As a commercialization and extension of content creation, the application of multi-modal models such as AI-generated video is expected to play a huge role in the field of IP content generation across categories, expanding the incremental space for IP monetization and accelerating the commercialization of IP.
The Zhitong Finance App learned that China Merchants Securities released a research report saying that many global tech giants have laid out the video generation field, and generative artificial intelligence such as Sora and Kimi is developing rapidly, empowering the content generation industry. As a commercialization and extension of content creation, the application of multi-modal models such as AI-generated video is expected to play a huge role in the field of IP content generation across categories, expanding incremental IP monetization space and accelerating IP commercialization.
The main views of China Merchants Securities are as follows:
Generative AI is developing rapidly, and generative AI such as Sora and Kimi have made great progress
With the advent of ChatGPT, image generation AI represented by Midjourney and video generation AI represented by Pika, Runway, and Sora developed rapidly and are being used in marketing, film and television, advertising, etc. Among them, Sora is a large video generation model based on the DiffusionTransformer model based on Patches' visual feature tags. It can generate videos and images of different lengths, aspect ratios, and resolutions. It can produce high-definition video of up to 60 seconds, with precise scene layout, vivid characters, and complex lens scheduling. Compared to other products on the market, Sora combines the Transformer and Diffusion models to achieve disruptive progress in terms of time, resolution, and realism.
The application of multi-modal models such as AI-generated video is expected to play a huge role in the field of cross-category IP content generation, expand the incremental space for IP monetization, and accelerate the commercialization of the content generation industry. On the domestic side, GM Big Model Company Step Star launched its product Step-1V 100 billion parameter multi-modal large model, and the KIMI released by the domestic big model manufacturer MoonShotAI (MoonShotAI) attracted widespread attention in the market with its long text processing capabilities.
Film and television: AI empowers the entire industry chain, innovates and increases efficiency to accelerate IP monetization
On the upstream side, AI can help with script creation and data analysis to discover popular screenplays. On the midstream side, generative artificial intelligence such as Sora has developed rapidly, further lowering the threshold for film and television creation, increasing the realism of virtual characters, improving film and television special effects, improving creative efficiency, speeding up the iteration of film and television content, and speeding up the implementation of classic film and television IP. For film and television production companies, the positioning of generative artificial intelligence such as Sora is still an auxiliary tool, and companies with quality IP resources are expected to benefit more in the medium to long term.
Gaming: The IP market has great potential, and AI helps reduce costs and increase efficiency
From January to September '23, China's mobile game IP market revenue accounted for 77.7% of the mobile game market, reaching 132.06 billion yuan. In '22, game revenue accounted for 37.2% of the total revenue of the entertainment IP market, making it an important source of revenue. AIGC can be applied to various aspects of the game field, including early design and testing, mid-term experience optimization and operation optimization, and derivative fields such as game peripherals and game live streaming, etc., so that creators' ideas can be implemented quickly, game production and IP monetization cycles are greatly shortened, production costs are significantly reduced, and game quality is improved.
Publication: Digital Reading Leads Industry Recovery, AI Helps Monetize High Quality Corpus
In the book retail market in 2023, the year-on-year growth rate of the market changed from negative to positive, and the growth trend resumed. The market volume reached 91.2 billion yuan, an increase of 4.7% over the previous year. The total revenue scale of the digital reading market reached 46.352 billion yuan, an increase of 11.50% over the previous year. The number of digital reading users reached 530 million, an increase of 4.75% over the previous year. The number of users led to the prosperity and development of the digital reading market. The development of IP derivatives of outstanding works has enriched the income of enterprises in the industry. At the same time, high-quality Internet corpus is scarce, and high-quality content may be exhausted within 3 years as soon as possible. High-quality corpus resources in the publishing industry can be used as training data for large models, and the corresponding copyright payment model is expected to bring new revenue to the publishing industry.
Related targets:
Optical Media (300251.SZ) (AI+ Animation), Shanghai Film (601595.SH) (implementing the Big IP Strategy and launching the “iPAI Planet Project”), Chinese Online (300364.SZ) (high quality IP+AI+ short drama, big models assist short story creation, Wensheng Video reduces production costs), Huace Film and Television (300133.SZ) (IP content library support, short drama and micro-video cooperation to actively explore new AI+ businesses), Baena Qiancheng (300291.SZ) (innovates the entire IP chain) Industry+AI Empowering Master Industry), Tencent Holdings (00700) (ToC Big Model, AI+ Game+Advertising+Fintech Cloud Service), Gomai Culture (301052.SZ) (investing in technology companies to develop AI proofreading big models), CITIC Publishing (300788.SZ) (AIGC Digital Publishing Platform Assisted Book Publishing), New Classics (603096.SH) (Bibi Zoo), Tom Cat (300459.SZ) (collaborates with companies such as OpenAI, Google, and stabilityAI to study multiple emotional models) Model), Mango Supermedia (300413.SZ) (AIGCHUB application platform launched), Giant Network (002558.SZ) (GianGPT completed the filing, space-killing AI gameplay, layout of traditional IP+AI such as “Journey”)
Risk warning: AI technology development falls short of expectations, commercialization is uncertain, and policy regulation risks.