Over the past two years, AI-generated videos have been everywhere, but can you name even one 'AI-produced feature film'? Unfortunately, as of now, the answer remains zero.

It's quite strange, isn't it? With so much hype around the technology, why has it hit a roadblock when it comes to 'feature films'? The reason lies in the production process: a theatrical-grade feature film consists of hundreds or even thousands of shots, requiring consistent characters, environments, and performances throughout, as well as passing the test of professional post-production. These challenges are not about generation capabilities but rather about the 'production system.'

Fortunately, there are signs that this status quo is beginning to change:

Utopai Studios, the world's largest AI film and television company, has set its theatrical plans for 2027: Three works will be brought to the big screen on a large scale. This will be the first time that AI-native feature films are released in theaters on such a large scale.

So, what makes Utopai Studios' AI works different? Why can it make rapid progress in the field of feature films?

What does AI and technology mean for a film and television company?

First, let's clarify Utopai's positioning.

Founded in 2022 in Mountain View, California, Utopai Studios was dubbed the 'world's largest independent AI-native film and television company' by Variety in a report in late August this year. Note that it is a film and television company, not a model company or a technology supplier. According to Forbes estimates, Utopai is valued at around $1 billion.

The reason for its positioning as 'AI-native' lies in two aspects: First, its self-developed production intelligence platform, PAI; second, its video generation model, Utopai X, customized based on MiniMax H3. We'll delve into these later.

When many people hear about AI film and television, they subconsciously assume it is a product entirely created by AI. However, in reality, AI is just a tool. Just as AI glasses focus on the glasses themselves, and AI rings focus on the rings, AI film and television focuses on the 'film and television' aspect. You must first possess the ability to tell stories; models and platforms are merely tools for filmmaking. Stories and creators are the ultimate goal. This is also Utopai's philosophy: The significance of technology is not to replace filmmakers but to empower them with new capabilities.

Therefore, although it is an AI-native film and television company, Utopai is not an AI-driven company but rather a content-driven one. It does not sell models, nor does it make its models adapt to all scenarios. What it aims to do is leverage AI to serve its own film and television projects, using models as productivity tools and integrating AI into its production system rather than riding the AI hype wave.

Perhaps specializing in its field, Utopai's custom video generation model, Utopai X, achieved the second-highest global ranking and the top spot in the United States on the text-to-video leaderboard of the independent evaluation agency Artificial Analysis in late September.

For your information, Artificial Analysis is a well-regarded independent AI evaluation agency in the industry. Its Video Arena employs blind evaluations, meaning audiences vote for two videos generated from the same prompt without knowing the model brands, with rankings determined solely by output quality. Utopai X ranked second with a score of 1150, just behind Wan 3.0, surpassing well-known models like Dreamina Seedance 2.5. It was also the highest-ranked model from a U.S. company on the list; its audio synchronization and physics capabilities even ranked first.

It is said that this is the first time a custom model from a film and television company has entered the top two since the blind evaluation mechanism was introduced. Of course, Utopai did not fine-tune its model to participate in evaluations; it did so to produce its films. Evaluation was just a byproduct, not their goal.

As seen from the leaderboard, the top spots are generally occupied by model manufacturers and tech giants, which do not lack computational power or funding. Why can a filmmaking company outperform specialized model teams in model evaluations?

The answer lies in the 'data flywheel.' Utopai X is used in real feature film production, where artists and technical teams provide feedback on the model based on real needs such as character design, environment construction, and shot organization. The model is repeatedly refined in real projects, forming an optimization path that is difficult to replicate in laboratory environments.

Ultimately, Utopai X excels in challenging areas of generative video systems, such as reflections, light spots, spatial understanding, and advanced consistency. These capabilities are crucial for filmmaking: the reflection of light on surfaces, light spots formed by water refraction, the reasonable positioning of objects in a three-dimensional environment, and the continuous movement of characters and cameras in space—these details determine whether a shot feels realistic.

