The restaurant business is changing beyond recognition
Originally written in Chinese. Translated with AI, reviewed by Gary. Read the Chinese original →
I recently read an article in The Economist titled "The restaurant business is changing beyond recognition", published on 30 July 2026. Its gist:
The restaurant industry is being transformed. Technology, inflation and consumer habits are reshaping it together.
I found it quite eye-opening, because some of its points are angles I rarely used to consider when looking at the restaurant business.
When we used to think about opening a restaurant, it was basically: find a good location, fit it out, hire people, cook in the kitchen, and wait for customers to walk in. But today the restaurant business is slowly becoming a business of data + supply chain + automation + delivery + experience.
The article is long, so I won't cover all of it. I'll just pick a few points I found especially interesting.
Can your food still sell once it leaves the restaurant?
First, the consumer is no longer the consumer of the past. After the pandemic, work from home changed foot traffic across whole cities. We used to think the CBD, financial district and office areas were always the best places for a restaurant, but now many suburban areas are busier, because people don't necessarily go back to the office every day.
OpenTable data shows that the dining peak in London and New York, which used to be Friday, is gradually moving to Thursday, and people are eating dinner earlier and earlier.
Even more striking: in 2025, of every four meals sold by US restaurants, about three were not eaten in the restaurant. The global online food delivery market is expected to have reached about US$473 billion, roughly double pre-pandemic levels, while physical restaurant visits in the UK are still about 15% below pre-pandemic.
So running a restaurant today, you can't just ask:
"Is my location good?"
You also have to ask:
"Once my food leaves this location, can it still sell?"
Location used to be perhaps the most important asset on its own, but today that isn't enough. Whether your food suits delivery, whether it still tastes good 30 minutes after leaving the restaurant, whether you can reach customers through different channels: all of these are becoming more and more important.
Second, supply keeps rising, but consumers are getting thriftier.
The restaurant business never had high barriers to entry, and during the pandemic many people switched careers and opened restaurants, so the number of US restaurant outlets actually hit a record. But the article cites National Restaurant Association data showing that about 42% of US restaurants were unprofitable in 2025.
China's situation is even more pronounced. Consumers are increasingly value-conscious. Average spending per customer at large restaurant chains fell by about 20% at one point in the first half of 2024, and dropped another 8.3% in 2025.
In short:
More and more restaurants, customers not growing with them, and each one spending less.
It's a classic case of rising supply under demand pressure. Competition gets fiercer, and eventually everyone turns to promotions, discounts, vouchers and price wars. But restaurant margins were never especially thick, so when rent, labour, ingredients and platform commissions all rise together, it easily ends up with sales still there and profit gone.
Restaurants are becoming factories
The third point, I think, is the most interesting in the whole piece: restaurants are becoming more and more like manufacturing businesses.
The article uses Japan's Saizeriya as an example. For decades, the chain has rarely raised prices on many of its dishes. Not because its margins are especially high, but because it has relentlessly restructured its supply chain.
It builds its own factories and processes its own ingredients, even producing béchamel sauce in Australia. Many ingredients arrive at the restaurant already prepared and cut, so outlet kitchens don't need as many skilled chefs. They barely even need knives.
Out front, there's QR-code ordering, robots delivering food and automated cashiers.
In a traditional restaurant, the customer orders, a waiter takes the order, a chef cooks, and the waiter brings it to the table. Now it's slowly becoming:
centralised factory → standardised food → automated kitchen → digital ordering → customer.
In the UK, Wagamama and Wasabi are doing something similar, moving preparation out of expensive city-centre restaurants into central kitchens with lower rent. Karma Kitchen even provides the infrastructure directly: industrial kitchens, cold storage, ventilation, electricity, cleaning.
On the surface it's still a restaurant, but the operating model behind it looks more and more like a factory.
We used to see restaurants as a retail business, because location, foot traffic and shopfront mattered most. In future, they may look more and more like a supply-chain business.
Automation is also starting to genuinely change restaurant unit economics. The pandemic accelerated apps, digital ordering, self-service, delivery and automation all at once. The article notes that real labour productivity in the US restaurant industry rose by about 15% after this wave of digitalisation. Sweetgreen's automated assembly line can make about 500 customised salads an hour.
So future competition among restaurants may be less about who hires the most people, and more about who can produce more meals with less labour, less kitchen space and less waste.
The same sales with fewer people; the same kitchen with higher throughput.
All of it comes back to unit economics in the end.
From reactive to predictive
Then there's a part I really like: AI coming into restaurants. But the AI it talks about isn't robots frying rice.
The real value may lie in prediction.
The article uses examples from Toast. When Taylor Swift played concerts in Miami, omelette sales at nearby restaurants rose 27%. When Beyoncé performed in Atlanta, demand for hard seltzer and late-night meals jumped, with a spending peak even at four in the morning.
What makes this data truly valuable is that, in future, a system can know: what concerts, football matches or conferences are happening nearby tomorrow, what the weather will be, what sold best at similar past events. Then it can forecast roughly how many customers to expect tomorrow, how much inventory to prepare and how many staff to roster.
In other words, restaurants can move from:
"People arrive, then I prepare supply."
to:
"Before demand happens, I've already started preparing supply."
From a reactive business to a predictive one.
The article ends with another point: the middle of the restaurant market may find it harder and harder.
On one side is cheap + fast + convenient, using central kitchens, delivery and automation to push efficiency to the limit. On the other is premium + experience.
If a consumer today has to make a special trip, sit in traffic, find parking and spend a few hours on a meal out, that meal has to give them a reason. Good food may be just the baseline; you also need ambience, service, story and a social experience.
The most dangerous place is stuck in the middle:
Not cheap, not fast, and no particularly good experience either.
And the most interesting part is that this shift isn't only happening in restaurants.
Fashion retail is on a very similar path.
It used to be: design, produce, open stores, wait for people to come. Today it's demand data, small batches, real-time inventory, content, delivery and automation.
Restaurants are the same.
They're slowly shifting from a location-driven business to a demand-driven business.
The Economist's headline, "changing beyond recognition", isn't an exaggeration at all.
What's really being changed may not just be restaurants, but our decades-old understanding of what it means to "open a shop and do business".