Table of Contents
ToggleSix months—that's the time span from OpenAI Sora's official launch to its complete shutdown . Sora announced this news without warning on social media this morning. Meanwhile, Disney had initially agreed to invest $1 billion in OpenAI and license classic characters like Mickey Mouse and Cinderella for Sora's film generation; ultimately, they received nothing, and the partnership dissolved silently.
This doesn't mean that AI video technology is immature, but rather that OpenAI made a wrong bet on a core assumption about consumer products.
Sora's "Social Platform" Waterloo
When Sora launched its standalone app at the end of September last year, it was positioned not as a simple video generation tool, but as a platform with social attributes: users can generate AI videos for free and browse other people's works, with a design logic closer to TikTok or Instagram.
The App Store data is honest: Sora initially climbed rapidly in the charts, but then dropped sharply and had no retention.
This result is not surprising. Currently, users download an AI video app because "I need to generate videos," not because "I want to watch AI videos generated by others." By packaging a tool-like need as a social product, it seems OpenAI has made a mistake this time.
Users don't need another TikTok; they need a faster and cheaper Premiere Pro.
Disney was taken aback
According to reports, Disney had not only agreed to invest but also planned to license its core IPs to OpenAI. The direct reason this collaboration ultimately fell through was that OpenAI reallocated its resources to coding tools and enterprise products in preparation for its IPO.
Sources familiar with the matter revealed that Disney had discussed a collaboration plan with OpenAI on Sora earlier this week. However, they were suddenly shocked to receive news that OpenAI had decided to abandon the tool entirely. One source, who requested anonymity, described the move as a surprise withdrawal.
In other words, OpenAI gave up on itself first.
OpenAI hasn't withdrawn from the video; they've just changed their wording.
However, it's important to distinguish one thing here: OpenAI shut down Sora as a standalone product, not the ability to generate videos itself.
Sora's research team has shifted its focus to world simulation research for physical simulation of robotics; the video generation function has been integrated into ChatGPT as an extension of the dialogue function, rather than a standalone application.
The business logic behind this shift is clear: ChatGPT has a base of paying users, and enterprise clients have specific needs for generating videos, so there's no need to maintain a separate app whose ranking is declining. Acquiring Sora doesn't mean OpenAI has lost in the video market to Google Veo 2, Runway Gen-3, or Pika Labs: it simply means it's no longer competing with standalone products.
Pre-IPO contraction is a rational move.
For OpenAI, the pre-IPO resource contraction is understandable. But this decision also acknowledges one thing: Sora's "quasi-social platform" hypothesis has never been validated by the market from the beginning.
As for whether a consumer market for AI-generated videos exists, and who will ultimately win, no one has yet provided a definitive answer based on available data. Sora's failure simply eliminates one possibility.