On the other hand, producing a feature film is entirely different from generating a 30-second video. A theatrical-grade feature film consists of hundreds or thousands of shots, requiring consistent characters, environments, and performances throughout a long narrative. Directors also need to repeatedly adjust shots, lighting, and scenes. These challenges cannot be solved by general-purpose video models because they do not understand 'projects.' Therefore, Utopai decided to create its own project-aware system, namely the production intelligence platform PAI (Production Intelligence), which we mentioned earlier.

In PAI's workflow, scripts can be broken down into scenes and shots, with characters, scenes, and props tied to the project throughout. AI agents act as 'production assistants,' helping filmmakers plan shots, experiment with creative ideas, and generate different versions. Utopai X is deeply integrated into PAI and can seamlessly connect with professional post-production tools like Premiere and DaVinci Resolve. However, in the most critical decisions, such as which version to choose, which shots to keep, and which details to modify, the final say always rests with the filmmakers.

As Zijian He, Utopai's Chief Science Officer, said, 'Technology should expand the canvas for filmmakers; the creative vision still comes from the people making the films.'

Zijian He is also an industry heavyweight with an impressive resume. He previously led the multimodal generation team at Meta's Superintelligence Lab, was the driving force behind the flagship video model MovieGen, and was deeply involved in the development of Llama 4. Having someone of his caliber build a production system for filmmakers and state that 'technology should expand the canvas for filmmakers' demonstrates who holds the real power at Utopai—AI or creators.

Three Films Hitting Theaters in 2027

No matter how strong the technology is, as a film and television company, it must prove itself on the big screen. In 2027, Utopai plans to release three films on a large scale in theaters, spanning from family-friendly animations to artistic dramas and historical epics—a wide range. As a company only three years old, it could have started with the easiest projects, but Utopai has set itself three increasingly difficult challenges.

The first film is The Most Serious Fart, a family animation adapted from a bestselling children's book in 2023, scheduled for release in the summer of 2027. Its most eye-catching aspect is its creative team: Oscar-nominated screenwriter Joe Stillman (Shrek), producer Russell Hollander (Despicable Me), Pixar and Laika veteran John Lee (Coraline, The Boxtrolls), storyboard artist Andy Brooks (How to Train Your Dragon), and casting director Scott Muller (Frozen, Zootopia)—a table full of familiar faces from Hollywood's animation system.

The second film is Half Moon, also known as the Korean version of The Farewell, a family drama directed by Berlin Silver Bear winner Hyo-joo Yang and co-produced by Germany's In Good Company and South Korea's Paper Barn. The film tells the story of a cross-cultural family living in Germany and the emotions long misunderstood and silenced among loved ones. There are no Deliberately created (deliberately manufactured) dramatic conflicts; instead, it focuses on familiar yet subtle moments between family members, all about love that was never spoken.

The third film is Cortés, a 16th-century historical epic with a screenplay by Oscar-nominated writer Nicholas Kazan, depicting the clash of two civilizations after Spanish conquistadors entered the Aztec Empire. This type of subject matter is notoriously difficult to get greenlit in Hollywood due to the potential for controversy in cross-civilizational narratives, where any one-sided portrayal of heroes or villains can spark debate. Traditionally, almost no one dares to touch such projects under the standard production model.

Utopai launched this project in 2025, spending significant time on historical research and script polishing, aiming to use its new production system to bring such 'worthwhile but untouchable' stories to the big screen.

All three films will be released on a large scale in theaters, which is also the biggest difference between Utopai and 'AI short film creators.'

On the series front, the sci-fi series Space Nation is the first AI-generated series to reach a partnership with traditional copyright buyers. Brazil's Globoplay and Germany's ZDF Studios have acquired distribution rights at market prices for non-AI content, with a 2027 debut. According to Forbes, assuming delivery is completed as agreed, the combined potential presale revenue for Space Nation and Cortés could reach up to $110 million.

