![]() |
| AI-generated illustration of a hand-drawn theme-park sketch beside a monitor showing the same sketch rendered as colourful concept art. |
The Walt Disney Company and Adobe announced on 9 November 2026 that they have signed a multi-year strategic alliance to build custom generative AI models trained on Disney's own characters, worlds and visual styles. The deal, announced at Adobe MAX in Miami Beach, makes Adobe Disney's Official AI Innovation Partner and puts Adobe's Firefly Foundry at the centre of Disney's creative pipeline, from theme-park design to its film studio and ESPN. It is a striking turn for a company that, just last year, sued an AI image generator and called it a "bottomless pit of plagiarism". But before anyone declares that Disney films are about to be made by machines, it is worth being precise about what this deal does and does not do: the models are private to Disney, no finished product has been announced, and the companies have disclosed no financial terms and no timetable for rolling the tools out across the business.
The core of the agreement is Firefly Foundry, Adobe's service for building proprietary AI models for a single enterprise customer. Disney will work directly with Adobe on generative AI models tailored to its properties, trained on what Adobe calls "unique world knowledge" — character attributes, environments and visual styles — to generate what the companies describe as commercially safe and production-ready outputs. In practice, the first uses are unglamorous and internal: speeding up theme-park project development, exploring concepts at the Walt Disney Studios innovation lab StudioLAB, and finding new ways to use ESPN's archive. Disney will also get early access to Adobe Firefly, Premiere and Photoshop, along with dedicated Adobe engineering and research support. What the deal does not announce is any new tool for the public: you will not be able to generate Disney characters yourself, and no Disney film or show made with these models has been named.
As an Amazon Associate, we earn from qualifying purchases.
Prices and savings can change. Check the retailer for the current price.
What the deal actually covers
The agreement touches four parts of Disney's business, and each one is at a different stage. The most advanced is Disney Experiences, the division that runs the theme parks, resorts and cruise lines. Walt Disney Imagineering, the creative and research arm behind the parks, is already using two custom models built through Firefly Foundry. One is a sketch-to-image model that turns a hand-drawn concept into a fully rendered piece of two-dimensional concept art. The other is a custom image model that generates creative assets designed to stay on-brand and accurate to a specific franchise. Disney's senior vice president of research and development technology and engineering, Kyle Laughlin, told Reuters that the tools compress work that used to take months into hours and days, and let teams run through many more iterations before anything is built, which reduces costly changes once construction starts.
The second area is the film studio. Adobe will join StudioLAB, the prototyping hub inside Walt Disney Studios, as a StudioLAB Innovation Partner, and the two companies will work together on concept exploration and on broadening internal creative workflows. Again, the emphasis in the announcement is on early-stage work — exploring ideas faster — rather than on finished films. No production has been named as using the models, and the announcement does not say the tools will generate finished frames or footage.
The third area is sports. Adobe will work with the ESPN Edge Innovation Center on using AI to boost engagement with sports fans and to bring more of ESPN's content archive to audiences. That is the vaguest part of the deal, and the most open-ended: an archive as large as ESPN's is an obvious place to apply AI search and summarisation, but neither company has described a specific product or feature.
The fourth part is access rather than a project. As Official AI Innovation Partner, Adobe will give Disney early access to its latest creative and media technologies — Firefly, Firefly Foundry, Premiere and Photoshop — plus dedicated engineering and research resources to support the Disney production pipeline. In other words, Disney gets to shape and test Adobe's tools before other customers do, and Adobe gets its technology embedded in one of the world's most recognisable creative companies.
Why Disney changed its mind about AI
The deal is easier to understand against Disney's recent history with AI, which has swung from legal confrontation to deal-making in the space of eighteen months. In June 2025, Disney and Universal sued the AI image company Midjourney for copyright infringement, and Disney's complaint described Midjourney as a "bottomless pit of plagiarism". That lawsuit is still part of the landscape: Disney has not withdrawn its claim that AI companies should not train on its characters without permission. What has changed is the method. Instead of fighting other companies' models in court while its own artists work without AI tools, Disney is now building models that train only on material it already owns, inside a closed system it controls.
That shift was already visible before this week. In June 2026, Adobe and Walt Disney Imagineering announced a narrower collaboration to bring Firefly Foundry into the design and pre-production visualisation pipeline for the parks. The two models Imagineering is already using came out of that work, and this week's alliance widens it from a single division to the whole company. And in December 2025, Disney agreed a reported 1 billion US dollar investment in OpenAI, alongside a deal to license its characters for OpenAI's Sora video generator. Taken together, the pattern is consistent: Disney is not opposed to generative AI as such. It is opposed to other companies training on Disney characters without a deal — and it is now signing deals, on its own terms, with the companies it chooses.
There is also a commercial logic that has little to do with art. Bloomberg reported earlier this month that Disney is considering cutting thousands of jobs and has asked some managers to target 20 percent in annual cost savings, a report Reuters noted alongside the AI deal without linking the two. The two stories should not be conflated — the companies have not said the AI tools are meant to replace staff, and Laughlin told Reuters the tools are "in no way meant to replace" roles at Imagineering. But the context matters: Disney is under pressure to cut costs at the same time as it is investing in AI tools that make parts of the creative process faster. Readers should keep both facts in mind, and judge the deal by what happens to Disney's creative workforce over the next year, not only by the press release.
