iQIYI Proposes Decentralized AI Production Model for Film and TV

iQIYI founder and CEO Gong Yu laid out a "decentralization" thesis for film and TV production in a 36Kr interview published via KrASIA, building on ideas he first raised publicly at the company's February 2026 earnings call. Gong frames AI's impact as a "one-one-two" law: unit content costs fall by roughly one order of magnitude, the number of creators rises by at least one order of magnitude, and the number of finished works rises by at least two orders of magnitude, a dynamic he ties to the Jevons paradox, where efficiency gains increase total resource consumption rather than reduce it. At iQIYI World Conference 2026 in Beijing (April 20-21), the company publicly detailed Nadou Pro, a professional film and TV production platform it had quietly launched in March, according to KrASIA. PR Newswire and ANTARA report the platform now integrates nearly 70 AI agents spanning scriptwriting, directing, art design, filming, editing, and marketing, plus script and scene analysis tools. iQIYI also announced a limited-time 20% subsidy for mid-length dramas and AI-generated content, per KrASIA.

Nadou Pro connects to iQIYI's proprietary Qizhi large model as well as third-party video-generation models, including ByteDance's Dreamina (Jimeng AI), Kuaishou's Kling AI, Shengshu's Vidu, MiniMax's Hailuo AI, and Alibaba's Wan, per KrASIA. Gong told 36Kr that Nadou Pro "currently carries almost no commercialization target" and grew out of AI tooling iQIYI built internally since setting a goal of industrialized film and TV production in 2021; more than 70 existing internal studios have adopted the platform alongside roughly ten new AI-focused studios. PR Newswire and ANTARA report iQIYI intends to open its IP library, talent network, and revenue-sharing mechanisms to outside creators, with Gong quoted saying "iQIYI will transform into a non-centralized social media-alike platform."

iQIYI's choice to orchestrate across five-plus third-party video models rather than commit to one is a notable production-pipeline pattern: it suggests studios building AI video workflows may get more value from model-agnostic orchestration layers than from single-vendor integration, at the cost of added complexity in asset versioning, multimodal consistency across shots, and rights metadata for cleared IP. None of the sources reviewed disclose measured output quality, latency, or cost-per-shot for Nadou Pro's pipeline.

• •Gong told 36Kr he expects "some good works" using Nadou Pro to appear around Q3-Q4 2026; watch for iQIYI's own disclosure of adoption and output volume against that timeline.
• •Licensing terms and scope for the IP library iQIYI is opening to outside creators.
• •Whether AI-generated content reaches iQIYI's top content tier this year, which Gong told 36Kr is "uncertain" but possible.