OpenAI confirmed plans to discontinue its Sora video generation platform, including the consumer app, developer API and ChatGPT video features. CEO Sam Altman notified employees Tuesday that video products would phase out to refocus resources on business tools and coding capabilities ahead of a possible fourth-quarter initial public offering. The move dissolves a December multiyear Disney licensing deal that permitted Sora users to generate content featuring Mickey Mouse and Yoda.
Sora launched in late September 2025 to immediate viral success, topping Apple's App Store charts and reaching 1 million downloads within five days. Lead engineer Bill Peebles described explosive growth forcing compute rationing as server demand overwhelmed capacity. Despite cinematic-quality output from text prompts, sustained operations proved financially unviable amid skyrocketing chip costs.
Executives cited compute scarcity as the decisive factor, redirecting Sora's research team toward robotics training where video simulations enhance real-world task learning. "Side quests" like consumer video generation diverted cycles from agentic systems capable of autonomous software development and enterprise analytics. OpenAI recently consolidated ChatGPT, Codex and browser functions into a unified desktop superapp.
The abrupt reversal stunned Hollywood partners exploring Sora for storyboarding and effects work. Observers noted energy-intensive video rendering—far exceeding text or image generation—created prohibitive scaling barriers absent subscription monetization. Chinese competitors quickly replicated capabilities, diluting first-mover advantage.
Internal documents revealed Sora's lead staffer labeling economics "unsustainable" weeks prior. OpenAI promised timelines for data export and API sunsetting, though professional users face workflow disruptions. The decision underscores maturing AI triage where GPU allocation favors high-margin enterprise applications over viral consumer experiments.
Broader implications ripple through generative media startups burning venture capital on similar video models. Investors now scrutinize path-to-profitability for bandwidth-hungry modalities as OpenAI prioritizes physical world applications over digital content creation.