What "artist-led" and "commercially safe" really mean
Adobe's announcement leans heavily on two phrases: the models are "human-centered and artist-led", and their outputs are "commercially safe and production-ready". Both deserve a careful reading, because they are the companies' own claims, not independently established results. "Artist-led" means the models are built around Disney artists' work and are meant to be directed by them, with humans choosing what to generate and what to keep. "Commercially safe" means the models are trained on Disney's own intellectual property rather than on the open internet, so the outputs should not reproduce another company's characters or copyrighted work. Those are real design choices, and they address the two objections Disney itself has raised against other AI tools: that they train on other people's work without permission, and that they sideline the artists.
But the phrases also have limits that the announcement does not spell out. "Commercially safe" is Adobe's assurance about its own training data, and it has not been tested in court. "Production-ready" describes the companies' ambition for the outputs, not a finding that every generated image is ready to use without review — the TechBooky analysis of the announcement makes exactly this distinction. And "artist-led" does not answer the question artists themselves keep asking about generative AI: whether faster concept work means fewer concept artists, or simply more concepts from the same artists. Laughlin's statement that the tools are not meant to replace roles at Imagineering is a genuine commitment from a senior Disney executive, and it is worth recording. It is also a statement about intent at one division, at the moment of announcement. The honest test is what Disney's artist headcount and freelance hiring look like a year from now.
What this means for creators and fans
For most people reading this, the immediate effect is close to nothing, and that is worth saying plainly. The models are private and proprietary to Disney. They will not be released, they will not appear in Adobe's consumer products, and they give the public no new permission to generate Disney characters. If you are a fan hoping to make your own images of a favourite character with an official tool, this deal does not give you one. If you are an artist worried that your work has been scraped into a public model, this deal does not address that either — it is about Disney's own library, not about the wider training-data debate.
For working artists, the implications are more mixed, and they depend on details the announcement does not provide. On the positive side, a closed model trained only on Disney IP is the version of generative AI that creators' groups have asked for: opt-in, licensed, and controlled by the rights holder, with the artists who made the original work still in the loop. If Disney's model becomes a template, it could push the industry towards licensed training data rather than scraping. On the cautious side, the history of creative-industry automation is that tools described as assistants gradually take over more of the pipeline, and the people who lose work are usually freelancers and junior staff, not the senior artists quoted in press releases. Disney has not said how it will credit AI-assisted work, whether artists whose work trains the models will be paid for that use, or whether the tools will be offered to outside studios working on Disney projects. Those are the questions to keep asking.
For Adobe, the deal is a proof point for Firefly Foundry, its enterprise service for building custom models. Adobe MAX, where the deal was announced, is the company's biggest showcase, and having Disney as its flagship enterprise AI customer is a stronger sales story than any demo. It also shows where Adobe sees its AI future: not only in tools for individual creators, but in private, licensed models for large companies that want AI without the copyright risk of public models. Expect other studios and brands to be offered the same pitch in the months ahead.
How to judge this deal over the next year
Announcements are cheap; the useful thing is to know what evidence would show whether this one mattered. Here is a short checklist to revisit over the coming year.
- Watch for the first production credit: the first Disney film, show or park attraction that credits AI-assisted concept work will tell you how far the tools have moved from experiments into real production.
- Watch whether the models stay private: if Disney ever offers a consumer-facing tool built on them, that would be a major change of policy, and it has not been announced.
- Watch Disney's artist employment: headcount, freelance hiring and any union statements will say more about "artist-led" than the press release does.
- Watch ESPN: the first visible AI feature on ESPN, whether an archive search tool or a new format, will show whether the sports side of the deal is real or decorative.
- Watch the Midjourney lawsuit: Disney is still pursuing its claim against an AI company it accuses of training without permission, and the outcome of that case will shape the rules for everyone else.
- Watch for the next deal: if another major studio signs a similar alliance with Adobe or a rival, this deal was the start of an industry pattern rather than a one-off.
Bottom line
Disney and Adobe's alliance is a significant deal, and it is also a narrow one. It is significant because the world's most protective entertainment company has decided that the answer to generative AI is not only to sue, but to build — private models, trained on its own characters and worlds, controlled by its own artists. It is narrow because nothing in the announcement gives the public a new tool, names a finished product, or puts a price or a date on the work. The right response is neither excitement nor cynicism: take the deal seriously as a signal about where the creative industries are heading, keep the companies' claims in proportion, and judge it by the evidence that appears over the next year — the first credited production, the state of Disney's artist workforce, and whether the "artist-led" promise survives contact with the company's cost-cutting plans.
Reference sources
- Reuters, "Disney, Adobe to co-develop custom AI models tailored for media giant", 9 November 2026
- Business Wire, "Adobe and The Walt Disney Company Announce Multi-Year Strategic Alliance to Support Disney Innovation with New AI-Powered Tools", 9 November 2026
- TechBooky, "Disney and Adobe Expand Firefly AI Partnership", 9 November 2026
- Mediaweek, "Adobe and Disney sign multi-year AI alliance", 10 November 2026

Leave a public comment