Eating in the 21st century

Here’s how I think most of us in the developed world will eat starting sometime in the next decade or so:

Restaurants and grocery stores will still exist, but most lunches and dinners we consume will be delivered to us as complete, fully-cooked (or mostly-cooked) meals. Fewer people will keep fully stocked kitchens, and many more people than today will forgo owning kitchen equipment beyond the basics like fridges and microwaves.

Of course, such a lifestyle has been feasible for decades via takeout, but eating takeout every night is considerably more expensive than cooking for oneself. We also could simplify our meals by microwaving frozen dinners, but these dinners range from merely passable to terrible. So most of us have stuck to cooking.

This should strike us as strange. In a world of ever-increasing specialization, why do most of us cook when we neither really wish to cook nor do so well? The answer is simply that no one has yet sorted out the logistics of having the specialized few cook for the many in a way that leverages scale to keep costs competitive with the status quo. The requisite information technology has been widely distributed for several years, but no one has yet put it together.

So here, in broad outline, is what I believe predominant consumer food services will look like quite soon:


  • First what we’ll need are large-scale kitchens geared to produce hundreds or thousands of meals at a time. What we won’t need are storefronts, so the kitchens will generally be located in light industrial zones, where real estate is cheaper. We’re aiming for a scale of production considerably larger than any restaurant but smaller than a factory. At some point in the future, surely many cooking tasks in these kitchens will be handled by robots, but in the near term, food preparation will be handled by chefs and their staffs.
  • Rather than having a fixed menu, the kitchen’s team of chefs devise the menu for each week anew. Several meal options are offered each day, but the options constantly change.
  • Unlike a conventional restaurant, where meals are mostly made only once ordered, these meals are prepared en masse and coordinated to be ready at designated times, e.g. ‘we need 100 of meal X at 6pm, 200 of meal X at 7pm, and 50 of meal X at 8pm’.


  • Customers browse the week’s menu and order via mobile apps or the web.
  • A sizable discount is offered for placing orders 12+ hours in advance because then the chefs know exactly how much to make. A discount is also offered for repeat business, let’s say 10% off any order placed within 5 days of your last order.
  • Different meals have different prices, but most meals are within the range of a middle-class American’s daily food budget ($7–10).
  • Delivery is free because the meals are delivered en masse rather than ad hoc (explained more below).
  • Customers can choose between portion options that vary the meal prices within a dollar or so.
  • A discount is also offered for ordering multiple instances of the same meal, e.g. a family might order three of a particular meal for delivery at the same time.


  • Rather than deliver meals to every customer’s door, the meals are loaded on trucks and taken to locations where the customers meet the trucks. (Getting permits for parking these trucks to distribute the meals is a notable regulatory hurdle.) The delivery locations are chosen such that most customers need travel a short ways to pick up their meal.
  • The general goal is to keep times between meal readiness and delivery under 30 or 40 minutes. This is of course much longer than at a typical restaurant, but unlike takeout food, the meals are generally designed to account for delivery time, e.g. a meal may be slightly under-cooked in the expectation that the customer is instructed to further heat the meal for some (short) length of time.
  • When ordering, customers select their pickup location and time (generally a 20 minute window). Customers receive updates on the status of their orders in case it does not precisely make its window.
  • The meals are delivered in reusable, sealed, hard-plastic trays. Each tray is tracked by an embedded id chip. The customer puts a deposit on each tray, and the deposit is refunded when the tray is returned. (Keep in mind we’re aiming for repeat business: returning trays is not really a hassle if customers pick up new meals every night.)
  • Customers are asked to rinse their trays before returning them so as to prevent food crusting on the material. Failure to do so may incur fees for damaged trays. Regardless, the trays are fully washed upon return.

Of the Silicon Valley startups currently doing “Uber for food”, the closest to the mark is Munchery. However, Munchery deviates from my prescribed formula in four ways:

  • Munchery produces and delivers meals upon demand rather than producing meals truly en masse.
  • Munchery deliveries go straight to the customers’ doors, and so customers pay an enormous $4.50 delivery fee for every order.
  • Munchery uses disposable packaging. Though the packaging is minimal (relative to ordinary takeout) and bio-degradable, it’s still offensively wasteful to generate all that refuse.
  • Worst of all, even without delivery fees, Munchery food is just too expensive. A single middle-class American generally spends less than $300 a month on food, but the average Munchery meal (with delivery fee) is over $15, meaning a customer would spend over $450 a month just for their dinners.

For mass adoption of food services as a replacement of cooking at home, food services will have to deliver each person’s lunch and dinner for less than $10 a day. With sufficient scaling that fully capitalizes upon today’s technology, that should be a viable price point for a successful business. I’ll be interested to see if any of the current players catch on.

Like what you read? Give Brian Will a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.