ByteDance’s Seedance 2.0 hits Cannes with 95-minute AI film ‘Hell Grind’

ByteDance’s cloud platform Volcengine used the 79th Cannes Film Festival to showcase its Seedance 2.0 model and premiere Hell Grind, a 95-minute AI-generated feature film billed as the world’s first full-length AI movie.

The film was produced by the US-based AI company Higgsfield, while its core video generation model, Seedance 2.0, was developed by ByteDance.

TechNode described the premiere as more than a short AI demo, saying Hell Grind arrived as a complete theater-scale narrative feature rather than a 15-second proof of concept.

Long-form generation beyond short clips

That matters here because long-form video generation has been one of AI filmmaking’s biggest technical bottlenecks.

The report said most mainstream AI video tools today can only generate clips lasting between 15 and 30 seconds.

Building a feature-length film usually requires stitching together tens of thousands of fragmented shots, a process that often creates inconsistent faces, unstable scenes, and broken visual continuity.

That makes the output difficult to use in a professional production pipeline.

TechNode said Seedance 2.0 appears to have moved past many of those limits. Hell Grind follows four street kids — Roko, Jaxx, Lulu, and Rein — who discover a mysterious artifact while exploring a museum.

The find awakens a dark force and grants them superpowers, forcing them to join forces and fight for survival as reality and illusion begin to blur.

After watching an early cut, Chuck Russell reportedly said the project made him genuinely empathize with the characters, something he described as rare in AI-generated cinema.

Fast production and lower cost

The production numbers were another focus of the Cannes showing.

According to the report, the movie was completed by a 15-person team in 14 days, with a total budget under $500,000.

A traditionally produced film of similar scale could easily cost tens of millions of dollars.

At the summit, Higgsfield co-founder Alex Mashrabov said the technical infrastructure for AI-native filmmaking is already mature enough to bring ambitious cinematic ideas to life at a fraction of traditional production costs.

Questions for film work and authorship

TechNode said the implications extend beyond production speed and cost.

If a 95-minute AI-generated feature can be made this way, narrative-scale generation may no longer be filmmaking’s main constraint, shifting the bottleneck toward creative direction rather than budget or team size.

For independent creators, that could lower the barrier to entry and widen access to feature-length storytelling.

At the same time, The report said the shift raises structural questions for the film industry.

Parts of the mid- and low-tier production workforce could face displacement pressure if feature films can be produced in roughly two weeks at a fraction of traditional budgets.

It also pointed to a deeper authorship debate: whether AI-generated emotional impact reflects genuine artistic intent or optimized patterning of human responses.

As these systems become more capable of producing coherent, emotionally resonant narratives, the role of human creators may move further toward defining intent, taste, and meaning.