Despite Sora's impressive performance at its initial launch, it ultimately lost the competition for computing power resources to productivity tools due to daily operating costs of one million US dollars, a sharp decline in users, and difficulties in monetization. This move not only rendered Disney's $1 billion cooperation investment worthless but also marked OpenAI’s strategic contraction ahead of its IPO, redirecting core resources toward 'super applications' such as AI agents with stronger commercial value.
OpenAI once positioned the video generation tool Sora as the most important consumer-grade product after ChatGPT, but it has abruptly halted its public access in less than six months. This outcome perhaps illustrates that in the AI race where computing power is scarce and competition intensifies, products that continuously burn cash without generating revenue have no room to survive.
The immediate trigger for Sora's shutdown was a grim financial ledger—global active users plummeted from a peak of about one million to fewer than 500,000, with daily operational losses reaching approximately one million US dollars. Meanwhile, OpenAI is urgently reallocating computing resources for a new model codenamed 'Spud' to support the development of programming and enterprise service products, while Sora was consuming too much of this most scarce resource.
This decision has directly impacted OpenAI's core partners. $Disney (DIS.US)$ Previously, Disney had signed a multi-year licensing agreement with OpenAI, pledging a $1 billion investment and granting access to over 200 of its IP characters for Sora’s use. Many Disney executives were informed of the closure decision less than an hour before the announcement, and the $1 billion investment has yet to materialize, effectively bringing the partnership to a standstill.
OpenAI stated that it would concentrate the released computing power on productivity tools to catch up with competitors like Anthropic, which have gained an early advantage in the enterprise market. This shift indicates that as the company approaches its IPO, monetization capability has become the primary criterion for resource allocation.
A Fleeting Bloom: The Growth Dilemma Behind Its Viral Fame
Sora’s creation stemmed from the academic ideals of two doctoral students from the University of California, Berkeley. Tim Brooks and Bill Peebles joined OpenAI in early 2023, dedicating themselves to building an AI model capable of generating high-quality videos from text and simulating the physical world. In February 2024, they named the system 'Sora,' inspired by the Japanese word for 'sky,' and unveiled it to the public. Sora astonished the industry with hyper-realistic videos such as mammoths traversing snowy plains and fashionable women strolling through neon-lit streets of Tokyo. Sam Altman subsequently invited users on the X platform to submit text prompts, collectively showcasing its generative capabilities.
In December of the same year, OpenAI officially launched the consumer-grade Sora application for the public. Within a week of its release, the app quickly climbed to the top of the App Store charts. Users simply needed to input commands to generate a ten-second video clip within minutes. Additionally, the feature allowing users to upload their faces enabled them to star in various fantastical scenarios. Altman himself contributed his likeness, sparking a wave of user interest.
However, the hype faded just as quickly. After peaking at around one million global users, the number steadily declined, shrinking to fewer than 500,000 in the following months. According to data from analytics firm Similarweb, usage had already stagnated by the end of the year. Altman once likened Sora’s launch to the historic moment of ChatGPT’s debut, but the app ultimately failed to realize its creators’ vision—the Wall Street Journal described its actual performance as 'more like AI dross than AI magic.'
A Black Hole of Computing Power: The Daily Cost of One Million Dollars
The root of Sora’s exorbitant operational costs lies in the technical characteristics of the video generation model itself. Unlike language models that learn from text, video models need to understand and reconstruct entire dynamic scenes, making their training and inference costs far higher than text-based products. Every user who embedded their face into WWII newsreels or Hollywood car chase scenes consumed a portion of limited AI computing power. According to media reports citing insiders, Sora incurred daily operational losses of approximately one million US dollars.
OpenAI has an internal dashboard to track the allocation of computing power among various teams. Some employees were surprised by the proportion of computing resources allocated to the Sora team—video generation tools neither generate significant revenue nor enhance language model capabilities, prompting internal questioning about this resource distribution.
External competitive pressures have made this issue increasingly urgent. Google Gemini has gained widespread consumer adoption, while Anthropic's coding tool Claude Code has rapidly captured the Silicon Valley software engineering community with its highly autonomous programming abilities, catching OpenAI off guard. In response, OpenAI hastily launched a new version of its coding product Codex but still struggled to close the gap.
