OpenAI is in discussions to convert its abandoned Sora investment into an equity stake in a startup, according to people familiar with the matter. The move would mark a strategic pivot for the AI company, which has been reevaluating its approach to filmmaking technology after shelving the Sora project. The talks are ongoing and could still fall apart.
From Shelved Project to Startup Bet
OpenAI had been developing Sora, a video-generation model that promised to upend how films are made. But the project was quietly abandoned, leaving the company with sunk costs and no clear path to recoup them. Now, instead of writing off the investment entirely, OpenAI is exploring a deal that would give it a stake in a startup — likely one that can carry the technology forward without OpenAI bearing the full burden of development and legal risk. The exact startup hasn't been named, and OpenAI declined to comment.
The shift is unusual but not unheard of. Companies often salvage value from failed projects by spinning them out or transferring assets to partners. For OpenAI, the calculation is simple: Sora's underlying tech still has potential, but the copyright issues surrounding AI-generated video are a minefield. By taking a stake in a separate entity, OpenAI can distance itself from the legal exposure while keeping a foot in the door.
Why Copyright Remains the Elephant in the Room
Sora's biggest obstacle was never technical — it was legal. AI models that generate video by training on existing films and clips have drawn intense scrutiny from studios, artists, and regulators. The unresolved question of whether training on copyrighted material constitutes fair use hangs over every company in the space. OpenAI has already faced lawsuits over its text and image models; a video model would multiply that risk.
That's likely why the company is reconsidering its direct involvement. A startup stake would let OpenAI benefit from any breakthrough while insulating itself from the worst-case legal scenarios. The startup, presumably, would take on the licensing battles and the responsibility of building a compliant product. It's a classic risk-shifting move, and it's one that other AI firms are watching closely.
The Startup Angle: Who Benefits?
If the deal goes through, the unnamed startup would gain access to Sora's technology — and possibly OpenAI's brand and resources — without having to build from scratch. That's a significant head start in a field where compute and talent are expensive. For OpenAI, the upside is a potential return on its original investment, plus a window into a market it still finds attractive.
But there are open questions. Would OpenAI license the tech outright, or would it retain some control? How would revenue be split? And what happens if the startup also gets sued? None of these details have been made public. The people familiar with the talks cautioned that no final agreement has been reached.
What This Means for AI in Filmmaking
OpenAI's pivot doesn't signal a retreat from AI in film — it signals a change in strategy. The company still believes generative video is a big deal. It just doesn't want to be the one holding the bag if courts rule against the training practices that make it possible. By backing a separate startup, OpenAI can keep experimenting while limiting its downside.
That could become a template for other AI companies sitting on similar projects. If the Sora deal works, expect more spinoffs and stake conversions as firms look to monetize abandoned work without taking on fresh legal risk. The copyright landscape isn't getting clearer anytime soon, so the incentive to offload that risk is only growing.
For now, the talks are private and the outcome uncertain. OpenAI isn't commenting, and the startup involved hasn't been identified. The next sign of progress would be a formal announcement — or silence, if the deal collapses. Either way, the way AI companies handle their failed bets is about to get a lot more creative.
