Will AI kill the Asian entertainment industry, or reinvent it?

AI-generated content is rapidly expanding in Asian film and video: more than 95% of roughly 128,000 Chinese microdramas released in Q1 2026 were AI-made, and some 100-minute dramas reportedly cost less than 20,000 yuan ($2,981) to produce. The shift is widening access and prompting iQiyi to pivot toward AI-generated content, but has also led to reported redundancies and copyright concerns; ByteDance agreed to strengthen safeguards after its Seedance model drew infringement allegations. Audience response remains mixed: Edenstone reported nearly 70,000 active users in three days for a short film, while the human-made film Niu Lai grossed 60 million yuan ($9.1 million) in China.

Analysis

The investable shift is not simply lower production cost; it transfers scarcity from filming capacity to audience attention, distribution, trusted IP and rights clearance. Cheaper content can reduce commissioning/acquisition costs for streamers, but also floods feeds, raises discovery costs and makes engagement—not volume—the binding constraint. If production savings are competed away through lower prices or higher inference and moderation spend, studios may see little margin expansion.

For iQIYI (IQ), the pivot toward AI/social content is strategically understandable but execution risk rises: it must prove retention and monetization, not just content throughput. Treat the reported pivot as an ambition, not evidence of improved unit economics. Alibaba (BABA) has model exposure through HappyHorse, but the article provides no revenue contribution or adoption economics; any read-through to consolidated earnings is premature. In the West, copyright safeguards create a meaningful friction point for global model deployment and could advantage distributors and rights owners able to license recognizable catalogs—while limiting the apparent cost advantage of unlicensed generation.

Near term, the signal is weak for broad media repricing. Over 1–3 months, watch IQ reporting for AI-content engagement, ad yield and subscriber retention, and BABA disclosures for model commercialization. Over 6–18 months, the key structural test is whether audiences accept long-form AI work and whether rights regimes permit scalable use. Contrarian risk: cheap, abundant content may increase the premium on visibly human-authored work and established franchises rather than displace them wholesale.

Key Decisions for Investors

• No broad AI-media trade on this evidence. Keep IQ on a catalyst watch; do not underwrite margin gains from production-cost claims without evidence of improved retention, monetization and total cost per engaged viewer.
• For IQ, a bearish thesis is falsified if subsequent reporting shows sustained improvement in engagement and monetization without deterioration in subscriber retention; reassess after the next disclosures rather than trade the announcement alone.
• Monitor copyright enforcement and licensing terms as a 1–3 month catalyst for global model economics. Stronger restrictions would weaken unlicensed-generation economics and support the relative value of licensed catalogs; broad, enforceable licensing would reduce that headwind.
• Over 6–18 months, track whether AI output gains feature-length audience acceptance. If engagement remains concentrated in short-form content, expect cheaper production to expand supply more than industry profits; if long-form engagement and ad/subscription yield validate, revisit selective exposure to platforms that can distribute and monetize at scale.