OpenAI Kills Sora After Six Months, Disney Walks Away from $1 Billion Deal

OpenAI is shutting down its controversial AI video app as downloads plummeted 67% and its biggest partnership collapsed.

Professional video camera on a tripod in a production studio

OpenAI announced Tuesday it’s shutting down Sora, the AI video generation app it launched just six months ago. The same day, Disney confirmed it’s abandoning a $1 billion investment deal that would have licensed over 200 characters, including Mickey Mouse, Cinderella, and Marvel heroes, for the platform.

“We’re saying goodbye to Sora,” OpenAI stated. “We’ll share more soon, including timelines for the app and API and details on preserving your work.”

The Numbers Tell the Story

Sora’s collapse was swift. Downloads peaked at 3.3 million in November 2025, then cratered. By February 2026, monthly downloads had fallen to around 1.1 million, a 67% decline in three months.

The app never gained the traction OpenAI expected, despite having 900 million ChatGPT subscribers to cross-promote to. Users who tried it often didn’t come back.

Why OpenAI Pulled the Plug

Video generation is computationally brutal. OpenAI said it needs to make “trade-offs on products that have high compute costs” as it refocuses on coding, reasoning, and text generation.

The company is also under pressure from Anthropic, whose Claude models have been gaining ground with businesses and engineers. Anthropic deliberately avoided image and video products to concentrate compute on text and code, a strategy that’s paying off.

OpenAI said its Sora research team will pivot to “world simulation research to advance robotics that will help people solve real-world, physical tasks.” Translation: the technology might resurface in a different form, but the consumer app is dead.

The Disney Disaster

In December, Disney announced a three-year partnership with a $1 billion investment. The deal would have licensed characters from Frozen, Star Wars, and the Marvel multiverse for Sora videos.

No money ever changed hands. With Sora shutting down, Disney issued a brief statement: “We respect OpenAI’s decision to exit the video generation business and to shift its priorities elsewhere.”

The collapse came before OpenAI’s expected initial public offering, removing what would have been a major marketing talking point.

Six Months of Controversy

Sora was plagued by problems from launch. Users immediately started generating videos featuring copyrighted characters, including McDonald’s Ronald McDonald appearing in Love Is Blind episodes.

OpenAI initially told copyright holders they needed to opt-out by submitting individual takedown requests for each video. Three days later, facing industry backlash, OpenAI reversed course and required opt-in permission before characters could appear.

The “cameo” feature, which let users upload videos of themselves and friends to insert their likenesses into AI-generated content, drew privacy concerns. When users created “disrespectful depictions” of Martin Luther King Jr., OpenAI had to block the civil rights leader’s likeness. Viral clips featuring Michael Jackson and Tupac Shakur raised similar questions about using the dead without consent.

What This Means

OpenAI built its reputation on launching ambitious products, from GPT-3 to ChatGPT to DALL-E. Sora was supposed to extend that streak into video. Instead, it became the company’s first major product failure.

The shutdown signals a broader recalibration. Generative AI video remains expensive to run and legally treacherous. Competitors like Runway, Kling, and Pika continue operating, but none have achieved mainstream adoption.

For users who created content on Sora, OpenAI promises details on “preserving your work.” History suggests acting quickly is wise. When companies shut down services, data preservation timelines tend to be shorter than expected.

The Bigger Picture

This is the second high-profile AI video product to face reality in 2026. Just weeks ago, Google scaled back Veo availability citing similar compute constraints.

The promise of AI-generated video captured massive attention, but delivering it profitably while navigating copyright, consent, and compute costs has proven harder than the demos suggested. The technology exists, but the business model remains elusive.