This is an English adaptation of a FoodBud historical article originally published on November 9, 2021.
FoodBud argued in November 2021 that Meituan’s next strategic lever in food delivery could be a new layer of infrastructure: smart meal lockers and related restaurant-side hardware.
The logic is familiar. Once a platform builds a high-traffic “road” and becomes structurally important to offline merchants, it can start charging for access to that road. The article noted that this dynamic applies not only to Meituan, but also to Alibaba and other large platforms accused of monopoly behavior. At the same time, FoodBud acknowledged that better infrastructure can also make the industry more efficient.
Meituan had already been moving deeper into restaurant operating infrastructure. Recent media reports at the time described “Meituan Smart Kitchen” on the merchant side.
According to on-site LED display information cited in the article, Meituan Smart Kitchen aimed to help merchants digitize store operations through hardware. Its listed products included smart burger machines, smart packing machines, and smart pickup lockers.
FoodBud’s view was that Meituan’s digitization of foodservice operations was becoming more granular: production time, cloud printers, merchant-side meal staging lockers, consumer-side pickup lockers, rider dispatch mechanisms, and other operational details. The expected result was higher end-to-end delivery efficiency.
The article then placed Meituan’s locker push in the broader context of cloud kitchens, ghost kitchens, and shared kitchens.
Cloud kitchens had become especially popular in India and North America, with some companies raising large financing rounds. Covid-19 changed how consumers interacted with restaurants, especially in the United States, where many states limited restaurants to outdoor dining or delivery. That helped accelerate a model where brands focus only on producing and delivering food ordered online.
Unlike traditional restaurants or fast-food stores, cloud kitchens do not need a full dine-in format. Multiple brands can also share the same physical operating space. Compared with conventional restaurants, the model can reduce labor and rent costs.
For startups and small businesses testing foodservice concepts, the model offered flexibility, speed, lower startup risk, and potentially attractive returns. Some restaurant operators created online-only brands, while others worked with DoorDash, Uber Eats, or providers of centralized kitchen space.
The article cited Sweetgreen’s Outpost Channel as one example of pickup infrastructure. Sweetgreen began testing pickup lockers in 2018 and had expanded to more than 1,000 by March 2020. The program was suspended during the pandemic in 2020, but as employees returned to work, the number of operating Outpost locations had recovered to 350 as of September 26.
FoodBud also cited Rebel Foods, an Indian cloud-kitchen startup, which had recently raised a $175 million Series F round. Its valuation rose from $800 million in the previous round to $1.4 billion after the financing.
Rebel Foods’ annual revenue had reached about $150 million, and the company planned an IPO within the following 18 to 24 months. Its brands included Faasos, Mandarin Oak, The Good Bowl, and Oven Story Pizza.
A founder working in India told FoodBud that India had a severe shortage of restaurant supply, which meant many delivery companies had to build their own central kitchens. Swiggy and Zomato had both explored similar approaches.
FoodBud compared this market context with OYO: OYO had performed well in India and had already planned to list, while its China business had largely exited the market. In FoodBud’s view, Rebel Foods’ success in India similarly reflected local market conditions, especially insufficient restaurant supply.
The article also noted that some companies overestimated how much demand would shift. Swiggy, backed by South African internet group Naspers and one of the first Indian delivery companies to make a large bet on cloud kitchens, had to close and relocate many unprofitable cloud kitchens in April.
In China, FoodBud saw some comparable elements in companies such as Shiheng and Weijie. Weijie’s PR-stated figures were 2,000 total stores and annual revenue above RMB 1 billion. But FoodBud argued that such companies were only rule-followers in front of Meituan, while Meituan was the rule-maker.
In dense restaurant areas, especially malls, Meituan had set up shelf-like consolidation points. Staff could move meals from restaurants to the pickup point, allowing riders to avoid running through the mall and instead collect orders directly from the hub.
On the consumer side, pickup lockers were being deployed around office buildings and office parks. Users could collect meals quickly, and riders did not need to wait downstairs.
