Pacific Securities: AI Video Model Accelerates Iteration, Tool and IP Value Highlighted
AI video tools serve as a carrier, highlighting the core value of the industry chain.
Pacific Securities released a research report stating that since 2025, domestic and foreign video models have accelerated their performance iterations, meeting the ability to produce L3 short film-level content, driving the global film and television industry into the era of widespread AI adoption. Currently, the penetration rate of AI in the film and television industry is still in the single digits, but with further iterations of models and video tools, the industry's penetration rate is expected to experience explosive growth. AI video tools, as carriers, highlight the core value of the industry chain. At the same time, IP companies are expected to benefit fully from this wave, realizing a revaluation of the value of content assets.
The main points of view of Pacific Securities are as follows:
Models accelerate iterations, L3 short film-level capabilities can already meet some film and television production needs.
In terms of models, overseas video models have taken the lead in breakthroughs in cutting-edge capabilities such as physical simulation and fidelity, while domestic models continue to catch up with leading overseas models. Specific to video models, the native resolution can reach up to quasi-2K, and the longest single-generation duration can reach 25 seconds. They also support synchronous output of audio and visual elements, improving the controllability of generated images and storytelling. They can more accurately simulate the physical laws of the real world and human movements and expressions, indirectly reducing costs through the release of lightweight models and improving quality and stability. In terms of content generation, mainstream models currently support L3 short film-level content creation, with some capabilities meeting the requirements of L4 feature film-level content, to a certain extent, meeting the production needs of the film and television industry.
The film and television industry is entering a period of widespread AI adoption, with significant room for improvement in penetration rates.
Leveraging the rapid iteration of video models, AI video tools are gradually empowering film and television production: 1) Animation dramas: AI application in content production accounts for 50%-80%, driving an explosive growth in animated dramas, with over 70% of AI animated dramas. 2) Live-action short dramas: From "AI + live-action" to full AI production, AI-simulated human dramas quickly increase in volume, with the top work "Zhanxiantai" reaching over a billion views in just 6 days of its release. 3) Movies and TV dramas: AI-assisted production is still the main trend. AI animated films have already been launched, while live-action films are still in the early stages. The institution believes that AI significantly reduces costs in the filming stage of film and television materials, while enhancing efficiency by compressing the content production cycle and enriching content themes and forms. Looking ahead, the global video production market is approximately $316.3 billion, while the current global AI video market size is about $4.5 billion, with a penetration rate of only 1.4%, indicating a relatively low level.
Video tools are core, with IP segments benefiting the most.
AI video tools are carriers that convert model capabilities into actual productivity. They rely on both model technology support and professional video content production practice, forming a pattern of coordinated development among video models, IPs and film and television, and third-party tool companies. In the future, AI video tool companies with leading technological advantages are more likely to rely on their creative capabilities and platform ecosystems to produce high-quality video content. Meanwhile, IP companies, with a massive resource of content that can be turned into videos, are expected to fully benefit from the maturity of AI video tools.
Risk warning
The risks of AI technology development falling short of expectations, increased content regulation, and intensified industry competition.
Related Articles

Q TECH (01478) announces annual performance, with shareholders' net profit of 1.494 billion yuan, an increase of 435.22% year-on-year.

CHINA LILANG (01234) will distribute a special dividend of 0.03 Hong Kong dollars per share on May 22nd.

CHINA LILANG (01234) will distribute a final dividend of HK$0.13 per share on May 22nd.
Q TECH (01478) announces annual performance, with shareholders' net profit of 1.494 billion yuan, an increase of 435.22% year-on-year.

CHINA LILANG (01234) will distribute a special dividend of 0.03 Hong Kong dollars per share on May 22nd.

CHINA LILANG (01234) will distribute a final dividend of HK$0.13 per share on May 22nd.






