Netflix disclosed GenAI workflows were used in roughly 300 titles in Q2.
Why it matters: The filing shows AI is already embedded in production work at a major streamer, not just in recommendations or marketing. For studios and labor groups, the 300-title figure suggests the technology is being used at scale across content pipelines, especially in post-production.
- Netflix said in its July 16 Q2 shareholder letter that GenAI workflows were used in roughly 300 titles.
- The company said the largest concentration of GenAI work was in post-production.
- Netflix said generative AI is also being used for recommendations, conversational discovery and promotional assets.
- Named examples in the filing included The Eternaut and Happy Gilmore 2.
Netflix disclosed in its Q2 2026 shareholder letter that "GenAI workflows have been used in roughly 300 of our titles, with the largest concentration of work in post-production." The filing was part of the company’s July 16 earnings release package.
The company did not identify which titles were counted, how much AI each project used or how many were originals versus licensed. That limits what outsiders can infer about the breadth of the rollout.
Netflix said AI was used for visual effects in the Argentine series The Eternaut and for de-aging effects and scene-building assistance in Happy Gilmore 2. In this context, scene-building assistance refers to helping create or refine the look of a scene, while de-aging effects are used to make actors appear younger on screen.
Netflix also said generative AI is being used to improve recommendations, test conversational discovery and improve promotional assets. Because the disclosure comes from the company itself and not from an outside audit or third-party report, it should be read as a corporate filing rather than independently verified industry data.
By the numbers
- 300 - Approximate number of Netflix titles where GenAI workflows were used in Q2.
- July 16, 2026 - Date Netflix filed the shareholder letter with the SEC.
Yes, but: Netflix did not say how much of the work in those titles was AI-assisted, which titles were included or whether the 300-title count reflects a wide deployment or a narrow set of tasks repeated across projects.