TL;DR
- In 18 months, Disney tried three very different approaches to AI. It sued, it licensed its characters to an outside platform, and now it is training custom models inside its own walls with Adobe.
- The lesson isn’t about Disney. It’s about what every business has to protect: its recipe, the thing nobody else can copy.
- Most businesses don’t know their own recipe. It sits in their archive: clients, questions, decisions, what worked and what didn’t.
- Consistency turns an archive into a legacy. Changing direction every year leaves you with the same amount of data, but no formula in it.
- Below: the timeline, a kitchen analogy, how to read your own archive, and six questions to find your recipe.
Disney, one of the largest entertainment companies in the world, agreed to a $1 billion deal with OpenAI. A few months later, the deal was gone, because OpenAI shut down Sora. Instead of looking for the next big AI partner, Disney made a decision that, in my view, shows very clearly how we should make decisions about our own businesses.
This article isn’t really about Disney. It’s about a question that goes beyond it: how do we build something of our own that nobody can take from us in an AI world?
The short answer
Three acts in 18 months
- Act one: sue, June 2025Disney and Universal sued Midjourney, in one of the first major lawsuits by Hollywood studios against a generative AI company. According to the complaint, a single sentence was enough to generate Darth Vader or Elsa. In September, Disney also sent a cease-and-desist letter to Character.AI.
- Act two: license out, December 2025Disney agreed to invest $1 billion in OpenAI and to license more than 200 characters for Sora. On March 24, 2026, OpenAI announced it was shutting down Sora, and the deal ended with it.
- Act three: train inside, autumn 2026Disney named the former CEO of Character.AI as its first chief technology officer. A few weeks later, Character.AI filed for bankruptcy, and Disney reportedly bought select assets and hired most of its staff. In November, Disney named Adobe its Official AI Innovation Partner, to build custom models trained on Disney's characters and worlds.
So the sequence was: sue, license out, then train inside. Partners and approaches changed the whole time. And here’s the question that interests me most: why didn’t a company with Disney’s resources build its own model from scratch, or simply look for another deal like the one with OpenAI?
Two ways to work with a great chef
Imagine a chef who has tasted the dishes of almost every restaurant in the world. That’s roughly what large AI models are: trained on a huge part of the public internet, including restaurants that never invited them.
You bring your recipe to his restaurant. He cooks it beautifully, but in his kitchen, for his guests, and only as long as he decides to keep the restaurant open. That's roughly what happened with OpenAI and Sora.
You invite the chef in. He cooks with your ingredients, by your recipe, for your guests, and you do everything possible so the recipe never leaves the kitchen. My reading is that this is what Disney is doing with Adobe.
You hire the skills and the brain of AI, but you find a way to keep the recipe inside. If I had to sum up this article in one sentence, it would be: learn to protect your recipe. But to protect it, you first have to know what it is.
Where Disney’s recipe was born
Disney’s recipe has a precise birth date. In 1927, Walt Disney created Oswald the Lucky Rabbit, and the character became a hit. A year later, Walt went to negotiate a new contract and discovered that the rights to Oswald belonged to Universal, not to him. He lost the character and part of his team.
His answer was Mickey Mouse. And with it came a decision: never again build on something that isn’t ours. Almost a hundred years of decisions in the same direction followed: films, series, merchandise, television, parks, streaming. Each new format feeds the others and brings people back to the same characters.
That’s why I don’t worry much when people say “someone stole my idea”. A good idea isn’t stolen when its value lies in the execution. And execution is a series of well-considered decisions that nobody else has made in the same order.
Your recipe is in your archive
Most of us aren’t Disney and don’t have a hundred years of history. What we do have is an archive. Every business has one. Smaller businesses have a smaller archive, larger ones a larger one.
| What’s in your archive | What it can tell you |
|---|---|
| The clients who came to you | Who really values what you do |
| The questions they ask | What problem you actually solve |
| What you tried and what worked | Where your strength is |
| What didn’t work | Where not to spend energy again |
| The decisions you made | The logic you follow, often without noticing |
Many people sit on a gold mine without realizing it. I work with companies, some of them very large, that have years of data and experience and still don’t know the formula of their own success. They don’t know why things worked out. And instead of finding out, they start running after every new technology, accepting every offer and scattering in every direction.
The same applies to many independent professionals who jump from project to project for years. You can be excellent at your work and still be completely unpredictable. And unpredictable people are rarely recommended, because nobody knows what exactly they would be recommending.
Searching, adapting or scattering
Changing direction isn’t a problem in itself. Right now it’s often how you adapt and survive. But there is a difference.
| Searching and adapting | Scattering | |
|---|---|---|
| Each new decision | Builds on the previous one | Starts from zero |
| What you’ve learned | Gets reused | Gets forgotten |
| Audience, site, team | Grow over time | Start over each time |
| Trust you’ve earned | Accumulates | Gets reset |
| Your data | Shows a pattern you can repeat | Holds only noise |
Disney is a good example of the first column: animation, parks, streaming, toys and now AI. The format changes, but the characters, the logic and the intellectual property stay.
What loses value and what doesn’t
For a long time, many businesses earned money because they knew something the client didn’t: where to buy, how something is made, who can do it. In many sectors, that’s no longer enough. The client asks AI and gets an answer in seconds.
This doesn’t mean intermediaries have no value. Curation, convenience and trust have very real value, and people keep paying for them. But information that everyone gets for free isn’t that kind of value anymore.
A business without values often chases only money, and that works while the market works in your favor. When it stops, you realize every decision was made by a single criterion. I'm not against numbers. They help you decide, but they don't tell you which opportunities are worth chasing.
You can get advice from anyone, even from an AI model. That's exactly why people increasingly look for the one person they trust. Trust takes part of the weight of a decision off your shoulders. Nobody pays for the answer. They pay for lower risk and for being understood.
The best comparison I have is a good therapist. They don’t give you information you don’t have. They listen and know how to bring out what’s already in you. In business, I can tell a project went well when, after enough time working together, the client forgets that at the start they stood on the opposite side, and reaches the conclusions on their own, from the data they’ve been sitting on.
Find your recipe
Six questions to find your recipe
That last point matters online too. A business that’s clear about what it stands for is easier for people, and AI assistants, to recommend. If you want to see how ChatGPT, Gemini and Perplexity describe your brand today, the AI Visibility Audit shows where you stand.
My take
Disney tried to sue, tried to give its characters to someone else’s platform, and finally invited the chef into its own kitchen. It hired the skills, but, as far as it can, it never let the recipe out.
Our recipe is our values, the trust we’ve earned, our understanding of people, the history of our decisions and the consistency with which we make them. AI can help us act faster, but it can’t write that recipe for us. Even with all the information in the world, and we already have it.
What do you think: do you know your recipe? I’d like to hear your view.
For a neutral overview of Disney’s AI decisions and what they mean for intellectual property, read Disney’s AI Strategy, From Midjourney to Adobe, Explained on Egorithms.


