OpenAI’s unexpected deprecation of Sora 1 on March 13, 2026, may signal the beginning of an AI bubble collapse across the entire generative media sector. The company has quietly shifted users away from its legacy video generation tool toward Sora 2, marking not just a technical upgrade but a tacit admission that the original product failed to sustain momentum in a crowded market.
Key Takeaways
- OpenAI deprecated Sora 1 on March 13, 2026, forcing all US users to Sora 2 as the single updated experience.
- Sora app installs fell 45% month-over-month to 1.2 million in January 2026; consumer spending dropped 32% in the same period.
- Competition from Google’s Gemini (Nano Banana model) and Meta AI’s Vibes video accelerated Sora’s decline.
- OpenAI CEO of applications Fidji Simo stated the company is “actively looking at which areas to deprioritize” to focus on coding and enterprise.
- Manifold markets predict a 27% chance OpenAI will retire the full Sora app entirely in 2026.
The numbers don’t lie: Sora’s collapse in real time
When Sora launched in October 2025, it dominated the App Store and generated genuine excitement about AI-powered video generation. Six months later, the reality is brutal. App installs plummeted 45% month-over-month to just 1.2 million in January 2026, while consumer spending tanked 32% in the same window. These aren’t gradual declines—they are freefall metrics that suggest the product lost appeal almost overnight.
What makes this worse is that OpenAI tried to fix the problem. The company shifted from an opt-out copyright model to an opt-in approach after Hollywood backlash, and even announced a Disney deal allowing character use. Neither move stopped the bleeding. A partnership with one of the world’s largest entertainment companies should have reversed the trajectory. It didn’t. That’s when you know a product is broken at its core.
Competition exposed Sora’s weakness
Sora didn’t fail in isolation. Google’s Gemini app, particularly its Nano Banana model, and Meta AI’s Vibes video feature both gained traction while Sora lost users. The market proved it wanted AI video generation—just not Sora’s version. This matters because it reveals the difference between a category problem and a product problem. The category is real. Sora’s execution was not.
OpenAI’s response has been to quietly kill the original product and consolidate around Sora 2, a move the company frames as “evolution” rather than failure. The language matters for PR, but the outcome is identical: Sora 1 is dead, and users have no choice but to migrate. This is what a soft shutdown looks like in the age of cloud-based software.
Is this the start of an AI bubble collapse?
OpenAI’s strategic shift tells the real story. CEO of applications Fidji Simo told Wall Street Journal that the company is “actively looking at which areas to deprioritize” because they “cannot miss this moment” on productivity and business tools. Translation: generative media is no longer a priority. The company is pulling resources away from consumer-facing creative tools and doubling down on enterprise and coding—where margins are higher and competition is less brutal.
This is textbook bubble behavior. When a category launches with massive hype, early entrants assume they’ve created a permanent market. Six months of declining metrics later, they realize they’ve created a crowded commodity market. The rational response is to retreat to defensible positions—enterprise, B2B, integration into existing products like ChatGPT. Consumer apps are left to die.
Manifold markets currently assign a 27% probability that OpenAI will retire the Sora app entirely in 2026. That’s not a sure bet, but it’s a meaningful odds-on proposition. When prediction markets start pricing in full shutdown, it’s a signal that serious observers see the writing on the wall.
What happens to the broader AI hype cycle?
Sora’s collapse matters because it tests whether the AI boom has real staying power or whether it’s another cycle of inflated expectations meeting market reality. The answer is complicated. OpenAI is not abandoning AI—it’s abandoning this particular consumer product and refocusing on areas where it can maintain defensibility: enterprise, coding integration, and ChatGPT ecosystem lock-in.
That’s not necessarily an AI bubble collapse. It’s a consumer generative media collapse. The distinction matters. Startups that bet everything on consumer video generation tools are in trouble. But companies that integrated AI into existing enterprise workflows or built moats around API access are fine. OpenAI will survive. Sora won’t.
The real question is how many other AI products are one year away from the same decision. How many startups are currently burning through venture capital on consumer AI apps that will never hit profitability? How many investors are about to realize that “AI-powered” is not a business model, it’s a feature? Those answers will determine whether this is a minor course correction or the beginning of a genuine reckoning.
Did the Disney partnership fail to save Sora?
OpenAI announced a deal with Disney allowing character use in Sora, a move designed to unlock entertainment value and differentiate from competitors. The partnership proceeded despite broader Hollywood copyright tensions, but it failed to reverse Sora’s decline. This suggests the problem wasn’t licensing or content—it was the product itself.
Why is OpenAI deprioritizing Sora now?
Fidji Simo stated OpenAI is “actively looking at which areas to deprioritize” to focus on productivity and business applications, describing the situation as a “code red”. With app installs and spending collapsing, Sora became a resource drain competing for attention with higher-margin enterprise initiatives.
What’s the difference between Sora 1 and Sora 2?
Sora 2 is described as a single, updated experience that reduces complexity from older models and infrastructure, available across web and mobile. Sora 1 supported older video and image generation but is no longer accessible in the US as of March 13, 2026.
OpenAI’s Sora shutdown is not the end of AI or even the end of AI video generation. It’s a reminder that hype and product-market fit are not the same thing. The market wanted AI video tools. It just didn’t want Sora. That distinction will define which AI companies survive the next two years and which ones disappear into the graveyard of venture-backed failures.
Edited by the All Things Geek team.
Source: TechRadar
