OpenAI’s unexpected decision to shut down its Sora app, which allowed users to generate video clips by describing them, may open the door for other companies to make it big in AI video. But it also highlights the many ways in which the technology may simply be more trouble than it’s worth.
The AI giant and ChatGPT creator closed Sora this week as part of a widespread effort to cut costs, ahead of a rumoured public offering. In the six months it was available, the app created almost as many headaches for OpenAI as it did copyright-infringing SpongeBob clips for its dedicated audience.
Sora was originally revealed in early 2024, and made available to paying OpenAI customers later that year. But it didn’t hit the mainstream until it was released as an app in October 2025, hitting a million downloads in a matter of days. Structured like TikTok but with AI content generated by other users, the app encouraged people to make models of their own faces that others could use. It was immediately filled with violent, racist content, sexualised deepfakes of real people, and famous copyrighted cartoon characters in compromising situations.
In a blog post at the time, OpenAI chief executive Sam Altman said he was “hearing from a lot of rights holders who are very excited for this new kind of interactive fan fiction, and think this new kind of engagement will accrue a lot of value to them, but want the ability to specify how their characters can be used”, which is surely a generous interpretation of the calls he’d received.
OpenAI did indeed begin blocking the use of copyrighted characters. And in December, Disney announced a $US1 billion ($1.44 billion) investment in OpenAI for a three-year deal that would see its characters in Sora, and AI-generated clips on Disney+. That investment has been withdrawn following this week’s closure.
It’s clear from the app’s trajectory that the economics simply didn’t stack up. The 10 million users creating videos of things that never happened and don’t look quite right were not providing money to OpenAI, while one Forbes estimate put the cost to the company of generating the videos at $US15 million a day – enough to burn through Disney’s entire investment in two months.
The system was also a significant admin and regulatory burden for the company, which had to watch for illegal, objectionable and copyright-infringing content, while also being the poster child for a technology some worry will threaten the film industry. If another company is to step up and claim the mantle, it might ideally be one that doesn’t mind bleeding money, and is relatively lax when it comes to ethical responsibilities.
Elon Musk said this week that his xAI would double down on generative video in the wake of Sora’s closure. Between reposting racist attacks on US mayors and pointing to social welfare as a “planned demolition of America” on X, the platform he owns, the world’s richest man said the next update for xAI’s Grok Imagine tool would be epic. He has also spent the last week sharing videos made with Grok, and encouraging users to download the app.
But Grok Imagine has faced even greater criticism than Sora since it was released publicly last August. Its “Spicy Mode” ensures the video output is suggestive or explicitly sexual in nature, while videos can also be animated based on photos a user uploads.
An update that allowed users to reimagine images and videos posted to X by simply replying to the post and tagging Grok led to a proliferation of non-consensual pornography and explicit content as people responded to practically any image with “@Grok make her take her top off”, or “@Grok show her from behind”. In response to complaints in January that the tool was being used to generate child sexual abuse material, an xAI staffer said guardrails were being tightened. Musk put the blame for the content on those using the tool.
Some of Grok’s biggest fans are unhappy with the app as well, as recent changes have cut back on free features, tightened daily generation limits, and failed to make progress in clip length (the videos only last a few seconds) and realism. More limitations may arrive in particular locations as Grok Imagine is currently under investigation by regulators in the European Union, the UK, Malaysia, India, France and the state of California over its nudification capabilities.
Like OpenAI, xAI has failed to find a profitable use for generative video that isn’t tainted with ethical and regulatory problems. But that still leaves several companies that could step up to fill the void left by Sora, especially since the OpenAI model had created cottage industries of AI prompters, ad agencies, aggregators and generated video sellers, now scrambling to find a new model.
- Runway caters to Hollywood by providing AI video effects in a much more controllable package than the get-what-you-get generation of Sora and Grok. Currently favoured by professional filmmakers and high-end ad agencies, it could consider a consumer push.
- Luma Labs is focusing on cheaper, faster generation with its Dream Machine that’s been favourably compared to Sora, with high-resolution clips suitable for hobbyists. Critics have noted its training data is unclear, which is a potential roadblock to commercial work.
- Kling has been referred to as the gold standard for AI video physics, producing relatively long clips (15 seconds) with consistent characters. Developed by the partly state-owned company behind Chinese video sharing platform Kwai, it could seize on Sora’s user base.
- Google could snap up corporate Sora users, as its Veo platform is integrated into the tools those workers already have via the Gemini chatbot, as well as the dedicated AI Studio app. Veo now supports 4K output and spatial audio generation.
- Adobe could similarly be seen as a “safe and scale” alternative to Sora. It claims Firefly is trained on clean stock imagery it has the rights to, meaning generations will be safe from copyright infringement, and it’s integrated directly into Premiere Pro.
Get news and reviews on technology, gadgets and gaming in our Technology newsletter every Friday. Sign up here.
Tim Biggs is a writer covering consumer technology, gadgets and video games.Connect via X or email.From our partners
