Sora Is Coming to ChatGPT. A Big Test for Disney's IP Now Looms.

[Author’s Note: A reminder I am away for some business travel and some time off for the second half of this week. Barring any major developments, the next essay will be on Monday.

This essay is free for all subscribers]

This morning The Information reported that “OpenAI plans to soon launch its Sora AI-generated video capabilities in ChatGPT.”

ChatGPT has over 920 million weekly active users, whereas Sora has 20.7M lifetime downloads and flat growth—only 2M additional downloads in roughly two months— since its launch last October, according to recent data from research firm AppMagic. That said, it reportedly will continue to operate the Sora standalone app, though its long-term plans for the app are “not clear”.

Now one-click away from a global audience of ChatGPT users, Sora will inevitably drive greater usage than it has seen, to date. Google’s data for usage of its video generation platform Veo—230 million videos have been generated with Veo 3, up 600% since June 2025—suggests the business rationale: OpenAI needs to expand distribution for Sora both because the app needs it and ChatGPT needs to compete with Gemini.

Will it scale? According to the AppMagic Mobile Market Landscape 2026 report I analyzed last week, downloads of chatbots like ChatGPT are around 4x as large as AI art generators and AI music generators. Both this data and Sora’s download history (below) suggest consumer demand for video generation may underwhelm relative to ChatGPT’s scale.

This confirms the IP bifurcation thesis I outlined in November. IP behind hyperscaler paywalls (e.g., ChatGPT, Google Gemini) retains value because licensing and generation can be controlled. IP distributed through cable syndication, free streaming platforms, or social media clips loses value because it becomes training data for open-source models.

Sora as an independent app offered a narrow pathway into OpenAI’s ecosystem. Its integration into ChatGPT pulls it deeper within that ecosystem and, like Veo 3 in Gemini, makes it core to the platform’s value proposition.

Good News for Disney?

On its surface, this appears to be a good development for Disney’s $1 billion deal with OpenAI which licensed 200-plus characters to OpenAI under a three-year agreement. The licensing terms apply to the Sora platform specifically. Embedding Sora in ChatGPT exponentially expands the existing audience and potential audience. Disney appears positioned to have more user-generated content to choose from for Disney+. OpenAI will have more content it will be able to monetize as it expands its advertising model.

That said, Sora’s integration also raises the question whether the constrained, Disney-approved use cases for its IP will generate enough engagement on ChatGPT to justify the partnership. Hilary Mason—Co-founder and CEO of generative AI roleplaying platform Hidden Door—predicts “Disney characters on Sora will be either so locked down as to be boring or will feature in humiliating, offensive, and pornographic content within hours, or maybe both.” The odds of those types of content emerging seemed smaller when Sora was smaller. Now, with 920 million users getting access to Sora, they seem much higher.

The other challenge is whether Sora within ChatGPT will compete with popular generative AI platforms Seedance and Veo 3. Beyond Disney’s IP, Sora has not demonstrated a compelling advantage over Seedance or Veo 3 for the creators—professional and amateur—who are actually making generative AI content. As I argued before, they may prefer celebrity likenesses to Disney IP.

That leaves both Disney and OpenAI facing some new questions about their partnership. If demand for the approved use cases is thin, Disney’s $1 billion will have bought positioning on a platform where its IP will be underutilized. If demand is strong, Disney will prove that licensing behind a hyperscaler paywall works—but only within that wall. Outside, content like “Rasta Wars” and the Stormtrooper vlogs suggest that creators are happy with Veo 3 and open source models.

Given these videos and the explosion of Seedance-produced videos, OpenAI and Disney will have to renegotiate or kill the Sora elements of their partnership. Exiting CEO Robert Iger made the deal. New CEO Josh D’Amaro inherits the consequences. How he handles them will be one of the closely watched decisions of his early tenure.

Market Coordination Failure

At 20 million users and low usage, Sora was a manageable threat to studios. At 920 million, that is no longer possible. Disney’s only options are to enforce selectively and hope the volume stays low, or to accept that its IP will circulate uncontrolled inside the largest consumer platform in the world. Neither involves outcomes that Disney will embrace. In the second instance, Disney could issue takedown notices while simultaneously benefiting from unauthorized use as earned media. However, if any version of Mason’s predictions proves to be accurate, D’Amaro will opt to kill the Sora elements of the partnership.

Disney is not alone in this bind. Every studio, insurer and investor described in my prior essay now faces Sora’s larger scale—and the same absence of shared frameworks to act on it.

Whatever the outcome, OpenAI’s move confirms that storytelling creates moments but infrastructure captures value. The scale of storytelling just changed within OpenAI. The question now is how the rest of the market will adapt if ChatGPT users do not flock to Sora. They may not. The chatbot still wins.