Ellison’s $4 Million Terminator
On Paramount Skydance’s earnings call, David Ellison made remarks about the future of moviemaking — and he brought a blockbuster with him.
Asked whether artificial intelligence would be able to make movies, the chief executive answered with a superlative and a sum. AI, he said, has the potential to be the “biggest creative unlock in modern history” — bigger, in his telling, than CGI, the visual-effects revolution that began with Jurassic Park and Terminator 2. And then he reached for Terminator itself to show his working.
The original film, Ellison said, cost $4 million to make. Remade today, with the visual effects a modern version would demand, it would cost close to $100 million. But with AI tools, he said, filmmakers are going to be able to make it for $4 million again.
It was a neat piece of arithmetic — a hypothetical difference of roughly $96 million, set out in a single sentence — and it came in the same call as a reassurance. AI, Ellison said, “will never be a replacement for the people and the artists that make a film.”
Both claims were made on the same call, to the same audience. That is the story. A studio chief predicting that a film costing close to $100 million to remake today could be made for $4 million again with AI tools, while saying AI will never replace the people and the artists who make films. Whether both can hold at once is a question the accessible reporting does not settle.
The math, and the fine print
Start with the numbers. The $4 million figure is real, in a sense: The Terminator was originally budgeted at $4 million in 1984. But that was the starting budget, not the finishing one. The budget was later raised to $6.5 million, and the film’s final cost came to about $6.4 million.
That does not make Ellison’s comparison false — the point was the scale of the gap, not the decimal places. But it is a reminder of how the comparison works: a round number from 1984, a round number from today, and a round number from the future. The $100 million remake figure is Ellison’s own hypothetical, reported by The Verge from the call; one secondary summary of the call put the figure lower, at $60–70 million. The exact wording will have to wait for a full transcript, and the accessible reporting does not settle the proposed budget or the feasibility of the comparison.
What is not in doubt is the direction of travel he was describing. Ellison said AI will significantly reshape Paramount Skydance’s business, that the company has begun implementing the technology across the business, and that he is looking forward to sharing more in the coming months. What he did not say, at least in the portion of the reporting that is publicly accessible, is what any of that means in practice: no specific AI products, tools, timelines, partners, or investment amounts appear in the accessible text of The Verge’s report, which is paywalled partway through.
So the promise arrives without a price tag, a schedule, or a product name. That matters for how the $4 million should be read. It is a prediction about what AI tools will one day make possible, not a report on something Paramount has already done.
Why CGI, why now
The comparison itself is the one Ellison offered: he said AI’s impact on film will be bigger than CGI’s, and that the visual-effects revolution started with Jurassic Park and Terminator 2. The accessible reporting does not say how the two technologies compare in speed of adoption, or in what happened to the people who worked with them — those questions sit outside what the call’s public record establishes.
What the reporting does establish is the context in which the comparison was made: a studio that has just told investors it plans at least $6 billion in annualised cost savings within three years. Whether CGI is a fair precedent for AI — and what the comparison implies for the people who make films — remains an open question.
That savings target was part of the same earnings report. Paramount Skydance reported $7.03 billion in revenue for the third quarter of 2026, up 5 percent year over year, and $73 million in profit — six cents a share, against the 20 cents analysts had expected. Shares fell more than 2 percent in after-hours trading. Paramount+ grew revenue 21 percent to $2.13 billion, with 82.2 million subscribers. The call was the company’s first earnings report since the Warner Bros. Discovery merger closed.
None of that is an AI story on its own. But it is the context in which the remarks were made: a newly merged company, a miss on earnings, and a multi-billion-dollar savings target set out in the same earnings report. The superlative arrived alongside a quarter in which the company missed earnings expectations.
The people in the sentence
Which brings us back to the reassurance. “AI will never be a replacement for the people and the artists that make a film.”
Hollywood has recent reason to worry. In the 2023 strikes, generative AI was a core demand: the Writers Guild contract that ended the dispute establishes that AI is not a writer, bars studios from forcing writers to use it, and requires disclosure when AI-generated material is given to a writer. The actors’ union, SAG-AFTRA, won requirements for informed consent and compensation before a performer’s likeness is scanned or recreated. The people Ellison was reassuring won contractual protections in the 2023 strikes covering AI’s use in writing and performance.
That history is what gives the $4 million figure its edge. If a Terminator can be made for $4 million again, the question is where the savings come from — and the accessible reporting does not answer it. The counterargument is that there may be no contradiction at all: AI could lower the cost of execution while artists keep the creative decisions, and the “never a replacement” line could be the thesis while the $4 million is only the tooling. That reading is consistent with what he said on the call, but it is one the reporting does not confirm, because it does not say which costs would fall.
What we still don’t know
The accessible portion of the reporting leaves several questions open. What AI tools or workflows has Paramount actually implemented, and where in production? Does the $4 million figure assume AI-generated effects at current quality, or at some future quality? Who bears the cost of the artists Ellison says are irreplaceable, in a film made that cheaply?
Ellison said he is looking forward to sharing more in the coming months. Until he does, what the accessible reporting establishes is a prediction made on an earnings call, not a demonstration — a vivid one, a confident one, and one that asks Hollywood to believe two things at once: that a film costing close to $100 million to remake today could be made for $4 million again with AI tools, and that AI will never replace the people and the artists who make films. Both may be true. The accessible reporting does not settle either.
