Some media reported, citing information disclosed by people familiar with the matter, that ChatGPT developer OpenAI obtained Disney character authorization, but did not pay a cash licensing fee. Instead, it adopted an extremely rare method similar to a gambling agreement.According to people familiar with the matter, OpenAI and Disney management’s copyright licensing deal for the official character of Sora, the popular Vincent video application around the world, is in the form of warrants/equity subscription rights instead of cash license fees; OpenAI will give Disney the option to further purchase shares of this high-speed growth AI startup in addition to its already announced OpenAI shares of up to US$1 billion.
These terms, which have not yet been officially disclosed by both parties, are intended to establish that if Sora developed by OpenAI is a big success and brings strong revenue generation to OpenAI, the financial incentives of both parties will be consistent; Disney is equivalent to giving up huge immediate cash income from IP licensing in exchange for potentially even greater upside in equity assets in the future.
According to the latest disclosures from people familiar with the matter, the official character authorization given by Disney this time is not a cash license fee, but in the form of OpenAI's "stock warrants" (stock warrants), and allows Disney to obtain important options to further purchase OpenAI's equity in addition to the announced investment. It can be said that Disney is exchanging the value of its strong IP list for an upward risk-return curve for OpenAI's equity. The specific potential returns depend on the speed of Sora's commercialization and the path of OpenAI's future liquidity events (such as further equity transactions/listings, etc.).
You know, for a number of super investment giants such as SoftBank and Thrive Holdings, they are vying to snap up the equity of OpenAI. Therefore, it is a very cost-effective risk-return path for Disney to obtain the equity value of OpenAI through Disney's official character authorization through gambling.
In the context of "OpenAI equity is highly scarce and institutions are vying to allocate it", Disney uses IP authorization to exchange (or obtain additional) OpenAI equity or strictly speaking "warrants". In essence, it is using content assets to make an "option-style bet bound to the success or failure of Sora." Its risk-return structure may indeed be more cost-effective for Disney.
OpenAI and Disney declined to comment.
It is understood that the two parties jointly announced last week that Disney has allowed the Sora video application to use its officially licensed Disney material library containing more than 200 animated and biological characters, including Mickey Mouse, Cinderella, Ariel and Simba; at the same time, Disney agreed to invest US$1 billion in OpenAI based on its current valuation of approximately US$500 billion.
The deal is billed as the largest equity investment to date by a major film studio in a large AI model and AI application startup. For OpenAI, this cooperation will help it gain a foothold in Hollywood and greatly enhance Sora's competitiveness in the competition with Runway, Google Nano Banana and other similar AI video generation services.
According to media reports, OpenAI has been in in-depth communication with major studios and film studios including Disney, Comcast's Universal Pictures, and Warner Bros. Discovery in the past few months about the creative and commercial potential of Sora applications.
As part of the latest deal, Disney officially agreed to use OpenAI's software technology to create new products and experiences, making it immediately one of OpenAI's best-known and largest customers.
The blockbuster cooperation with Disney comes as Google and OpenAI stage a "peak AI showdown"
As ChatGPT developer OpenAI launched the "ridiculously powerful" GPT-5.2 in deep reasoning and code generation to compete with Gemini 3 launched by Google in late November, following the recent strong showdown between Sora2 and Google's Nano Banana Vincent video application to win more new users, this "peak AI showdown" between Google and OpenAI has reached the climax of the story between the two sides.
This major cooperation with Disney basically takes place at an important stage when the competition between OpenAI and Google heats up significantly, and the battlefield spills over from "model capabilities" to "product ecology and content/distribution": On the one hand, OpenAI continues to promote multi-modal products such as ChatGPT and Sora; on the other hand, Google accelerates the launch of stronger agents and AI research capabilities around the Gemini system, and strengthens its multi-modal and infrastructure advantages.
Disney officially allows Sora to use more than 200 characters (with supporting equity investment and betting), which directly enhances Sora's appeal and commercial imagination in entertainment and short video scenes. Therefore, this is undoubtedly a reinforcement of Sora's "AI Wensheng video product moat", borrowing top IP and Hollywood endorsement to fight against Google. Making Sora's character library "legally callable" through authorization is equivalent to establishing a compliance channel in high-risk areas of AI content generation (copyright, publicity rights, trade unions and platform rules). And this authorization innovatively uses "warrants as the core of the consideration instead of cash license fees", which essentially allows Disney to exchange "current licensing income" for upward exposure to the success of OpenAI/Sora; for OpenAI, it reduces the pressure on cash expenditures and exchanges for a super content party to become a customer and ally.
Google Gemini3 will undoubtedly set off a new wave of AI applications around the world in the near future. The Gemini3 series of products brought an extremely large amount of AI token processing once they were released, forcing Google to significantly reduce the free access to Gemini 3 Pro and Nano Banana Pro, and also imposed temporary restrictions on Pro subscribers. Recent trade export data from South Korea shows that demand for HBM storage systems and enterprise-level SSDs continues to be strong, further verifying that "the AI boom is still in the early construction stage when computing power infrastructure is in short supply" as Wall Street calls it.
OpenAI, the developer of ChatGPT, is determined not to admit defeat, and this wave of AI applications is expected to spread rapidly to various industries amid the fierce competition between Google and OpenAI. It is bound to bring about a new round of surge in global demand for AI computing power.
In order to urgently combat Google's Gemini 3, OpenAI has just launched GPT-5.2 against the backdrop of a "red alert" being sounded within the company. GPT-5.2 is OpenAI's most advanced artificial intelligence (AI) large model to date. It is fully optimized for professional work scenarios and has set the strongest records in the industry in multiple benchmark tests. Among them, GPT-5.2 Thinking has set a new record in the history of the SWE coding ability test. It is also OpenAI's first model whose performance reaches or exceeds the level of human experts. In addition, GPT-5.2 is superior to its predecessor in creating spreadsheets, making presentations, image recognition, code writing, and long text understanding, aiming to "create more economic value for people."
In the scientific research field focused by some AI researchers, GPT-5.2 Pro achieved an accuracy of 93.2% in the GPQA Diamond test, followed by GPT-5.2 Thinking at 92.4%. On the expert-level mathematics test FrontierMath, GPT-5.2 Thinking solved 40.3% of the problems, setting an unprecedented record. OpenAI calls GPT-5.2 Pro and GPT-5.2 Thinking “the best scientist assistant models in the world.”