THURSDAY, MAY 21, 2026 · BRISBANESUBSCRIBE →

THE AI POST

INTELLIGENCE. CURATED.

Empty film production set with camera and lighting equipment
BusinessApril 20, 2026

OpenAI Is Killing Sora in 5 Days. Nobody at OpenAI Seems to Care.

On April 26, OpenAI shuts down Sora after just 18 months. The $1 billion Disney deal is dead. The video generation wars continue without the company that started them.

Five days from now, OpenAI will pull the plug on Sora. The app goes dark on April 26. The API follows on September 24. The company that kicked off the generative video arms race in February 2024 with a 60-second shot of a woman walking through neon Tokyo is quietly abandoning the category it invented.

The official line from OpenAI, delivered to the BBC: they want to focus on robotics and "real-world, physical tasks." Translation: video generation does not pay. Not at the scale OpenAI needs to justify spending $14 billion a year.

The $1 Billion Disney Deal Is Dead

The most expensive casualty is a partnership that was supposed to change Hollywood. OpenAI had signed a $1 billion deal with Disney to integrate Sora into film and animation pipelines. Multiple outlets have since reported the deal is now scrapped. Eighteen months of executive meetings, legal wrangling, and creative workflow design. Gone.

Disney is not commenting. OpenAI is not commenting. But the silence itself tells you everything. Studios spent 2024 terrified Sora would replace VFX houses. Studios spent 2025 terrified Sora would put animators out of work. Studios will spend the back half of 2026 wondering why they ever took the threat seriously.

The Market Is Fine Without Sora

Here is the awkward part for OpenAI. The AI video market is not dying. It is growing. Runway Gen-4, Google Veo 3, Kuaishou's Kling, Pika, and Luma are all still shipping, still raising, still getting used. Google is reportedly integrating Veo directly into YouTube as a premium feature later this year. Runway's enterprise pipeline with major studios is humming.

Sora was not killed by a market that rejected AI video. Sora was killed by a company that decided video was the wrong bet for its IPO story. Fidji Simo's "kill the side quests" memo lands differently when you remember Sora was the product Sam Altman personally demoed on the Joe Rogan podcast. Two years later it is a side quest.

What Users Lose

Sora users have five days to download everything they generated. After April 26, the content is gone. OpenAI is not offering a migration path. There is no Sora Lite, no API continuation for existing customers, no partnership with another provider. Just a delete button with a calendar date attached.

This matters because Sora was the first major AI product shutdown to happen at scale. Anthropic killed older Claude versions but kept the brand. Google sunsetted Bard but rebuilt it as Gemini. OpenAI killed GPT-3.5 but made it a seamless upgrade. Sora is the first real goodbye. The first product that shipped to millions of users and just ends.

The Enterprise Lesson

Forbes columnist John Sviokla nailed the takeaway weeks ago: if you built your creative workflow on Sora, you are now in a scramble. Thirty days to migrate, re-prompt, re-train your team, and rebuild your library on a competing platform. Runway, Kling, and Veo have all published migration guides. The guides exist because this is going to happen again.

Every AI product decision from here forward happens under a new rule: the vendor might kill this tomorrow. No amount of API commitments or enterprise contracts will save you when the strategic calculus shifts. OpenAI had every reason to keep Sora alive. They had the models, the users, the Disney deal, and the headline value. They killed it anyway.

Who Wins

Runway is the obvious beneficiary. They have the Hollywood relationships, the Gen-4 model, and a product roadmap that never depended on a consumer app. Kuaishou's Kling just got a gift from the American market: the dominant Chinese video model now has a clean shot at enterprise accounts that were teed up for OpenAI. Google Veo inherits whatever YouTube-adjacent creative market Sora was supposed to capture.

OpenAI wins too, in its own way. One less distraction from the $500 per seat enterprise play. One less category to defend in the IPO roadshow. One less product that lost money every time someone pressed generate.

The losers are everyone who believed the AI video revolution would run through San Francisco. It will not. It runs through New York (Runway), Mountain View (Google), and Beijing (Kling). OpenAI just handed them the category and walked away.

Five days. Then Sora is gone. And the video generation wars keep going without the company that named them.

OpenAISoravideo generationDisneyRunwayKlingenterprise AI