Wonder is building a system in which a restaurant can be imagined by a creator, developed with AI and produced through someone else’s kitchens. The interesting question is not whether this makes restaurants more efficient. It is what happens when the restaurant itself becomes programmable.
The room, unphotographed — the thing that gets more valuable as restaurants become software.
Wonder has raised $650 million at a $9 billion pre-money valuation. It is a significant amount of money for a restaurant business, although increasingly that description seems inadequate.
The company develops and operates restaurant brands, owns Grubhub and Blue Apron, is investing in robotics and autonomous delivery, and has expanded from 46 locations in May 2025 to around 140 by July 2026. Its latest funding will accelerate that physical expansion while supporting further investment in artificial intelligence and the infrastructure beneath it.
All of which is interesting. But the more consequential idea may be Wonder Create.
The proposition is deceptively simple. Bring an idea for a restaurant, along with your own recipes if you have them, or use AI to help develop the identity and menu. Wonder can then match those dishes against its ingredient library, produce them through its kitchens and distribute them across Wonder, Grubhub and Seamless.
The creator retains the concept and intellectual property and controls the menu, pricing and margins. Wonder supplies the machinery underneath it.
That removes much of what has traditionally stood between an idea for a restaurant and an actual restaurant: the lease, the kitchen build, the equipment, the brigade, the procurement system and the considerable operational headache of getting the doors open.
In Wonder’s language, what you need is an idea.
That changes something more important than restaurant efficiency.
The scarce thing is changing
For most of restaurant history, scarcity has been physical.
A suitable site was scarce. Capital was scarce. Skilled chefs and staff were scarce. So were the knowledge and operational capability required to produce hundreds of meals consistently while somehow maintaining a viable margin.
These things were not simply infrastructure around the restaurant. They were part of what defined one.
Digital platforms have spent the past two decades separating industries from many of their old physical constraints. Music escaped the CD, television escaped the broadcasting schedule and retail separated the transaction from the shop.
Food has resisted the process because, however sophisticated the interface becomes, somebody still has to cook dinner.
Wonder does not change that. What it potentially changes is who needs to own the means of producing it.
The kitchen remains physical, but the person creating the restaurant no longer necessarily needs to build, own or operate one. Production becomes infrastructure: something that can be accessed rather than assembled from scratch.
Once that happens, scarcity begins to move elsewhere.
Away from property and equipment. Away, to some extent, from operational capability. And towards attention, audience, ideas and taste.
That is the point at which Wonder becomes something more than a story about restaurant technology.
It becomes a story about the economics of appetite.
When the creator owns the appetite
The creator economy has been circling food for years.
Creators determine where people eat, make dishes viral, sell cookbooks, empty supermarket shelves and introduce cuisines to audiences that traditional food media might once have taken years to reach.
The commercial response has been predictable: endorsements, collaborations, limited editions and branded products. Put a creator’s name on a sauce, invite them to create a burger or turn their following into a range of packaged goods.
Wonder Create proposes something structurally different. Instead of asking a creator to promote somebody else’s restaurant, it gives them the possibility of creating one.
A creator who has spent years developing an audience around regional Italian cooking, bodybuilding food, Korean street food or enormous sandwiches may already possess something a new restaurant would spend heavily trying to acquire: a community with a recognisable, shared appetite.
Until now, translating that attention into a restaurant meant entering an entirely different business. Property, labour, procurement, food safety, production, delivery, margin and waste all sat between influence and execution.
Wonder wants to compress much of that distance.
The creator brings the idea and, crucially, the audience. The platform brings the kitchen, technology and distribution. The restaurant begins to emerge somewhere between the two.
That is a subtle but important shift. The creator economy is no longer simply influencing what we eat. It is beginning to acquire the infrastructure required to make what we eat.
And that starts to change the restaurant itself.
A traditional restaurant is difficult to reproduce because it is an accumulation of human and physical variables. The room matters. The chef matters. The equipment, ingredients and team matter. Tuesday night is not quite the same as Saturday night, and the same dish cooked six months apart will never be perfectly identical.
Wonder is attempting to translate far more of that complexity into a repeatable system.
