OpenAI is shutting down Sora, its TikTok-style AI video app that launched just six months ago, according to TechCrunch AI. The company gave no reason for the shutdown and didn’t share a timeline for when the app will officially go dark.
This is a notable retreat. When Sora launched as an invite-only platform, demand was massive. But the hype didn’t translate into staying power. The app peaked in November with roughly 3.3 million downloads, then slid to about 1.1 million by February. For context, ChatGPT pulls 900 million weekly active users. Over its lifetime, Sora generated an estimated $2.1 million from in-app purchases: pocket change for a company already burning through cash at a staggering rate.
What Sora Was (and Why It Got Weird Fast)
Sora cloned TikTok’s vertical video feed, but every clip was AI-generated using OpenAI’s Sora 2 model. Its headline feature let users scan their faces and create realistic deepfakes of themselves. Those digital avatars could be made public, meaning anyone could generate videos starring your face.
Predictably, things went sideways almost immediately:
- Deepfakes of public figures flooded the platform despite OpenAI’s guardrails. Videos of Martin Luther King Jr. and Robin Williams prompted their daughters to publicly ask users to stop.
- Sam Altman became a meme, starring in AI-generated clips walking through slaughterhouses and stealing Nvidia chips from Target.
- Copyrighted characters showed up everywhere: Mario smoking weed, Pikachu doing ASMR, Naruto ordering Krabby Patties.
What stands out here is how quickly the app became an under-moderated deepfake playground, despite OpenAI’s stated policies against generating videos of non-consenting public figures.
The Disney Deal That Died With It
The copyrighted character flood led to an unexpected twist. Rather than sue, Disney offered OpenAI a $1 billion investment and a licensing deal covering Disney, Marvel, Pixar, and Star Wars characters. It looked like a landmark moment for AI content generation.
But with Sora gone, that deal collapsed too. TechCrunch AI reports that no money actually changed hands before it fell apart. Disney offered a diplomatic statement about continuing to “engage with AI platforms” going forward, corporate speak for “we’ll talk to whoever’s next.”
Why This Matters
Sora’s shutdown signals something important: an AI-first social feed isn’t a product people actually want to use daily. The technology behind it is genuinely impressive. Sora 2 can generate realistic video and audio that would’ve seemed impossible two years ago. But impressive tech doesn’t automatically mean product-market fit.
This follows a familiar pattern. Meta’s Horizon Worlds faced the same problem: technically ambitious, heavily promoted, but lacking the sustained engagement that justifies the investment. Building a social network is hard enough. Building one where all the content is synthetic turns out to be even harder.
The broader implications are worth watching:
- The deepfake risk hasn’t gone anywhere. Sora 2 still lives behind ChatGPT’s paywall, and competitors are building similar tools.
- AI social apps may need a different model. Pure AI-generated feeds might not hold attention the way mixed human-AI content does.
- Licensing deals remain fragile. The Disney partnership collapsing shows how dependent these agreements are on platform viability.
OpenAI is likely better off focusing its compute and resources on ChatGPT, where the user base and revenue actually live. But someone else will try this concept again. The technology is too accessible and the format too tempting. The question is whether the next attempt will solve the moderation problem Sora never could.
More details on the shutdown and the Disney deal fallout are available at TechCrunch AI.