How Supermarkets Cut Food Waste with AI

How much food waste do supermarkets generate?

On June 30, 2026, Japan's Ministry of Agriculture, Forestry and Fisheries (MAFF), the Consumer Affairs Agency, and the Ministry of the Environment released their FY2024 food waste estimates. Japan's total food waste came to 4.61 million tonnes, a slight decrease from the previous year. Published annually, this figure — tracked against data going back to FY2000 — serves as a key indicator of how far Japanese society has come in reducing food waste.

Of the business-sector food loss, food retail (supermarkets, convenience stores, etc.) accounts for 480,000 tonnes — less than food service (700,000 tonnes) or food manufacturing (1.10 million tonnes), but it draws particular attention because it occurs closest to consumers. Overordering to avoid stockouts, and removal of items past internal sell-by deadlines, are the main causes.

How much has it dropped compared to FY2000?

Per the published data, business-sector food loss has fallen 57% compared to FY2000, while household food loss has fallen 48%. The steeper decline on the business side likely reflects earlier, more organized efforts by major retail and food-service chains — more sophisticated ordering systems, revised best-before labeling, and so on. Household-sector reductions have progressed more slowly, suggesting there's still room to build stronger "eat it all, use it all" habits at home.

Where food waste occurs in retail

• Removal of items whose best-before or use-by dates are approaching, past the store's internal sell-by deadline
• Unsold fresh food and prepared meals due to bad weather or lower-than-expected foot traffic
• "Extra" ordering driven by staff wanting to avoid the lost sales from running out of stock
• Removal for reasons unrelated to actual quality, such as damaged packaging or mislabeled dates

Prepared foods, bread, and boxed lunches in particular have a short window between production and sale, making demand especially hard to forecast. Foot traffic swings sharply with weather, nearby events, and the day of the week, so orders based on experience alone tend to run short or long — which is exactly where AI demand-forecast ordering delivers the most value.

The Food Loss Reduction Act and supermarkets' role

Japan enacted the Food Loss Reduction Promotion Act in 2019, establishing that national and local governments, businesses, and consumers each have a role to play in reducing food waste. Under this law, MAFF and the Consumer Affairs Agency publish annual estimates for both the business and household sectors, and collect and publicize initiatives by industry. As the point of direct contact with consumers, supermarkets play a key role in putting the law's goals into practice — through in-store awareness campaigns, markdown policies, and smarter ordering.

How AI demand forecasting is transforming ordering

The measure delivering the biggest results in supermarket food waste reduction is AI-based demand-forecast ordering. By learning from past sales alongside weather, day-of-week, and local event data, AI automatically suggests the right order quantity for each store — turning what was once a veteran staffer's intuition into a quantified, streamlined process.

Challenges with traditional ordering

Before AI, ordering was typically done manually by store staff based on past sales ledgers and personal experience. While a veteran could order with high accuracy, judgment varied widely between staff, and know-how was often lost when someone was transferred or retired. On top of that, with supermarkets carrying anywhere from thousands to tens of thousands of SKUs, judging demand item by item took an enormous amount of time.

What kind of data AI learns from

• Several years of POS sales data (by item, time of day, day of week)
• Weather data (temperature, precipitation probability, typhoon information, etc.)
• Calendars of nearby events and school activities
• Promotional schedules such as sales and flyer distribution
• Inventory data and delivery lead times

By combining these data sources, AI can make judgment calls close to human intuition — for example, "fewer bento customers tomorrow because of rain" or "more demand for rice balls because of a nearby sports day."

Concrete results from adopting companies

These figures are all drawn from each company's own published case studies, and they share a common pattern: shorter ordering time and reduced food waste achieved together. By forecasting demand variability in advance, AI cuts down on both "sold out, lost sales" and "overproduced, thrown away."

What shorter ordering time really means

The benefits of AI demand-forecast ordering go beyond waste reduction alone. When ordering time drops from 3 hours to 25 minutes, as in Yaoko's case, staff can redirect that time toward customer service, shelf-stocking, and store cleanliness. For a retail industry grappling with labor shortages, automating ordering is seen as a way to tackle food waste and improve working conditions at the same time.

Hurdles to adopting AI ordering

Despite the clear benefits reported, adopting AI demand-forecast ordering comes with real hurdles: integrating it with existing core systems, preparing historical data for each store, and the transition period needed for staff to trust and rely on AI's suggestions. Newly opened stores or those that have just been renovated, in particular, have little historical data to draw on, so it can take time before forecast accuracy stabilizes.

Spreading to small and midsize supermarkets

AI demand-forecast ordering initially spread mainly among major chains, but the rise of cloud-based services has made it more accessible to smaller supermarkets that lack the resources to build their own systems in-house. Subscription-based services with lower upfront costs have also emerged, and the technology is expected to become a standard ordering tool regardless of store size.