The fact that mainstream distribution systems are willing to pay real money for AI film and television—with Globoplay and a German TV station purchasing distribution rights for Space Nation at market prices for non-AI content—carries more weight for a film and television company than any trophy or evaluation score.

From Home Studios to a Global Collaboration Network

Of course, the theatrical film lineup is just one aspect of this company. Over the past two months, Utopai has clearly been 'evolving' toward becoming a full-industry-chain company, with a Iconic action (symbolic move) being the external commercialization of PAI. Utopai's strategy for PAI is to collaborate with strategic partners to enable high-quality AI long-form production.

For example, in early September, Utopai announced an AI-assisted animation collaboration with Japan's DeNA. To provide some context, DeNA is a well-known Japanese mobile gaming and entertainment company that co-developed mobile games like Pokémon Trading Card Game Pocket with The Pokémon Company.

In this collaboration, DeNA's Entertainment Development Division will introduce PAI into its experimental animation production process. This marks PAI's first step outside Utopai's own studios as a commercialized product, making Utopai the first company to bring an AI production system into the Japanese animation industry. This move is expected to significantly impact Japan's traditionally hand-drawn animation industry, though the premise of the collaboration remains that story, style, and final creative control always rest with the artists.

In China, Utopai announced a partnership with Huace Film & TV in April this year to jointly develop the animated series Journey to the West: The Lost 500 Years. For Chinese audiences, this is also its closest appearance to date.

Its sports and entertainment footprint is also rapidly globalizing. Utopai's investors include Tom Ryan, former president of Paramount+, and director Roland Emmerich. In May this year, NBA Hall of Famer Carmelo Anthony's Creative 7 completed a strategic investment; it also collaborated with James Harden on an AI animation short film.

Coupled with its Asia-Pacific layout (layout), such as acquiring South Korea's Alquimista Media and appointing Hyun Park, former head of Studio Dragon's Global Business Division, as APAC president (focusing on Asian production branches with operations in South Korea and Japan), it is clear that Utopai is not just indulging in self-congratulation. A production and distribution network spanning East and West is taking shape, whether in Hollywood, Japan, South Korea, or China.

At the same time, Utopai is gradually being accepted by the mainstream system. The film Bedford Park, with Hyun Park as executive producer, won the U.S. Dramatic Special Jury Award for Debut Feature at the Sundance Film Festival, with global distribution rights subsequently acquired by Sony Pictures Classics. At this year's Toronto International Film Festival, Utopai co-presented the 'Future in Focus' panel with Variety, featuring participation from well-known filmmakers like James McAvoy and Rebel Wilson. Notably, Utopai is also the first AI film and television company to join the Writers Guild of America and the Screen Actors Guild.

When you put all these moves together, I wonder if you've noticed: As the world's largest AI film and television company, Utopai is gradually being accepted by the entire film industry under the standards of 'a normal film company.' Yet it differs from 'a normal film company' in that the tech circle recognizes its fine-tuned models, the distribution circle recognizes its creative content, and industry associations recognize its unique identity. This is the true value behind the title of 'the world's largest independent AI-native film and television company.'

2027: The True Test on the Big Screen

Undeniably, discussions around AI film and television are currently in a strange phase: Technology news breaks daily, with everyone weighing in, but truly engaging works that people sit through are rare. Most of the conversation revolves around 'whether it can be done,' not yet 'whether it looks good.'

But 2027 will be different. Utopai's three films will hit theaters, facing ordinary audiences in countless screening rooms who don't follow tech news. They won't care about company rankings or which model is stronger; they'll do one thing after leaving the theater: tell their friends whether the film was worth watching.

Leaderboards prove technology; theaters prove films. 2027 might just be the first year of AI film and television, when discussions move from tech circles into screening rooms. And the answer to 'whether it's worth watching' in those screening rooms has never been determined by models—it's determined by stories.