back to top
HomeTechThe $500K AI Film That 'Premiered at Cannes' Didn't Actually Premiere at...

The $500K AI Film That ‘Premiered at Cannes’ Didn’t Actually Premiere at Cannes

- Advertisement -

Last week an AI startup called Higgsfield announced it had premiered a fully AI-generated feature film at Cannes. The Wall Street Journal covered it. The founder posted on LinkedIn that “for decades, Cannes has been the room where new cinema gets legitimized.” The story spread fast.

There was one problem. Cannes said it never happened.

According to Futurism, which reached out to festival organizers directly after failing to find the film on the official Cannes schedule, a festival spokesperson confirmed that “Hell Grind was not screened as part of the official Festival de Cannes program.” The film was presented during an industry event organized by third parties in Cannes. That’s a meaningfully different thing and the distinction matters because the entire credibility of the announcement rested on the Cannes name.

It’s a clean example of how AI hype gets manufactured and how quickly it travels before anyone checks.

What Hell Grind actually is

Higgsfield, a San Francisco startup valued at $1.3 billion, made a 95-minute action film called Hell Grind in two weeks using AI video generation tools including Google’s Veo 3. Total cost was $500,000. Of that, $400,000 went to compute costs, which tells you something about where AI filmmaking economics currently sit.

The film follows four street thieves whose heist goes wrong when an ancient artifact pulls one of them into the underworld. It’s campy, action-heavy, and exactly the kind of spectacle you’d expect from a proof-of-concept designed to sell Hollywood studios on AI video tools rather than win awards.

The technical process was more involved than the “just prompt an AI” framing suggests. Each prompt averaged 3,000 words. Every generation produced about 15 seconds of footage which then had to be generated multiple times with tweaks to get a usable shot. The first 25 minutes of the film required 16,181 initial video generations that became 253 final shots. Maintaining visual consistency across a feature-length film is genuinely hard with current AI tools and the team had to build detailed style prefixes into every prompt defining lighting, camera type, physics behavior, and more to avoid the over-lit artificial look that gets dismissed as slop.

“You can’t go into AI and say make me a 95-minute cool video,” said Adil Alimzhanov, a content lead at Higgsfield. That’s honest and worth crediting. The work was there even if the marketing around it wasn’t.

How the Cannes claim fell apart

The venue where Hell Grind screened was the Marché du Film, which has a business relationship with the Cannes Film Festival but operates as a separate commercial marketplace with no meaningful selection process. It will screen any film that pays the fee. It has screened Sharknado. Calling it a Cannes premiere is roughly equivalent to buying an ad in the New York Times and describing yourself as a Times journalist.

When Higgsfield’s founder posted on LinkedIn implying the film had premiered at Cannes proper, a director named John Washburn replied directly. “This isn’t screening at the Festival de Cannes, which is what you’re implying. The suggestion that paying for a screening at some random theatre in the same town and at the same time as a major festival is somehow the same thing as being selected by that festival is misleading at best. Spurious bullshittery, really.”

Higgsfield later defended itself by saying the MarchĂ© du Film was an accredited component of the Cannes ecosystem. That’s technically true in the same way that a hotel gift shop is part of the hotel ecosystem. The Wall Street Journal’s original article made no mention of the MarchĂ© du Film and left readers with the clear impression the film was part of the festival proper. The paper later added a correction at the bottom of its original story clarifying the film screened at the MarchĂ© du Film, not the official festival program.

The pattern this fits into

Hell Grind isn’t an isolated case. It’s a chapter in a longer story about how AI companies build credibility through association.

Earlier this year a video of Tom Cruise and Brad Pitt fighting on a rooftop went viral for its apparent AI-generated quality, with people declaring it proof that Hollywood was finished. It turned out to be an AI reskin of existing footage of two human performers in front of a green screen. The underlying capability was real but the demonstration was theater.

The Cannes situation follows the same structure. The underlying capability is real. Higgsfield did make a feature-length AI film and the technical challenges they solved are genuine. But the claim that it premiered at Cannes was designed to attach prestige that the work hadn’t earned through the process that prestige is supposed to represent.

Demi Moore said at the festival that AI is here and fighting it is a battle that will be lost. Tilda Swinton said AI doesn’t have a chance. Guillermo Del Toro said something considerably more direct to thunderous applause. The debate at Cannes was real and substantive. Hell Grind was adjacent to it geographically and nowhere near it in terms of what the festival actually validated.

The $500K budget and two-week production timeline are striking numbers if they’re accurate. Given the marketing around the screening it’s reasonable to hold them loosely until independently verified.


Don’t miss any Tech Story

Subscribe To Firethering NewsLetter

You Can Unsubscribe Anytime! Read more in our privacy policy

LEAVE A REPLY

Please enter your comment!
Please enter your name here

YOU MAY ALSO LIKE
Anthropic Secretly Tracked Claude Code Users. Then Called It an Experiment

Anthropic Secretly Tracked Claude Code Users. Then Called It an “Experiment.”

0
There's a version of this story where Anthropic was trying to protect itself from large-scale model theft. There's another where one of the AI industry's biggest privacy advocates quietly crossed a line its own users never expected. What makes this headline important isn't just that hidden tracking code existed. It's that the company behind it was Anthropic. Just months ago, Anthropic publicly refused to let the Trump administration use Claude to surveil American users. The company defended that position in court, arguing that AI companies shouldn't become tools for government surveillance. That stance became part of Anthropic's identity. Then came a very different decision. In March, Anthropic quietly added hidden tracking markers to Claude Code that flagged users' timezones, proxy connections, and potential ties to Chinese AI labs. The code remained unnoticed until security researcher Thereallo discovered it last week. After the discovery went public, an Anthropic engineer confirmed it on X, described it as an "experiment" intended to combat account abuse and model distillation, and said the company had already planned to remove it. The tracker was taken down shortly afterward. The bigger question isn't whether Anthropic had a reason. It's whether a company that built its reputation on privacy can afford to hide surveillance from the very developers it asks to trust its tools.
Leanstral AI

Leanstral 1.5: Mistral’s AI Built to Prove Math Ended Up Finding Real Software Bugs

0
Mistral built Leanstral to do something most AI models don't attempt, write formal mathematical proofs that a compiler can verify as correct. Not "pretty sure this is right" correct. Mechanically, provably, no-exceptions correct. That's a narrow use case, and the audience for it is small. What nobody expected was that a model trained on IMO-level math problems and abstract algebra benchmarks would end up running against open-source codebases and finding bugs that testing and fuzzing had both missed. Five of them previously unreported on GitHub.
Open Source AI Coding Agents That Don't Need a Subscription

7 Open Source AI Coding Agents That Don’t Need a Subscription

0
Open almost any "best AI coding tools" list and you'll see the same names: Cursor, GitHub Copilot, Claude Code. They're good tools but they're also closed source and paid. What's changed over the past year isn't the quality of those products, it's how quickly the open-source alternatives have caught up. Some can orchestrate multiple agents, remember your projects across sessions, and automate complex development workflows. Many let you bring your own model, whether that's a local LLM, OpenRouter, OpenAI, GLM-5.2, Ornith, DeepSeek, or something else entirely. More importantly, you're in control. You decide where your code runs, which model powers it, and how your workflow evolves without being locked into a single company's ecosystem. If you've only looked at the paid options, these are the open-source AI coding tools worth knowing about.