Hollywood Stopped Arguing About AI. Now It Is Rebuilding the Studio Around It.

The numbers arrived without much fanfare, which is how you know things have changed. At the end of September, reporting around Bloomberg's Screentime conference in Los Angeles revealed that Netflix used generative AI in roughly 300 of the 1,000 titles it released this year. Not three. Not thirty. Three hundred. When the world's biggest buyer of screen content touches nearly a third of its slate with AI, the technology has stopped being an experiment and started being infrastructure.

The confession landed in a week when Hollywood was already in an unusually reflective mood. Bloomberg's annual Screentime event, held September 30, was dominated by two subjects: Paramount Skydance's pending acquisition of Warner Bros. Discovery, and the industry's embrace of artificial intelligence. Kathleen Kennedy, the former Lucasfilm president who stepped down in January after 14 years, used the stage to warn that both forces threaten Hollywood's economy. "To have a studio disappear and fewer movies get made, fewer content gets made," she said. Consolidation and automation, arriving together, is the plot twist nobody wrote.

Netflix built the machine

The clearest signal of intent came back in March, when Netflix acquired InterPositive, the AI filmmaking technology company founded by Ben Affleck in 2022. A later regulatory filing disclosed that Netflix paid approximately $587 million in cash for an acquisition announced in March (the filing did not name the company, but Netflix had announced the InterPositive deal that month). The 16-person team of engineers, researchers and creatives was folded into Netflix, with Affleck joining as a senior advisor.

What InterPositive builds is deliberately different from the text-to-video generators grabbing headlines. Its tools work directly with existing production footage: replacing missing shots, reframing existing ones, correcting lighting, and replacing or enhancing backgrounds. The model is trained only on dailies from the specific production it is working on, which keeps its output relevant and fences off the copyright exposure that comes with training on other people's work. It slots in alongside Eyeline, Netflix's in-house visual effects research group, and a dedicated animation lab. Co-CEO Greg Peters told analysts the three units are now "working in concert to drive production speed."

The financial logic is not subtle. Netflix reported second-quarter revenue of $12.56 billion, up 13 percent from a year earlier, with net income of $3.4 billion, and tightened its full-year 2026 revenue guidance to $51.0 billion to $51.4 billion. Co-CEO Ted Sarandos moved preemptively to address the labor concerns that have shadowed AI adoption since the 2023 strikes: "Great stories require great artists, and AI doesn't change that," he said, insisting that cost savings from AI are more likely to be reinvested into content than pocketed as margin. The market heard the AI part loud and clear either way.

Every studio is spending

Netflix is not alone; it is just the furthest along. In June, Lionsgate took a stake in Runway, the AI video company, to develop creative projects together, deepening a partnership that began in 2024. The same month, A24 struck a research partnership with Google DeepMind alongside a reported $75 million investment, the first time Google took a financial stake in a Hollywood studio, focused on storyboarding and production workflow tools. Amazon MGM Studios launched Project Nara, its own AI production platform built on AWS, with a GenAI Creators' Fund that greenlit three animated series for Prime Video, including Punky Duck from Jorge R. Gutierrez. Amazon MGM and Lionsgate have both appointed dedicated AI chiefs. Google has also built a foothold in entertainment through 100 Zeros, its initiative with Range Media Partners.

The debate about whether AI belongs in filmmaking is over. The debate now is who owns the models, who trains them on what, and who gets paid.

The fight over consent

The money is moving faster than the rules. Three years of labor disputes over AI have reshaped the industry's contracts: SAG-AFTRA's 118-day strike in 2023 established foundational protections, and a strike by video game performers that ended in July 2025 produced new rules on consent and pay for digital replicas, the AI copies of a performer's face or voice. AI protections are now a standard part of industry bargaining, and those contracts are coming up for renewal, which sets up fresh pressure over pay and consent just as studios deepen their AI partnerships.

The studios know it, which is why every announcement comes wrapped in the same language. Amazon MGM says "human creativity leads, and AI supports." Affleck said he built InterPositive with restraints to protect creative intent, "so the tools are designed for responsible exploration while keeping creative decisions in the hands of artists." He even used the startup's tools on his own new film, Animals, training models only on proprietary assets. That framing is partly principle and partly strategy: after 2023, nobody in Hollywood wants to be the executive who described a machine as a replacement for a guild member. And the parallel story in music streaming shows how fast synthetic content becomes a fraud and consent problem when the guardrails are missing.

The cheaper rival across the Pacific

The competition is no longer just between studios and American startups like Runway and Luma. Chinese platforms, including ByteDance's Seedance, Kuaishou's Kling and Alibaba's HappyHorse, are closing the gap on realistic, cinema-quality video generation and undercutting their US rivals on cost. Hollywood spent two years arguing about which American lab to partner with; the real price war may arrive from Beijing and Shenzhen before the guilds finish their next round of negotiations.

What this means for the next decade

Step back and the shape of the next era is visible. Studios that own their AI stacks control their cost structure and their data; studios that do not will rent both. Netflix paid $587 million for 16 people and a toolkit because it believes production speed is now a moat. Lionsgate, A24 and Amazon MGM are making the same bet in their own ways, through equity stakes, research partnerships and proprietary platforms.

The audience may never notice, and that is the point. AI in film is not going to arrive as a genre of obviously synthetic movies. It is arriving as an invisible layer inside ordinary production: the relit scene, the fixed background, the missing shot nobody had to reshoot. The credits will keep getting longer. The crews may keep getting smaller. Hollywood's AI era is not coming. Looking at the 2026 slate, it is already here.