东方证券:视频生成进入精准控制时代 创作平权带动BC两端加速渗透
Zhi Tong Cai Jing·2026-02-09 02:24

Core Viewpoint - The report from Dongfang Securities emphasizes the importance of vertical multi-modal AI application opportunities, highlighting that technological breakthroughs and cost optimization will accelerate industry trends, leading to user growth, increased payment penetration, and enhanced commercialization [1] Group 1: Industry Trends - Since the beginning of the year, the domestic multi-modal video generation sector has seen accelerated iteration of models, significantly narrowing the technological gap with overseas counterparts [2] - The most notable change is the introduction of intelligent storyboarding, which lowers the entry barrier for users, while a unified multi-modal architecture enhances the efficiency and flexibility of creative intent expression [2] - The firm predicts substantial progress in both B-end and C-end expansions by 2026, with a focus on observing AI penetration in the content sector [2] Group 2: Technological Advancements - The acceleration of model development among domestic video generation companies has led to significant improvements in foundational attributes such as physical realism, motion fluidity, and instruction adherence [3] - Recent model releases have enhanced capabilities in storyboard functions and audio-visual synchronization, addressing previous gaps in functionality [3] - The competition in the video generation sector is now akin to the state of large language models (LLMs) in 2025, with companies achieving high baseline capabilities, suggesting that future differentiation will depend on specific application scenarios [3] Group 3: User Accessibility - The video generation sector has transitioned to a "dashboard era" characterized by precision and control, with recent models supporting multi-modal input architectures [4] - The generation duration has become more user-friendly, increasing to approximately 15 seconds per generation, which further lowers the creative barrier for both B-end and C-end users [4] - The models now allow for detailed editing of generated content, facilitating quick adjustments and enhancing the efficiency of creative expression [4] Group 4: Investment Recommendations - Relevant investment targets include Alphabet Inc. (GOOGL.US), Kuaishou-W (01024), MINIMAX-WP (00100), and Meitu Inc. (01357) [4]

Orient Securities-东方证券:视频生成进入精准控制时代 创作平权带动BC两端加速渗透 - Reportify