China building new AI-powered film industry

Summary

Cities in China are deploying compute subsidies, rent exemptions, and special funds to build an AI film supply chain. In the first half of 2026, the per-minute production cost of AI-assisted short dramas dropped from approximately $747 to a few hundred yuan. This move mirrors the industrial policy strategy China used to grow its electric vehicle and solar energy sectors. The underlying video generation engines include ByteDance's Seedance 2.0 and Kuaishou's Kling.

Why it matters

If China successfully industrializes AI-generated cinema at this cost level, it could flood global streaming and theater markets with cheap, high-volume content. This could squeeze the profitability of Hollywood and independent studios. The article also highlights the risk of overproduction. On Douyin, 221,900 new AI-generated pieces of content were published in six months, but only 1,055 reached 100 million views. This indicates that the subsidy model may create a glut rather than a sustainable creative economy. For the broader AI sector, this is a concrete example of state-directed capital reshaping a creative sector within a single year.

Highlights

• In the first half of 2026, the per-minute cost of AI-assisted dramas dropped from approximately 5,000 yuan to a few hundred yuan. A wedding scene that cost around 60,000 yuan using traditional methods now costs about 1,400 yuan.
• Beijing launched a 260 million yuan audio-visual technology fund. Shanghai, Shenzhen, and iQIYI are layering on incentives for compute, rent, and visual effects.
• 221,900 new AI-generated pieces of content were uploaded to Douyin in six months, but only 1,055 exceeded 100 million views. This signals an impending overproduction glut.
• 'Sanxingdui: Future Memories' is poised to become the first major studio-produced AI film licensed for release in Chinese cinemas.
• While actors continue to voice concerns about unauthorized use, copyright and likeness protection frameworks remain lacking.

Critical take

The article relies on a single Reuters/CCTV data stream and presents the cost reduction as a linear success story. However, the 221,900 vs. 1,055 view ratio cited in the text undermines the 'new industry' narrative. Massive production yielding almost no audience engagement is a sign of a subsidy bubble, not a competitive sector. Furthermore, the text does not independently verify the claim of being the 'first AI cinema film' or provide quality comparisons. This leads readers to accept that volume and state support are sufficient substitutes for creative sustainability.