control who sells and where you buy, all in ChatGPT

A few days ago we talked about how OpenAI is turning ChatGPT into the Windows of AI. It’s not really about whether OpenAI is building an operating system. It’s about what kind of power you’re looking for. And the answer lies in two models that define exactly what OpenAI pursues: Apple’s App Store and WeChat in China.

• The App Store controls distribution. Decide what apps exist, how they are promoted, what commissions are charged.
• WeChat centralizes functionality. Within a single app you do absolutely everything: you call, you shop, you order taxis, you pay bills, you reserve restaurants.

OpenAI wants both things at the same time. And so far no one has achieved it.

The moment. Apple actively blocked superapps on iOS. The documents from the Epic v. Apple trial made it very clear: allowing one app to do everything “would threaten the monopoly” because “adherence to the system would decrease.” Apple called it “letting in the barbarians.” Meta tried it with Facebook. Musk, in his own way, tries with X. They failed because Apple didn’t let them and because they didn’t have the right platform.

The play. Ben Thompson, of Stratecheryhe sees it clearly: OpenAI follows Microsoft’s original strategy with Windows. Apple integrates hardware and software, controls everything, but leaves room for competitors because it cannot serve all markets. Microsoft dominated without having to sell hardware. He only controlled the platform. OpenAI does exactly that:

1. Control the interface.
2. Control who users reach.
3. And let others build on top.

The difference with 2023-2024 is enormous. So They launched the GPT Store. Failure. Our own GPTs were and are very useful, but those of others were isolated in a store that no one visited. Now the Spotify, Canva, Zillow, Uber or Booking apps are not in a separate store. They are integrated into the core of the experience. They appear when they are relevant. They work within the chat.

Who is this good for?: For developers, they receive immediate access to 800 million weekly users without going through Apple or Google. Without building an audience from scratch. But in exchange, they have absolute dependence on an OpenAI that decides which apps are approved, which ones stand out, or how they are monetized. If your app competes with something OpenAI wants to do, you’re out. It is the classic dilemma of platforms. Rapid growth in exchange for zero control.

For companies that sell products, the rules of the game change completely. Before you would type “hotels in Valencia” in Google and browse through ten results. Now you have to explicitly say “search on Booking” or “search on Airbnb”.

• Strong brands in people’s minds will survive.
• Those that lived off organic traffic from Google will disappear.

There are entire businesses that prospered thanks to SEO in the Google era. Those businesses were never really strong because they were optimizations for a specific technological era. The wave of snippets several of those businesses were loaded in the search results. The next wave will sweep away a few more.

The problem. Microsoft is OpenAI’s largest investor and infrastructure partner. But now they compete directly. Microsoft launched Copilot Studioits own platform for businesses to create custom agents. And that competes head-on with what OpenAI has just launched. If OpenAI had remained a simple model provider, Microsoft would have turned it into commodity.

This app platform is a declaration of independence. It is pure coopetition: partners who must compete for the dominant role.

What’s at stake. Sam Altman summed it up this week: “Most people will want to have a single AI service, and that service has to be useful throughout their lives.” Your entire life not understood as “from today until you die”, but rather as “all areas of your life.” Work and pleasure, leisure and business.

What OpenAI pursues is not to be one platform among many. It is to be the platform:

• The place where everything happens.
• Where users spend their time.
• Where companies have to be to exist.
• Apple without the hardware. WeChat without regulatory restrictions. The App Store without the bottleneck of manual approval. Everything at once.

Maybe they will get it. Maybe not. But they are executing the perfect play at the perfect time. And that, in technology, is sometimes enough. Another thing is that Excel resists.

In Xataka | The more money they lose, the more they are worth: ten AI startups have skyrocketed their valuation by $1 trillion in 12 months

Featured image | Xataka with Mockuuups Studio