An idea becomes a brand. The brand becomes a menu. The menu becomes a series of instructions that can be executed through an existing production network.
At that point the restaurant starts to share characteristics with software.
It becomes programmable.
And when something becomes programmable, we generally make far more of it.
The infinite restaurant shelf
Media has already shown us what happens next.
Spotify increased access to music to such an extent that the problem stopped being finding something to listen to and became deciding what deserved our attention. Netflix did something similar to television. YouTube expanded video almost infinitely, while TikTok built an entire cultural system around an endless competition for a finite amount of human attention.
Food may be approaching its own version of the same problem.
Imagine restaurant brands created around individual creators, internet communities, diets, fandoms or passing cultural moments. A restaurant could appear around a viral dish or television series, reach its audience while the moment remained culturally relevant and disappear when attention moved somewhere else.
The restaurant would no longer necessarily aspire to become an institution.
Some could behave more like media releases: seasonal, responsive and deliberately temporary.
The constraint would cease to be how many restaurants we can physically create. It would become how anyone chooses between them.
Food acquires its own infinite shelf.
The next turn is when the feed leaves the screen and an agent chooses which brand on that shelf you ever see.
That should make anybody who has watched the development of digital media slightly cautious. Greater access to content did not automatically produce better culture, and reducing the barriers to launching restaurant brands will not automatically produce better food.
Removing the friction of creation does not remove the requirement for judgement.
If anything, it makes taste more important.
Which is why the most interesting question about automation in kitchens is probably not whether AI and robotics will replace chefs.
The better question is where human value moves when repeatable production becomes easier.
Consistency matters enormously in restaurants, but if technology becomes increasingly capable of supplying it, consistency alone becomes less distinctive. The chef who can operate a system remains valuable; the chef capable of creating something the system would not otherwise have arrived at becomes more valuable still.
Judgement, memory, curiosity, cultural understanding and the ability to taste something and know it is wrong before the spreadsheet can explain why are rather harder things to industrialise.
AI may therefore expose a distinction the restaurant industry has sometimes preferred to blur: the difference between cooking and authorship.
What remains difficult to reproduce
There is an irony underneath all of this.
The easier restaurants become to generate digitally, the more valuable the genuinely physical restaurant may become.
Not every inefficiency in hospitality is a problem waiting for technology to remove. Some of the things that make restaurants meaningful are inefficient almost by definition.
The waiter who remembers what you drank last time. The chef changing a dish because the tomatoes happened to be particularly good that morning. The noise of the table beside you. The smell coming from the kitchen. The room itself, full of people you do not know, collectively turning dinner into something more than a transaction.
That is also why some operators are trying to deliver the room when the diner never walks through the door.
None of these things scales especially well.
That may eventually be their value.
We have seen the pattern elsewhere. Recorded music became ubiquitous and live performance became more important. Online retail made purchasing almost frictionless while the best physical shops increasingly justified themselves through experience, theatre and human service. Digital abundance repeatedly increases the value of things that feel scarce, specific and difficult to reproduce.
Restaurants may be heading towards a similar divide.
At one end will sit extraordinarily efficient systems capable of generating and producing an almost limitless number of food brands. At the other will be restaurants whose value lies precisely in their inability to be reduced to a system: places tied to a chef, a room, a neighbourhood, a moment.
Between them will be a new category we are only beginning to understand.
This is what makes Wonder worth watching. It is not simply automating the restaurant. It is beginning to separate the idea of a restaurant from the infrastructure required to operate one.
That could allow creators, chefs, entrepreneurs and eventually perhaps anyone with sufficient imagination and an audience to create restaurants in something approaching the way we already create other forms of media.
Which brings us back to the question.
If a restaurant can be prompted into existence, constructed against an ingredient library, produced through a distributed kitchen network and delivered to customers without its creator ever operating a restaurant themselves, what are we still calling a restaurant?
Because once restaurants can be generated like content, the interesting question is no longer how many we can create.
It is what will still make one worth eating.
Eat Your Screen — The New Culture of Appetite.
We used to eat from the earth. Now we eat from the feed.