart fight shows it’s actually an operations story.
Here’s the simple truth: delivery gets expensive when you’re far away and your basket size is small. The business that reduces distance and increases basket efficiency wins.
Walmart’s playbook is operational compounding
Walmart didn’t “catch up” by copying Amazon’s website. It used:
- stores as fulfillment hubs
- membership (Walmart+) to increase retention
- strong unit economics for pickup and local delivery
The Reuters piece cites 28% e-commerce growth in Walmart’s most recent quarter and 26.5M Walmart+ members. Membership matters because it’s a behavioural contract: customers shift more of their shopping to the brand to “make the membership worth it.”
Amazon’s playbook is data compounding
Amazon’s edge is still data and automation. Even its physical ambitions are ultimately about feeding a system that can predict demand, allocate inventory, and shorten delivery promises.
For Singapore SMEs, you don’t need a mega-store to apply the same logic. You need:
- accurate demand signals
- inventory visibility
- smarter replenishment and pricing
- better customer personalisation
That’s where AI business tools in Singapore become practical, fast.
Where AI actually helps: 4 omnichannel problems to solve first
If you’re running retail or e-commerce in Singapore, you can treat Amazon vs Walmart as an audit checklist. The goal is to make decisions based on data, not gut feel.
1) Demand forecasting and replenishment
Answer first: AI reduces stockouts and overstock by forecasting demand at SKU-level using sales, seasonality, and promotion signals.
Grocery is brutal: margins are thin, and waste is real. AI forecasting tools can incorporate:
- day-of-week patterns (weekend spikes)
- campaign uplift (CNY bundles, Ramadan, year-end gifting)
- weather-sensitive demand (cold drinks, ready-to-eat)
- store vs online channel differences
Practical Singapore example: If you run a speciality grocer or minimart chain, forecasting can tell you which outlets should get more fresh items ahead of weekend condo traffic, while reducing spoilage at slower locations.
2) Inventory management across store + online
Answer first: The fastest omnichannel wins come from a single inventory view and AI-driven allocation rules.
Many retailers still operate “two inventories”: one for store shelves and one for online orders. That creates ugly outcomes—online oversells, store stockouts, and staff time wasted checking back rooms.
AI tools can help by:
- predicting which SKUs should be reserved for online
- recommending inter-store transfers
- optimizing safety stock by location
- flagging shrinkage anomalies
Benchmark mindset: Walmart’s physical footprint is a fulfillment advantage only because their store inventory is operationally usable. If your stock accuracy is weak, your store becomes a liability, not an asset.
3) Personalisation that doesn’t feel creepy
Answer first: AI personalisation works best when it’s tied to customer intent—replenishment, dietary preferences, household size—not just “products you might like.”
This is one area where Singapore retailers can punch above their weight. With the right CRM and recommendation engine:
- suggest replenishment reminders for staples
- bundle complementary items (sauce + noodles + proteins)
- personalise promotions by neighbourhood or outlet
Tip: Start with 3 segments (not 30). For example:
- value seekers
- premium/health-conscious
- convenience-first
Then refine.
4) Omnichannel marketing and measurement
Answer first: AI improves ROI when it connects ad spend to downstream outcomes like repeat purchases, basket size, and store visits.
A common mistake: treating online ads as “online-only.” In reality, many people discover products online and buy in-store—especially for groceries where last-minute decisions are common.
AI-driven attribution and analytics can:
- match customer profiles across channels (within privacy rules)
- estimate incremental lift by campaign
- optimise budget allocation across Meta, Google, marketplaces, and SMS/WhatsApp
Contrarian stance: If you can’t measure repeat rate and basket margin by channel, spending more on ads is usually just paying to learn nothing.
What this means for Singapore retailers in 2026
Singapore’s retail environment pushes businesses toward omnichannel by default:
- customers expect fast delivery (same-day is increasingly normal)
- labour costs make manual ops painful
- physical space is expensive, so every square metre must “earn”
Amazon’s mega-store plan and Walmart’s footprint advantage highlight a shared lesson:
You don’t win by choosing online or offline. You win by building a system that makes fulfilment and retention cheaper every month.
“People also ask”: Do I need physical stores to run omnichannel?
Not necessarily.
You can create an omnichannel experience with:
- a showroom + pickup point
- partner pickup lockers
- pop-ups and community activations
- dark stores / micro-fulfilment spaces
The requirement isn’t a big store. It’s operational integration.
“People also ask”: What’s the first AI tool I should implement?
Start where you bleed money quietly:
- demand forecasting (waste + stockouts)
- inventory visibility (overselling + dead stock)
- retention automation (repeat rate)
A simple rule: If you don’t trust your data, fix data capture before buying more AI.
A simple omnichannel AI blueprint you can copy
If you want a practical way to move from “we sell online and offline” to a real omnichannel strategy, use this 30-60-90 approach.
30 days: Get one version of the truth
- unify product catalogue and SKU naming
- clean customer data (duplicate profiles)
- baseline KPIs: stockout rate, shrinkage, delivery cost/order, repeat rate
60 days: Automate one high-impact workflow
Pick one:
- AI replenishment recommendations
- personalised replenishment campaigns
- promotion planning based on predicted lift
90 days: Connect operations to marketing
- segment customers by behaviour
- measure campaign lift on repeat purchase and basket margin
- optimise inventory allocation for promoted SKUs
This is how you turn AI in retail and e-commerce into profit, not dashboards.
Where the Amazon vs Walmart story is heading
Amazon is trying to buy itself something Walmart already has: proximity. Walmart is trying to keep up with what Amazon is best at: data-driven efficiency and customer lock-in.
If you’re running retail in Singapore, the useful takeaway is the middle ground: build a tighter loop between demand signals, inventory, and customer engagement. AI tools make that loop faster and cheaper, whether you have one outlet or fifty.
If you’re exploring AI business tools in Singapore for inventory management, customer personalisation, or omnichannel analytics, the next step is straightforward: map one workflow where you lose margin today, and automate it end-to-end.
What would change in your business if you could predict next week’s demand by SKU and location with enough confidence to buy less, waste less, and still stay in stock?
Source article (landing page): https://www.channelnewsasia.com/business/amazons-physical-grocery-push-deepens-its-fight-against-rival-walmart-5906961