A major streamer just bought an AI-generated franchise film
In mid-October 2026, Variety broke the news that Prime Video had acquired global rights to Cukur: Vartolu, an AI-generated spin-off film from the Turkish hit series Cukur, known internationally as The Pit. The film is produced by Ay Yapim through its Singapore-based AI studio, AI Yapim, and the deal has continued generating coverage from outlets including Yahoo Entertainment and MovieWeb through late October.
This is not AI Yapim's first project on Prime Video. In August, the studio launched Castle Walls, an AI-generated historical drama described as the first show made end-to-end with AI to debut on a major global streamer. But Castle Walls was small in scope, just two 15-minute episodes, essentially a proof of concept. Cukur: Vartolu is a different scale entirely: it's built on an established franchise with a massive built-in audience, which Ay Yapim itself described as taking the approach 'from an original story to an established drama franchise.'
That distinction matters. One AI-generated show is an experiment. A major streamer licensing an AI-generated spin-off of a franchise that has aired in 105 countries since 2017 and racked up 6.4 billion cumulative YouTube views is something closer to a business model.
How the film was actually made
The production details matter more than the headline. According to Ay Yapim's own statement, 'every frame of the film was generated with AI, from the characters' faces to every setting.' That sounds like a fully synthetic, prompt-generated movie. It wasn't.
The same statement clarifies the actual workflow: 'The acting is human: actors performed the scenes, and AI generated the characters over their movement, timing and expression. All dialogue was recorded by human voice actors.' In other words, this is a performance-capture pipeline. Real actors did real scene work, real blocking, and real line reads. AI was then layered on top to generate the final faces, bodies, and environments, driven by that human performance data rather than by text prompts alone.
The brief reviewed for this piece doesn't confirm whether the original Cukur cast members reprise their roles in Vartolu, who else is in the cast, or whether performers are credited under their own names. Those are open questions the press hasn't answered yet, and it's worth saying plainly rather than guessing. What is confirmed is the production approach, directed by Umut Aral and produced by Kerem Catay, with Dafne Eksioglu and Yuce Zerey as executive producers.
Why performance-driven AI beats prompt-only generation
For short-form creators, the technical choice Ay Yapim made is the real story, more than the acquisition itself. Text-to-video generation without a human performance reference tends to drift: faces shift slightly between shots, expressions feel generic, timing feels off because there's no real actor setting the pace of a line or a reaction.
A performance-driven pipeline, where a real person acts out the scene and the AI generates on top of that movement, timing, and expression, solves the consistency problem at its source. The AI isn't inventing a performance from a text description; it's rendering over one that already exists and already has natural timing, blocking, and emotional beats. That's a big part of why a studio was willing to scale this approach from a 30-minute proof of concept to a full franchise film.
The lesson scales down. A creator building a recurring character or cast across episodes faces the same consistency problem Ay Yapim faced, just at a smaller budget and smaller crew. Prompting for the same character over and over tends to produce a character who looks subtly different each time: different face shape, different proportions, different vibe. Starting from an actual human performance (even if it's just you or a friend reading lines and acting out the scene on camera) and generating AI visuals on top of that reference gives the AI something consistent to lock onto, scene after scene.
The disclosure standard is coming either way
One detail in Ay Yapim's statement is easy to skim past but worth noting: they explicitly separated what was AI and what was human in their own press materials. Faces and settings: AI-generated. Acting and dialogue: human. They didn't bury that distinction or let the marketing imply a fully synthetic product.
That's a credibility choice, not a legal requirement mentioned anywhere in the coverage reviewed here. But it's a reasonable bet that audience expectations and platform policies will move toward requiring this kind of disclosure, not away from it. Creators who get used to labeling what's AI-generated versus human-performed now, before it's mandated, build an audience relationship based on trust rather than catching up after the fact when a platform forces the issue.
What this means for your next episode
You're not producing a franchise film with a Singapore-based AI studio and a nine-figure IP behind it. But the underlying technique that let Ay Yapim scale from a 15-minute proof of concept to a feature film is available to creators working at a much smaller scale: perform the scene for real, then generate on top of that performance instead of starting from a blank prompt.
If you're building a recurring cast for a short-form series, tools like Episodo are built around that same logic, generating consistent characters from a performance or reference rather than reprompting from scratch each time, which is the same problem Prime Video just paid to solve at a much bigger scale.
The other habit worth adopting now is disclosure. State plainly what's AI-generated and what's a real performance, the way Ay Yapim did in its own press statement. It costs you nothing and it's the direction this entire format is heading.