The role and limits of markdowns and clearance sales

Marking down items as their best-before or sell-by date approaches — known as "mikiri hanbai" (clearance selling) — has long been a standard food waste countermeasure. In recent years, more supermarkets have adopted "dynamic pricing," using AI and sensors to automatically adjust discount levels based on time of day and inventory levels.

How dynamic pricing works

• Electronic shelf labels (e-paper) allow remote, bulk price updates, enabling fine-grained control over discount timing
• Discount rates are raised in stages as the sell-by deadline approaches
• Discounts are linked to inventory data, prioritizing items likely to be thrown away

That said, markdowns alone have limits. Applying discount stickers takes manual labor, and if shoppers start waiting for discounts before buying, it can actually reduce opportunities to sell at full price. That's why many supermarkets combine AI ordering — to avoid overproducing in the first place — with markdowns to sell off whatever is left over.

The link with "front-shelf picking" campaigns

Alongside markdowns, Japan's Ministry of the Environment and MAFF encourage consumers to practice "temae-dori" — picking items from the front of the shelf, which are closer to their sell-by date, when the food will be consumed soon anyway. This reduces in-store waste without waiting for a discount. Markdown policies on the retailer's side and this kind of purchasing behavior on the consumer's side work together to maximize the effect of clearance selling.

Adjusting how much prepared food is made

Another approach that doesn't rely on markdowns is scaling back production of prepared foods and boxed lunches based on AI forecasts in the first place. While large last-minute discounts near closing time can generate buzz, pricing plans built around discounting can also erode profit margins. Many chains are therefore investing in better demand forecasting with the ultimate goal of "selling out without relying on markdowns."

How electronic shelf labels improve efficiency

Traditionally, applying markdown stickers meant staff picking up each item by hand and attaching a sticker one by one. With electronic shelf labels, prices can be updated in bulk from a headquarters or backroom terminal, allowing fine control over discount timing while cutting the labor needed for sticker application. In an industry facing severe labor shortages, electronic shelf labels are gaining attention as an investment that tackles food waste and staffing efficiency at once.

Markdowns and understanding freshness labels

For clearance selling to work smoothly, consumers also need to correctly understand what "safe to eat by" really means. A best-before date indicates when food is "at its best," not a hard safety cutoff like a use-by date. Without this distinction being well understood, shoppers may avoid items simply because the best-before date is approaching, undermining the effectiveness of markdowns. In-store signage and municipal awareness campaigns play a role in communicating this basic knowledge alongside markdown policies.

Delivering surplus food through food-sharing apps

Beyond in-store markdowns, another growing approach is "food sharing" — delivering surplus food to consumers via smartphone apps. A leading example is TABETE, operated by CoCooking Inc., where restaurants, bakeries, and supermarket deli counters list items they want to sell before closing, and consumers pick them up at a discounted price.

The advantage for supermarkets is that they can cut disposal costs while gaining a touchpoint with new customers, all at once. Consumers benefit from buying food at a lower price, and this mutual benefit is why municipalities also encourage residents to use these services.

Municipalities promoting adoption

Several municipalities — including Tokyo's Adachi Ward, Utsunomiya City, and Sakai City — encourage residents on their official websites to use food-sharing services. For local governments aiming to reduce waste and food loss, partnering with private services offers awareness benefits without additional cost, and this kind of collaboration is likely to keep expanding.

Benefits and caveats for users

For users, the appeal of food-sharing apps lies not only in getting food at a lower price, but also in discovering bakeries and delis they wouldn't otherwise visit. On the other hand, listings appear irregularly and pickup windows are limited, so there's no guarantee you'll find exactly what you want. It's best used as a way to "take advantage of whatever good deal happens to show up that day," rather than as a guaranteed shopping channel.

How well food sharing fits supermarkets

Food-sharing apps were initially used mainly by independently run bakeries and restaurants, but in recent years supermarkets' deli and bakery counters have increasingly adopted them too. Rather than a bulk discount right before closing, announcing surplus items in the app ahead of time and spreading out pickup slots also has operational benefits — avoiding checkout congestion while selling through inventory in a planned way.

The structural challenge of ehomaki and Christmas cakes

Large-scale disposal of seasonal products has long been one of the most closely watched aspects of supermarket food waste. In January 2019, MAFF issued a notice to food retail industry associations ahead of ehomaki season, calling for sales matched to actual demand and encouraging a shift toward reservation-based or fully made-to-order sales.

Why reservations alone don't eliminate waste

• Demand from same-day, non-reserved customers is hard to forecast due to weather and foot traffic
• Stores, fearing stockouts, tend to produce extra on top of confirmed reservations
• Long production and preparation lead times make it difficult to respond to last-minute demand shifts

7-Eleven continues to encourage advance reservations through in-store reservation boards and awareness materials featuring MAFF's mascot "Rossnon." Even so, recent surveys have found a certain amount of unsold stock even at stores using reservation systems, leaving improved forecast accuracy and a higher reservation ratio as ongoing challenges.

