Skip to content
Briefings are running a touch slower this week while we rebuild the foundations.See roadmap
Sora
ProductUS

Sora

OpenAI's text-to-video product; launched 2024; discontinued 26 April 2026 for robotics reallocation.

Last refreshed: 10 May 2026 · Appears in 1 active topic

Key Question

What does Sora's sudden shutdown mean for studios that built AI workflows around it?

Timeline for Sora

View full timeline →
Common Questions
Why did OpenAI shut down Sora?
OpenAI closed Sora on 26 April 2026, citing a decision to reallocate the underlying compute resources and research team to robotics development. The closure was announced with limited notice and surprised industry partners including Disney.Source: event
What was Sora used for before it shut down?
Sora was OpenAI's text-to-video generation product, capable of generating up to 60-second video clips from text prompts. It was used by filmmakers, advertisers, and studios for concept visualisation and prototyping, and Disney had announced plans to use it for production.Source: event
Is there an alternative to Sora for AI video generation?
Following Sora's shutdown, alternatives include Runway's Gen-3, Google's Veo 2, Pika, and Meta's Movie Gen. Adobe Firefly Video is emerging as a Creative Cloud-integrated option. None have matched Sora's previous output fidelity according to most evaluations.Source: event

Background

OpenAI shut down Sora on 26 April 2026, less than 18 months after its public launch, reallocating the compute resources and research team toward robotics development. The closure was announced with limited notice and surprised media and entertainment industry partners including Disney, which had announced a $1bn OpenAI equity stake and a Sora-based production workflow just weeks before the shutdown. Disney's AI strategy was publicly premised in part on Sora access, making the discontinuation an immediate commercial and reputational problem for both companies.

Sora was OpenAI's text-to-video generation model, publicly previewed in February 2024 and launched to ChatGPT Plus and Pro subscribers in December 2024. It could generate up to 60-second video clips from text prompts with notably high visual fidelity, and was widely cited as the most capable publicly-available video generation model at launch. It was priced as a premium ChatGPT feature rather than an API product, limiting professional integration.

The shutdown illustrates the platform risk inherent in building creative workflows on closed-source AI products from a single vendor. OpenAI's decision to prioritise robotics over video generation reflects its own shifting compute allocation and commercial priorities, not any assessment of Sora's market success. For the media industry, the episode is a cautionary data point on the fragility of AI vendor commitments, particularly for long-cycle production workflows that depend on stable tooling.