FoodBud framed these lockers as nodes in a hub-and-spoke model: merchant-side hubs on one end, consumer-side hubs on the other, and riders handling transportation. The core capability in optimizing that transportation was Meituan’s algorithm. Hub-and-spoke operating models are widely used in aviation and logistics.
Meituan had recently disclosed parts of its delivery allocation rules.
According to Meituan, rider delivery time was not calculated simply as nearest distance divided by fastest speed. The system included a dynamic time estimate because actual delivery has many uncertainties.
When Meituan receives a new order, its order-allocation algorithm estimates the delivery time a rider would need if assigned that order, based on the rider’s current location and existing order load, while leaving buffer time. It also evaluates whether the new order would cause existing orders to be late. After analyzing riders within the delivery area, the algorithm assigns the order to a rider with sufficient time.
For route fit, Meituan said a rider’s path changes after accepting a new order. When a new order appears, Meituan calculates the extra delivery distance for multiple riders and assigns the order to the rider with the smallest additional distance.
Meituan said it used four delivery-time assessment methods: historical data model estimates, estimates based on city traffic conditions, cumulative estimates across scenarios such as food preparation and store pickup, and delivery-distance estimates. The actual estimated arrival time uses the longest of the four. As a result, even orders on the same route can have different delivery windows on different days.
Meituan also described two ongoing adjustments: a pilot for dispatch after food is ready, and an active reassignment function.
The dispatch-after-preparation pilot targeted the problem of riders waiting for food when merchants are slow to prepare orders. In pilot areas, Meituan gave merchants a free smart terminal hardware product called Chucanbao. Restaurants could report food-preparation status, and the system would dispatch riders only after the order was ready.
According to Meituan’s announcement, 2,400 stores nationwide were already participating in the pilot. In Nanchang, the first 32 pilot stores showed a 51% decline in average rider waiting time, and about 72% of merchants said the overall experience improved.
FoodBud’s conclusion was that merchant-side smart dispatch hardware and consumer-side pickup lockers would become increasingly important in this operating model.
FoodBud compared Meituan’s smart meal lockers with parcel lockers used in residential communities.
It argued that Meituan was probably not very interested in shared-kitchen businesses, which resemble a second-landlord model. By contrast, installing food storage lockers at both merchant and consumer ends could create infrastructure that improves operating efficiency and creates new monetization routes.
According to Meituan franchise materials cited in the article, Meituan divided Chinese cities into three tiers: seven cities including Beijing, Shanghai, and Shenzhen were S-level; cities including Nanjing, Xi’an, and Wuhan were A-level; and the rest were B-level. Meituan would provide different subsidy levels to franchisees depending on city tier and effective locker order volume.
The user experience was similar to Hive Box parcel lockers. After delivery, riders place meals into locker compartments. Users receive a notification and scan a code to collect the order.
The compartments used carbon-fiber heating wires and could reach 55 degrees Celsius. They also had real-time ultraviolet disinfection. FoodBud said lockers could reduce lost or mistaken pickups and improve food safety. Early feedback suggested relatively high rider and consumer acceptance in schools, office buildings, and similar locations.
Meituan described smart pickup lockers as a new delivery handoff model that strengthens contactless delivery safety, reduces property-management burden, and improves rider efficiency.
A Shenwan Hongyuan research report cited in the article said meal lockers could sharply reduce rider waiting time and improve last-mile handoff efficiency. As last-mile handoff improves, rider delivery efficiency during peak periods was expected to rise by 30% to 50%.
FoodBud compared the likely business evolution with Hive Box parcel lockers. As locker networks mature, free usage can shift toward paid usage. Hive Box’s revenue items included courier fees, overdue parcel fees, storage fees, membership fees, and advertising.
FoodBud’s core argument: if the trend continued, Meituan would have built another “highway” inside its delivery business, with another route for extracting fees from the foodservice delivery chain.
Note: IPO plans, financing figures, valuation figures, subsidy references, and efficiency forecasts are historical statements from the 2021 source article.