The same issue applies to Christmas cakes and osechi

The same structure seen with ehomaki also applies to Christmas cakes and osechi New Year dishes. Both are items that "won't sell if you miss the day or season," yet require days to weeks of preparation lead time. Many supermarkets and department stores work to raise reservation ratios so that production and procurement volumes move closer to confirmed reservation numbers, but eliminating the portion produced speculatively for walk-in customers remains difficult, leaving a structural pattern of annual waste.

Combining forecasting with reservation systems is key

A realistic approach to reducing seasonal product waste doesn't rely on any single measure. Reservation systems secure confirmed demand in advance, while AI demand forecasting — accounting for weather and local events — keeps same-day supplementary production to a minimum. This two-pronged approach is currently considered the best available solution for seasonal products like ehomaki and Christmas cakes.

What happens to unsold seasonal items

For surplus that occurs even with thorough reservations and forecasting, retailers take a multi-step approach: same-day markdowns, switching to frozen sale for items that can be frozen, donation to food banks, and recycling into animal feed or compost. The basic philosophy of seasonal food waste countermeasures rests on two pillars: preventing waste from occurring in the first place, and making the best possible use of whatever surplus does occur.

What consumers can do to help reduce food waste

Even as supermarkets advance their own initiatives, household food loss (an estimated 2.24 million tonnes in FY2024) still accounts for nearly half of the total. Depending on how you shop, consumers can directly contribute to reducing supermarket food waste too.

1. Front-shelf picking: choose items from the front of the shelf that are closer to their sell-by date
2. Make active use of clearance and "imperfect" product sections
3. Buy surplus items through food-sharing apps
4. Check your fridge before shopping to avoid buying more than you need
5. Only pre-order the quantity of seasonal items you actually need

For more specific ways to reduce food waste at home, see our related article "Japan's Household Food Waste: 2.36 Million Tonnes a Year — The Three Main Causes and What to Do Today."

Supermarket measures and consumer behavior are two sides of the same coin

No matter how accurate AI demand-forecast ordering becomes, if consumers keep reaching for items at the back of the shelf, the ones at the front will go unsold and ultimately be thrown away. Conversely, if consumers actively choose front-shelf items or clearance products, supermarkets can sell through inventory with smaller markdowns, which in turn improves order optimization. Food waste reduction only reaches its full potential when businesses' technology investments and consumers' everyday choices work in sync.

The three main causes of household food waste

Household food loss is generally attributed to three main causes: direct disposal (throwing away unopened food after it expires), leftovers, and excessive trimming (such as peeling away too much of a vegetable). Rethinking how you shop at the supermarket is also a first step toward preventing these forms of waste at home. Buying only what you need and using ingredients according to a plan helps reduce food waste across both the business and household sectors.

Parallels with food loss in seafood

Supermarket food waste countermeasures apply not only to produce and prepared foods, but also to the seafood counter. Fish and shellfish, which spoil quickly, are often discarded before ever reaching the sales floor due to being off-spec or misshapen. For more, see "Reducing Seafood Food Loss: Spoilage, Off-Spec Waste, and Frozen Food Utilization."

Growing use of underutilized fish and off-spec produce

Alongside more sophisticated demand forecasting, some supermarkets are actively turning previously discarded underutilized fish and off-spec vegetables into products. Using misshapen vegetables as ingredients for prepared foods, or packaging abundant but little-known fish species for sale, offers a dual benefit: less waste and a broader product lineup. Related efforts are also covered in "The Distribution Systems Rescuing Off-Spec Produce and Underutilized Fish."

Conclusion: cutting waste through both AI and consumer behavior

As the FY2024 statistics show, food waste from the retail sector remains substantial at 480,000 tonnes. As cases from Yaoko, Ito-Yokado, and Lawson demonstrate, AI demand-forecast ordering is a powerful tool that can improve ordering efficiency and cut waste at the same time. On top of that, markdown sales and food-sharing apps like TABETE serve as a safety net for surplus that AI alone can't prevent.

As seasonal products like ehomaki show, technology alone can't bring waste to zero. Food waste reduction moves forward steadily when supermarkets' systemic efforts combine with small everyday actions from consumers — like front-shelf picking and reserving seasonal items in advance.

Going forward, we can expect AI demand-forecast ordering to spread further among small and midsize supermarkets, more stores to adopt food-sharing apps, and reservation ratios for seasonal products to keep rising. Beyond just tracking the statistics, each individual choice we make in daily shopping — which item to pick today — adds up to support the retail sector's broader push to cut food waste.