Netflix Just Bought Hollywood's Most Interesting AI Company — And Ben Affleck Built It

How Affleck got here

Ben Affleck is not the person you'd expect to launch an AI company. He's an Oscar-winning filmmaker (Argo, Good Will Hunting), a sometimes-maligned Batman, and a tabloid fixture. But he's also someone who's directed enough films to know exactly where money gets wasted in production. In 2022, Affleck started paying attention to how AI technology was being applied — and misapplied — in the film industry. Most of the tools he saw were either trying to do too much (generate entire scenes from text prompts) or solving problems filmmakers didn't actually have. The tools that would genuinely help day-to-day production work didn't exist yet. [1][2]

So he built them. Affleck partnered with a team of engineers to create a proprietary dataset on a controlled soundstage — essentially a real production environment where they could capture the messy, complicated reality of filmmaking under controlled conditions. They trained their AI model on this data to understand what Affleck calls "visual logic and editorial consistency." [2] What that means in practice: the model knows what a correctly lit scene looks like. It understands that if a character is wearing a blue shirt in one shot, they should be wearing it in the next. It can identify when a background element changes between takes and suggest corrections. These are the kinds of problems that VFX teams spend thousands of hours fixing manually — and that InterPositive's model can flag and often resolve automatically. [2][3] "We also built in restraints to protect creative intent," Affleck wrote in a statement, "so the tools are designed for responsible exploration while keeping creative decisions in the hands of artists." [2]

Why this isn't just another AI acquisition

Hollywood is in the middle of a genuine identity crisis over artificial intelligence. SAG-AFTRA, the actors' union, is currently negotiating a new contract with major studios. New AI provisions are at the center of those talks — specifically, protections against studios using AI to create synthetic versions of actors or generate performances without consent. Just last month, several studios sent legal threats to ByteDance over its Seedance 2.0 platform, which they allege uses copyrighted material to train AI video generation models without permission. [1] In this environment, any company acquiring AI technology for filmmaking is stepping into a minefield. The wrong framing — "AI replaces expensive human labor" — could trigger union backlash, talent boycotts, and regulatory scrutiny.

Netflix and InterPositive appear to have been very deliberate about avoiding those land mines. Everything about InterPositive's positioning emphasizes what the technology does NOT do: it does not create AI actors. It does not generate synthetic performances. It does not write scripts or replace creative judgment. It fixes lighting. It catches continuity errors. It fills in missing shots using footage from the same production. [2][3] Elizabeth Stone, Netflix's chief product and technology officer, framed it carefully: "The InterPositive team is joining Netflix because of our shared belief that innovation should empower storytellers, not replace them." [1][2] That's a deliberately chosen stance, and it's smart. By positioning InterPositive as a post-production efficiency tool rather than a creative replacement technology, Netflix stays on the right side of SAG-AFTRA's red lines while still capturing enormous cost savings.

The real story: Netflix's margins

Strip away the Hollywood drama and this is fundamentally a business efficiency play. Netflix spent over $17 billion on content in 2025. That number includes everything from actor salaries to marketing, but a substantial portion goes to production and post-production costs. VFX work, reshoots, editing, color correction, continuity fixes — these are line items that add up fast, especially across the volume of original content Netflix produces. If InterPositive's tools can reduce post-production costs by even a modest percentage across Netflix's entire content pipeline, the math gets interesting quickly. A 5% reduction in post-production spending across hundreds of productions per year would represent savings in the hundreds of millions. Whatever Netflix paid for InterPositive — and the financial terms were not disclosed — the ROI timeline is likely short. [1][3]

This is also an exclusivity play. By acquiring InterPositive rather than licensing its tools, Netflix ensures that competitors like Disney+, Amazon Prime, and Apple TV+ can't access the same technology. In a streaming landscape where content spending is converging and differentiation is increasingly difficult, having a proprietary cost advantage in production is a legitimate competitive moat.

The SAG-AFTRA question

There's an elephant in the room, and it's wearing a union badge. InterPositive's "we help filmmakers, not replace them" positioning is smart — but it's also untested under real scrutiny. When AI tools start handling work that was previously done by VFX artists, editors, and post-production crews, those are union jobs being displaced regardless of how the press release describes it. The International Alliance of Theatrical Stage Employees (IATSE), which represents post-production workers, hasn't weighed in publicly on the InterPositive acquisition yet. But they will. And the conversation will inevitably shift from "AI that empowers storytellers" to "AI that reduces the number of people on the payroll." Netflix is betting that by keeping Affleck — a SAG-AFTRA member himself — as the face of the technology, and by drawing a clear line between "post-production efficiency" and "synthetic performance generation," it can navigate this tension. It's a reasonable bet. But it's still a bet.

What this means for the rest of Hollywood

Netflix's acquisition of InterPositive sends a clear signal to every other studio and streamer: the AI tools arms race in production is real, and the companies that build or buy proprietary technology early will have structural cost advantages that are hard to replicate. Disney has been experimenting with AI in its theme parks and animation pipeline. Amazon has AWS-based tools for content analysis and recommendation. Apple has deep custom silicon expertise. But none of them have made a move this explicit in the production space. The InterPositive deal also validates a specific theory about AI in creative industries: the most commercially viable applications aren't the flashy ones that generate content from scratch. They're the mundane ones that make existing workflows faster and cheaper. Nobody writes breathless headlines about AI that fixes lighting inconsistencies between takes. But that's where the money is.

Affleck, to his credit, seems to understand this better than most technologists. He didn't try to build the most impressive AI model. He tried to build the most useful one. And then he sold it to the company with the biggest content pipeline in the world. That's either shrewd business instincts or very good timing. Probably both.