OpenAI is in negotiations with an unnamed startup over deep access to Sora, its text-to-video generative AI model, with the aim of building AI video tools designed specifically for professional filmmakers.

The identity of the startup has not been disclosed, nor have the terms or timeline of the potential deal—standard practice for talks that remain in progress.

What Is on the Table

The negotiations center on granting a third party privileged access to Sora so it can build specialized products on top of the model. Those products would serve people who make films for a living, rather than casual users generating clips for entertainment. That distinction matters: professional filmmaking runs on established pipelines—storyboarding, previsualization, visual effects—where tools are expected to slot into existing workflows rather than simply produce standalone clips.

The arrangement also fits an established pattern. Rather than opening the floodgates, OpenAI has pursued selective partnerships around Sora. The company has previously held discussions with major studios, including Disney, Universal, and Warner Bros., about custom versions of Sora for their own internal use. And Sora is not alone in courting this audience: rival text-to-video systems such as Runway's video models and Google's Veo have also been positioned toward filmmakers and studios, making professional production one of the more competitive corners of the generative landscape.

How Sora Got Here

OpenAI first previewed Sora in February 2024. At that stage, access was limited to a select group of researchers and artists who tested the model and provided feedback.

The public launch followed in late 2024. Sora arrived positioned as a cutting-edge tool for generating video from text prompts, part of a broader wave of generative AI that moved from writing essays to producing moving images.

Since the initial announcement, various Hollywood factions have raised concerns about copyright, rights to likeness, and fair compensation for creators whose work may surface in the AI's output. OpenAI has responded by adjusting Sora's content controls and exploring licensing frameworks, while continuing to push toward integrations and enterprise collaborations.

What It Means for Hollywood and AI Video

For the film industry, a startup-built tool layer on top of Sora could reshape how AI reaches working creatives. Studios negotiating custom versions for internal use represent one path; a specialized startup selling tools to filmmakers more broadly is another.

For OpenAI, the strategy has a clear logic. Partnering with a startup allows the company to reach a specialized market without building every filmmaker-facing feature itself—the startup handles the workflow details while OpenAI supplies the underlying model. It is the same division of labor now common across the AI industry, where model developers license access to companies that build the domain-specific tooling.

Several questions are worth tracking as negotiations continue. The first is the startup's identity, which would reveal what kind of company OpenAI trusts with privileged model access. The second is what “deep access” actually means in practice—whether that involves customization, priority capacity, or something else entirely. The third is how any resulting product handles the copyright and likeness concerns that have trailed Sora since its February 2024 debut.

Reports emerging in early 2026 indicate that OpenAI is shifting its focus regarding Sora, with deployments of app features and adjustments in partnerships expected as negotiations progress.