The new model, codenamed 'Spud,' urgently required more computing power. Meanwhile, the company also planned to train a separate new model for video generation within ChatGPT. After cost calculations, OpenAI ultimately decided to cancel the training plan and completely shut down Sora.
Meta's Talent Poaching and Internal Isolation
Sora once came close to an early demise due to the intense talent war. In the spring of 2025, Meta CEO Zuckerberg initiated a large-scale poaching campaign targeting OpenAI, personally reaching out to dozens of top researchers with lucrative offers to join his AI lab. Bill Peebles, co-founder of Sora, was among those contacted and seriously considered the offer. According to media reports citing insiders, OpenAI ultimately retained Peebles through a salary increase and subsequently expanded his responsibilities in the Sora project, placing him in charge of training the next-generation video generation model and developing consumer-grade applications.
However, internally, Sora maintained a highly closed operational status. The project fell under the World Simulation Team led by Aditya Ramesh and remained independent from the core research team responsible for ChatGPT’s language models. Progress on the project was kept highly confidential from other departments. Some former employees described Sora as a 'company within a company.' This isolated operational model made it difficult for Sora to gain broader internal recognition of its strategic value, leaving it in a relatively vulnerable position in the competition for company resources.
Disney's Dream Shattered: $1 Billion Investment Goes Up in Smoke
Sora’s most representative partnership came to an equally abrupt end.
In December 2024, OpenAI and Disney announced a multi-year licensing agreement covering over 200 well-known IP characters from Marvel, Pixar, and others. Disney also committed to investing $1 billion in OpenAI and becoming its primary corporate client. Disney CEO Bob Iger told CNBC that this collaboration allowed Disney to participate in the rapid growth of AI and the shaping of new media entertainment formats. Notably, just one day before announcing the deal with OpenAI, Disney had sent a cease-and-desist letter to Google over copyright infringement—a move widely seen as an endorsement of AI licensing business models.
In February this year, Iger revealed during an earnings call that short videos created using Sora would soon appear in Disney+'s vertical video feed. The two parties were also negotiating the introduction of ChatGPT across Disney’s entire organization.
However, many Disney executives were informed of OpenAI's announcement to shut down Sora with less than an hour's notice. The $1 billion investment has yet to materialize, and the partnership between the two parties has effectively stalled. Disney subsequently issued a statement saying: "As the AI field evolves rapidly, we respect OpenAI’s decision to exit the video generation business and shift its focus to other areas, and we appreciate the productive collaboration between both teams." Under the leadership of new CEO Josh D'Amaro, Disney is currently engaged in active discussions with more than ten partners regarding the introduction of alternative AI tools.
Betting on the 'Super App': Pragmatic Logic Takes Center Stage
The closure of Sora does not merely represent a product failure but is emblematic of OpenAI’s accelerated narrowing of strategic focus just before its IPO.
In an internal letter to employees, Altman characterized this decision as a 'difficult but necessary sacrifice for the overall goals of the company' and expressed appreciation for employees' willingness to make 'tough trade-offs.' A spokeswoman for OpenAI stated that the company is conducting an 'uncompromising prioritization' of computing resources based on the principle of maximizing long-term economic value, noting that 'this focus allows us to grow and innovate more rapidly while serving businesses and developers more efficiently.'
The company’s current focus is shifting toward a planned 'super app,' which will integrate so-called 'agent' AI tools capable of autonomously performing tasks such as coding, data analysis, and travel booking for users. Such productivity-oriented products are gaining rapid traction among businesses and developers, though OpenAI lags behind Anthropic in this market thus far. Altman noted that the original Sora team would redirect their efforts toward areas with greater long-term potential, such as robotics.
For OpenAI, the journey of Sora represents a costly strategic misstep. This product, once envisioned by Altman as a means to reshape popular culture and unlock new revenue streams, ultimately faltered due to the absence of a viable business model and improper resource allocation. At a critical juncture when the company is focusing on profitability, it became a burden that had to be shed.
Editor/KOKO