A film entirely created by artificial intelligence was shown in Cannes. $400,000 went to AI computing power
To create every 15 seconds of the film, a 3000-word prompt with tons of details had to be written for the artificial intelligence.
The 95-minute film "Hell Grind" cost its creators $500,000, of which $400,000 was spent on artificial intelligence computing power. The entire process took two weeks. The Wall Street Journal shares the details.
A still from the trailer of the film "Hell Grind"
The topic of artificial intelligence (AI) has been one of the main ones at the Cannes Film Festival for several years. Previously, discussions around artificial intelligence were mostly anxious: actors, directors, and screenwriters feared that technologies would start replacing people. But this year, as WSJ notes, the mood has become less panicked.
For many in the industry, the question now is merely how to use this tool. An example here is the 95-minute film "Hell Grind" (the title can be translated as "Road Through Hell"), created by the San Francisco-based startup Higgsfield AI. In this adventure action movie, absolutely everything — characters, locations, lighting, and even small props — is a product of generation.
The film's plot tells the story of four thieves who literally descend into hell. The main character, Rocco, must navigate an anti-utopian wasteland to save his beloved Lulu.
As WSJ writes, Higgsfield AI took only two weeks to create such a large-scale visual story. The project's budget was $500,000, with 80% of that amount going to computing power costs.
However, the company itself does not develop base video generation models. Instead, it uses existing systems, such as Google Veo 3, and focuses its efforts on tools that allow images to be made consistent and stylistically uniform.
The creators of "Hell Grind" insist that AI does not eliminate the profession of director. Even with full film generation, a true understanding of cinematography is required. One needs to know the principles of shot composition, shot sequencing, the basics of editing, and camera work. AI can generate an image, but it is the human who determines how a scene should look and in what sequence it should proceed.
Lights! Camera! Prompt!
According to Adil Alimzhanov, Head of Content at Higgsfield, each individual request (prompt) for the model generated about 15 seconds of material. These short clips had to be generated many times, constantly adjusting the prompt to get the best version. For the first 25 minutes of the film, 16,181 preliminary video generations were needed, from which 253 final shots were ultimately selected.
A particular challenge in creating a feature-length AI film was maintaining visual consistency. AI models can be unpredictable, and in a major film, scenes cannot be allowed to look different in each new shot.
That is why each prompt had to be extremely long and detailed. All parameters were specified: image style, lighting, lens type, camera placement. In addition, the artificial intelligence was reminded of the need to adhere to the laws of physics: "maintaining gravity and inertia — mass must have real weight, correct contact shadows, absence of props floating in the air."
As a result, on average, each individual prompt consisted of 3,000 words. Higgsfield even created a separate tool that helps generate such gigantic prompts. A user can upload a script page, after which the system will generate a prompt of several thousand words, optimized for creating professional-quality video.
Despite the speed of work, the process turned out to be very expensive and labor-intensive. According to the creators, they had to discard hundreds of generated clips due to minor defects: incorrect eye movement of the character, unnatural facial expressions, or an unsuccessful camera movement.
As a result, the company's bill for AI computing power amounted to $400,000. To prevent costs from rising even further, the creators worked with "non-cloud" providers such as Nebius and CoreWeave, avoiding the services of major tech giants. Such services primarily rent powerful graphics cards and computing resources for neural networks and often offer more favorable terms for AI projects.
As the producing company notes, their project demonstrated that even in the age of artificial intelligence, filmmaking remains a complex craft. It's impossible to simply ask a neural network to "make a cool 95-minute film" — and get a finished movie.
Comments