OpenAI shut down its standalone Sora video generation app on Tuesday, ending a roughly six-month consumer experiment that produced impressive launch numbers and a swift, irreversible decline. The Sora team posted a farewell on X: We are saying goodbye to Sora. To everyone who created with Sora, shared it, and built community around it: thank you. The Wall Street Journal, citing people familiar with the matter, reported that OpenAI is discontinuing its video AI model efforts across the board — not just the consumer app, but the underlying Sora 1 web experience, retired March 13, 2026, and the Sora 2 infrastructure that replaced it.
Sora launched as OpenAI's second-ever iPhone app in September 2025, alongside Sora 2, and hit 100,000 installs on its first day, climbing to the top of the US App Store faster than ChatGPT had. By October it had crossed 1 million downloads. An Android app followed in November. Total downloads across both platforms reached 9.6 million, with $1.4 million in consumer spending — $1.1 million of that from US users alone, according to Appfigures data reported by TechCrunch. The early trajectory looked like a genuine consumer hit.
It did not hold. Downloads dropped 32 percent month-over-month in December 2025, even accounting for the typical holiday app-downloading lift. January 2026 was worse: installs fell 45 percent month-on-month, bringing the monthly total to roughly 1.2 million, Appfigures reported. Consumer spending fell in tandem — down 32 percent month-over-month in January, to $367,000, from a December peak of $540,000. By early 2026, Sora had fallen out of the top 100 US App Store apps entirely, sitting at No. 101 on the free chart and No. 181 on Google Play. The app's highest remaining rank was No. 7 in the Photo & Video category.
The Disney partnership was supposed to reverse the trend. On December 11, 2025, The Walt Disney Company announced a $1 billion investment in OpenAI alongside a three-year licensing deal allowing Sora users to generate videos featuring more than 200 Disney characters. The deal gave Sora something competitors could not easily match: Marvel heroes, Pixar characters, and Star Wars aesthetics baked into the creative tool. The partnership did not move the needle. Usage data showed no lasting boost, and by March the Sora 2 service itself was riddled with reliability problems — videos stuck at 99 percent completion, persistent under heavy load errors, and new accounts having video generation features disabled entirely while their access was reviewed. The underlying reason, widely reported by users and documented on community forums, was a content moderation system that OpenAI described as protecting intellectual property but which users experienced as a blunt instrument that blocked entirely ordinary prompts.
The IP restrictions OpenAI tightened in response to the moderation problem effectively killed what remained of the app's appeal as a creative tool, as 9to5Mac put it. The app had been positioned as an AI-powered TikTok competitor — a place to cast yourself and friends as main characters, remix shared videos, add music and dialogue. Once those prompts started getting rejected at scale, the creative loop broke. Hank Green, the YouTube educator and author who has been one of the more thoughtful chroniclers of AI's impact on creative work, called the product SlopTok — a portmanteau of TikTok and AI slop. The name stuck.
The contrast with Runway, which remains active in the same space, is worth sitting with. Runway has pursued the creative-industry angle deliberately and without the kind of product whiplash Sora experienced. It signed a first-look development deal with filmmaker Harmony Korine's production company EDGLRD and partnered with AMC Networks across marketing and TV development. Where OpenAI built Sora as a consumer app and then pulled the plug after two quarters of declining metrics, Runway has been building relationships with professional creative studios — a slower, more expensive, and more durable path. Whether that strategy pays off is an open question. But the divergence in approach is real, and it explains why the two companies are now in very different positions.
OpenAI's pullback fits a broader organizational reorientation that became visible last week and is now taking clearer shape. Fidji Simo, OpenAI's chief product officer, told employees at an all-hands meeting in March that the company needed to stop being distracted by side quests and orient aggressively toward high-productivity use cases, according to ALM Corp, which reported on an internal Simo memo dated March 19, 2026. That memo, describing a unified desktop superapp bringing together ChatGPT, the Codex coding platform, and the Atlas AI-powered web browser under one roof, landed five days before the Sora team posted its goodbye. Greg Brockman, OpenAI's president, will temporarily oversee the product revamp while Simo leads the commercial side, ALM Corp reported.
The structural signal is unmistakable: OpenAI is folding discretionary consumer products into a concentrated productivity stack as it approaches a potential IPO. The financial calculus behind that pivot is not subtle. OpenAI is carrying roughly $100 billion in debt, Al Jazeera reported, citing the company's investor documentation. The company has raised more than $168 billion to date. Nvidia CEO Jensen Huang said his company is set to invest another $30 billion, though he suggested it might be the last pre-IPO check, per Al Jazeera — a notable qualifier from the chipmaker that has been one of OpenAI's most consistent backers. Nvidia itself is valued at $730 billion. OpenAI has hired Cynthia Gaylor, the former DocuSign CFO, to lead investor relations as it prepares for a potential public offering as early as Q4 2026, ALM Corp reported. The company's consumer products — the ones that generate the most press coverage and the least predictable revenue — are being consolidated or cut accordingly.
ChatGPT, meanwhile, reported 900 million weekly active users as of February 2026, more than double its count of 400 million a year earlier, per ALM Corp — a growth trajectory that makes bundling Sora into an existing platform substantially more attractive than maintaining a standalone app with its own install base, retention challenges, and moderation overhead. Integrating Sora into ChatGPT sidesteps the question of whether standalone AI video apps have a market at all. Bundling it into a platform with 900 million weekly users solves a distribution problem, not necessarily a product one.
The question worth asking — and that the next six months of AI video development will answer — is whether AI video generation is actually ready for mass consumer adoption the way image generation was in 2022 and 2023. Sora's answer, handed down on a Tuesday in late March, was apparently: not yet, and not that way. Sora was named after the Japanese word for sky), a reference to limitless creative potential. Six months of user data, Disney's best characters, and a $1 billion deal could not save it from the harder verdict of the market.
