The Future of AI · Industry by Industry
The First AI-Made Movie Megahit: Matt Says 30% by 2030. I Say 11%.
A feature film whose moving images are predominantly produced by generative AI — rather than by live-action photography or a conventional animation pipeline — will gross more than $200,000,000 in worldwide theatrical box office on or before December 31, 2030.
The technology is close to solved and the money bar is not. $200 million worldwide would have been roughly the fourth-biggest film of 2026. The binding constraint is a distributor willing to spend real marketing money on a film audiences may actively resist, and then a mass audience showing up anyway. Four years is not long for both.
Call added September 23, 2026, after publication.
Three things push this faster than FRED thinks. Audiences will find the first one cool and novel, and novelty sells tickets. Hollywood will fight it like crazy, which is real resistance but also free publicity. And the next generation is growing up comfortable with AI creativity in a way the current one is not.
19 points apart. A meaningful split.
Resolves HIT if a single feature film reaches a cumulative worldwide theatrical gross above $200,000,000 as reported by Box Office Mojo or Comscore on or before 2030-12-31, AND its moving images are predominantly generated by AI models rather than captured by camera or built in a conventional animation pipeline, with that production method either stated by the distributor or producer or established by credible trade reporting in Variety, The Hollywood Reporter, Screen Daily or Deadline. Resolves PARTIAL if such a film reaches $100,000,000 but not $200,000,000 by the deadline. Resolves MISS in every other case. The following do NOT qualify at any level: AI visual effects, AI de-aging or digital doubles inside a conventionally photographed film; AI-written or AI-edited scripts; AI dubbing, translation or voice cloning; AI used for storyboards, previsualisation or marketing. Streaming, premium video on demand and home video revenue do not count. Figures are nominal US dollars, not adjusted for inflation.
Written before the outcome. Not reinterpreted after.Tom Cruise told a BAFTA audience in London this week that AI in film “is coming and it’s going to happen.” He also said people want to see real things. Both can be true, and the interesting question is what happens at the box office.
So we put a number on it.
The claim, and why the wording matters
A feature film whose moving images are made by generative AI grosses more than $200 million worldwide, on or before December 31, 2030.
That wording does a lot of work, and it has to.
AI already touches most big films. Visual effects, de-aging, cleanup, previsualisation, marketing. If any of that counted, this prediction would already be settled and it would have measured nothing at all.
So the claim requires the images themselves to be generated rather than filmed or conventionally animated. No AI scripts, no AI editing, no AI dubbing, no digital doubles inside a normal shoot. Box office only, so streaming does not rescue it.
A forecast that can be won by choosing a definition afterward is not a forecast. This is the same trap we found in Elon Musk’s claim about doubling the global economy, where the whole bet turned on a word he never said.
Where this actually stands today
The technology is much further along than most people realise.
- Hell Grind — Higgsfield unveiled it at a Cannes industry event in May 2026. Ninety-five minutes, produced entirely on its platform, no live-action footage. The company later open-sourced the prompts and assets.
- The Cully Hill Boys — 110 minutes, built in four weeks by a team of 28, screened to a private room in New York in August.
- Sanxingdui: Future Memories — Bona Film Group won approval in April 2026 to put a 90-minute AI-generated science fiction feature into Chinese cinemas, made with ByteDance video tools. Regulators filed it as animation because there is no category for AI film yet.
Feature-length AI films exist. One has real theatrical approval in a major market.
What none of them has is money.
No AI feature has had a wide release with a real marketing campaign, and none has reported a gross near blockbuster scale. For scale: a Chinese independent film made for about $3,000 crossed roughly $2.1 million this year and out-earned essentially every AI-generated movie released in 2026.
The distance from there to $200 million is about a hundredfold.
Why 11%
I priced this originally at 16% with a 2032 deadline. Matt chose 2030, which is a harder bar, so the honest re-price at his date is lower. Three gates:
Does generative video hold together at theatrical quality for a full feature? 85%. Mostly solved. Ninety-five minutes already exists. What remains is the difference between a film critics describe as impressive for AI and one an audience forgets is AI.
Does a distributor give one a genuine wide release with real marketing spend? 50%. This is where I part company with the optimists. A studio has to commit tens of millions in marketing to a film that a meaningful share of the audience, and most of the industry’s talent, will publicly oppose. China is the more likely path than Hollywood, and it already has regulatory approval.
Given that release, does it clear $200 million? 25%. Most wide releases never get near $200 million. Novelty cuts both ways — it draws curiosity and it invites a boycott. And a film has to be genuinely good, which no AI feature has yet demonstrated at length.
85% × 50% × 25% = roughly 11%.
Where Matt lands, and why it matters
Matt is at 30% on the same date. Nearly three times my number.
His reasoning is three factors, and the second one is the interesting one:
Novelty sells. The first AI film that looks genuinely good will be an event, and people will buy tickets to see what it is. Curiosity is a real box office driver and I may be underweighting it.
Hollywood will fight it like crazy. He counts this as resistance and as fuel. A public fight with the industry is free marketing that no studio could buy, and controversy has sold a great many tickets.
The next generation is growing up with it. This is the strongest part of his case. The discomfort is generational, and the audience that decides what a megahit looks like in 2030 is not the audience that decides today.
I think the money bar binds before any of that matters. He thinks the audience arrives faster than the industry can hold the line.
One of us is going to be wrong in public, on a date, and the page will say so.
The annual scoreboard
We are not waiting until 2030.
Each year this page records the highest worldwide theatrical gross achieved by any qualifying AI-generated feature, along with whether any major distributor has committed to a wide release.
| Year | Best AI film worldwide gross | Wide release committed? |
|---|---|---|
| 2026 | Low single-digit millions | No |
| 2027 | — | — |
| 2028 | — | — |
| 2029 | — | — |
| 2030 | — | — |
That first column is the whole forecast. A jump from a few million to nine figures does not happen quietly. If the number is still in the millions at the end of 2028, Matt’s 30% was too high. If a studio commits to a wide release in 2027, mine was too low.
What would change my mind
- A major studio or a top-three Chinese distributor announcing a wide theatrical release for an AI-generated feature
- Any AI feature clearing $25 million worldwide, which would be a tenfold jump from today
- A generative film winning a significant festival prize in open competition against conventional films
- Audience polling showing the “made with AI” label stops depressing intent to buy a ticket
Two of those together and 11% is too low.