For the first time, Netflix has publicly disclosed how much of its content has already involved generative artificial intelligence. And the number is higher than most people expected. In its second-quarter results for 2026, the company acknowledged that AI had assisted with roughly three hundred titles since the beginning of the year. Most of the work took place in post-production, the stage after filming when editing, effects, and image fine-tuning come into play.
Yet just last summer, Netflix was talking about a single series. It was the Argentine sci-fi series The Eternaut, where the team used AI to create a scene of a building collapsing in Buenos Aires after being hit by toxic snow. Company co-CEO Ted Sarandos told investors at the time that the sequence was created ten times faster than with traditional visual effects methods. In the space of a year, one experiment has therefore become a routine part of post-production.
Enhancing films in half the time
Netflix named three live-action productions: the Indian crime thriller Glory, the Brazilian documentary Brasil 70: A Saga do Tri, and the American documentary series The American Experiment. In all three cases, the tools were used to create things that would otherwise have cost enormous amounts of money: expanded crowds, historical battle scenes, and world-building shots.
Sarandos spoke most extensively about the American series. The American Experiment contains seventeen minutes of footage enhanced by artificial intelligence. It was created twice as fast and at half the cost compared with the options previously available. “It allowed us to expand the scope of the series in a way that would not have been feasible before,” he said during an earnings call with investors.
Sarandos added a point he has been making for a long time. Without these scenes, many productions would simply have gone without them because they could not afford them. “In many cases, creators would have omitted those key shots because they simply would not have had the money for them,” he explained. “They would not have been able to complete them within the timeframes they work with, either.”
AI is used throughout production
The company explained that its use of generative AI is not limited to effects. It deploys it across the entire production process, from the initial idea and visualization through post-production to the final version. Creators use it for scene reference materials, shot sketches, filming plans, and sequence preparation. Sarandos summarized this by saying that it allows the entire production process to run smoothly, faster, and more efficiently.
However, Netflix did not say exactly what it considers “working with generative AI,” nor how extensively the tools were used on individual projects. Across three hundred titles, this could mean anything from a single enhanced shot to a major overhaul of entire sections. It is also interesting that not a single animated production was among the works named.
That is a sensitive issue. In May, Netflix was recruiting for positions at its INKubator studio, which it described in job postings as an animation workshop built around generative AI. After the report was published, the company clarified that it views INKubator as an artist-led incubator for developing stories from its existing franchises. At the same time, it emphasized that films from Netflix Animation Studios would continue to be created using traditional animation methods. However, the figure of three hundred titles will cast doubt on such assurances among some observers. Analysts are questioning this approach following the disclosure that three hundred titles had used AI.
Artificial intelligence is a tool, not a replacement
During the call, Sarandos pushed back against the idea that AI should replace filmmakers. “It takes great artists to create something great, and AI does not change that,” he said. “Films are made by the people who make films. AI gives them better tools to make those films even better.” He made similar comments to Politico in March, saying that faster and cheaper is pointless unless it is also better.
Netflix did not buy some of this technology outright, but obtained it through an acquisition. In March 2026, it acquired InterPositive, a startup co-founded by actor and director Ben Affleck, for up to $600 million. The deal brought a sixteen-member team into the company, while Affleck himself took on an advisory role focused on developing AI tools tailored to filmmakers. Sarandos acknowledged that the acquisition is still in its early stages, but its initial impact on productions is already visible. InterPositive is only one piece of the puzzle. Netflix also has its own teams, such as Eyeline and its animation lab.
The market reacted sourly
The AI disclosures came as investors gave the company’s results a fairly lukewarm reception. Revenue rose 13 percent to $12.56 billion, narrowly missing Wall Street estimates. Net income reached $3.4 billion. Shares fell by more than 8 percent after the results were released and extended their losses on Friday morning, mainly due to a weaker outlook for the third quarter.
The decision to publish its viewership report only once a year instead of twice also contributed to the reaction, with some in the market seeing it as a step backward in transparency. Netflix defended the move as an effort to keep investors focused on its main financial indicators, namely revenue and operating profit.
Sarandos reiterated a statement that Netflix has made for years. According to him, savings from AI will be reinvested in creating new content. “The cost of production will go down, quality will go up, so we will probably reinvest the savings into more content for the service,” he said. AI at Netflix began as a peripheral experiment, but it has become a firmly established part of title production.
Sources: theverge.com and variety.com
