Core Viewpoint - The film and television sector is experiencing a significant resurgence driven by the upcoming 2026 Spring Festival holiday, which features an extensive lineup of films and advancements in AI technology that enhance production efficiency [3][4][5]. Short-term Dynamics - The 2026 Spring Festival will have a record-long holiday of 9 days, leading to a high number of effective screening days and a strong consensus in the market regarding exceeding box office expectations [4][5]. - The AI video generation technology has reached a critical point where it can produce near-movie quality content, significantly transforming the content production process [4][6]. - The recent performance of the film ETF (516620) indicates a shift in institutional investor perception, moving from a speculative rebound to a focus on recovery in the film sector [6][5]. Medium to Long-term Dynamics - The film industry is witnessing a recovery driven by a combination of policy relaxation and technological advancements, marking a departure from previous cycles [7][9]. - The AI technology revolution is fundamentally reshaping the industry's cost structure and growth potential, moving beyond mere efficiency improvements to becoming a core production capability [10][11]. - The demand for quality content remains strong, as evidenced by the recovery in box office revenues and the increasing popularity of live performances and other entertainment forms [9]. Investment Implications - The film ETF (516620) serves as a comprehensive tool for investors, covering the entire industry chain from production to distribution, thus allowing for efficient capital allocation in the recovering sector [12][13]. - The ETF mitigates individual stock risks and smooths out performance volatility by holding a diversified portfolio, making it a more stable investment option in a sector characterized by high dependency on single blockbuster films [13][14]. - The current valuation of the ETF is at historical lows, presenting a favorable risk-reward scenario for investors looking to capitalize on the industry's recovery and long-term growth potential [14].

1． 多模态输入：从"文本描述"到"所见即所得" Seedance 2.0支持文本、图片、视频、音频四种输入模态，最多可组合12个文件（9张图片+3段视频+3段 音频），通过@素材名精准控制生成内容。例如： 这种多模态方式大幅缩小了"想象与输出"的差距，无需复杂的提示词工程，即可实现"所见即所得"的效 果。 炒股就看金麒麟分析师研报，权威，专业，及时，全面，助您挖掘潜力主题机会！ 来源：人机与认知实验室 一、Seedance的基本定位与技术背景 Seedance是字节跳动旗下Seed团队研发的多模态AI视频生成平台，是当前全球领先的AI视频生成技术之 一。其核心定位是通过扩散模型（Diffusion Model），将静态素材（文本、图片、视频、音频）转化为 动态、连贯、有叙事性的视频内容，实现"从0到1"的电影级视频生成。 从技术迭代来看，Seedance经历了快速进化： 二、Seedance 2.0的核心优势：技术突破与应用价值 Seedance 2.0的竞争力主要体现在多模态输入、导演级控制、专业级输出三个方面，彻底降低了视频制 作的门槛，推动内容生产从"专业团队"向"个人创作者"转移。 2． 导演级控制 ...

智东西 作者 | 李水青 编辑 | 心缘 智东西2月12日报道，千呼万唤始出来，今天午间，字节跳动终于正式发布其视频创作模型Seedance 2.0。 ▲Seedance 2.0已正式上线豆包 字节称，Seedance 2.0采用统一的多模态音视频联合生成架构，支持文字、图片、音频、视频四种模态输入，集成了目前业界最全面的多模态内容参考和编 辑能力。该项目的更多的技术测评细节及大量案例Demo也随之公布。 在过去几天里，这款在灰度测试阶段的模型已经火遍全球。在国内，《黑神话·悟空》背后的游戏科学创始人兼CEO、制作人冯骥盛赞其为"地表最强"，并 称"AIGC的童年时代，结束了。"知名导演贾樟柯在昨晚也发微博称："Seedance 2.0确实厉害，我准备用它做个短片。"在各大平台，网友用Seedance 2.0制 作的视频作品已出现井喷之势。 ▲知名导演贾樟柯评论Seedance 2.0（图源：新浪微博） 在国外，a16z的合伙人Justine Moore感叹"AI视频模型的图灵测试已经被攻克了"；就连马斯克今日也跳出来赞叹"事情正快速发生"，意指技术发展太快了； 许多电影制作领域专业人士表示"这个模型让人感到 ...

猫咪暴揍哥斯拉、叶问拳打钢铁侠、孙悟空棒打奥特曼、小学生隔扣詹姆斯……最近，Seedance 2.0相关视频在不少短视频平台刷屏。 字节跳动新一代AI视频生成模型Seedance 2.0上线内测后，凭借多模态创作方式、自带运镜的呈现效果，迅速在全球范围内引发高度关注。海外社交平台 上，不少用户称赞其生成效果媲美好莱坞大片，相关话题迅速冲上热门，多条演示视频播放量破百万。 据模型内测介绍文档显示，Seedance 2.0搭载双分支扩散变换器架构，用户只需输入文字提示词或上传参考图像，即可产出带有完整原生音轨的多镜头视 频内容。 其最具突破性的"单指令生成连贯多场景"能力，能让AI自动解析叙事逻辑，生成环环相扣的镜头序列，在角色、光影、风格与氛围上保持高度统一，实现 近乎"零后期"的成片效果，彻底突破了早期AI视频工具仅能生成碎片化片段的局限，大幅降低了专业视频创作的门槛与成本。 值得注意的是，2月12日，知名导演贾樟柯昨晚发文称：Seedance 2.0确实厉害，我准备用它做个短片。对此，罗永浩表示，"接下来，拍……做一个电影 只需要导演一个人了。" 马斯克谈Seedance 2.0 最近，Seedance 2 ...

字节跳动新一代AI视频生成模型Seedance 2.0上线内测后，凭借多模态创作方式、自带运镜的呈现效果，迅速在全球范围内引发高度关注。海外社交平台 上，不少用户称赞其生成效果媲美好莱坞大片，相关话题迅速冲上热门，多条演示视频播放量破百万。 每经编辑｜陈柯名 最近，Seedance 2.0彻底火了！ 猫咪暴揍哥斯拉、叶问拳打钢铁侠、孙悟空棒打奥特曼、小学生隔扣詹姆斯……最近，Seedance 2.0相关视频在不少短视频平台刷屏。 Seedance 2.0画面。图片来源：锦观新闻 据模型内测介绍文档显示，Seedance 2.0搭载双分支扩散变换器架构，用户只需输入文字提示词或上传参考图像，即可产出带有完整原生音轨的多镜头视 频内容。 其最具突破性的"单指令生成连贯多场景"能力，能让AI自动解析叙事逻辑，生成环环相扣的镜头序列，在角色、光影、风格与氛围上保持高度统一，实现 近乎"零后期"的成片效果，彻底突破了早期AI视频工具仅能生成碎片化片段的局限，大幅降低了专业视频创作的门槛与成本。 Seedance 2.0演示画面 值得注意的是，2月12日，知名导演贾樟柯昨晚发文称：Seedance 2.0确实厉害，我准备 ...

Core Viewpoint - Seedance 2.0, a new video generation model by ByteDance, is rapidly gaining popularity globally, significantly lowering the barriers and costs for professional video creation [4]. Group 1: Product Features - Seedance 2.0 supports four modalities of input: image, video, audio, and text, allowing users to generate multi-shot videos with complete original soundtracks in just 60 seconds [4]. - The model has gained traction not only in China but also on overseas social media platforms, with users creating imaginative content such as "Wukong vs. Marvel heroes" and "pets dunking on NBA stars," with several demonstration videos exceeding one million views [4]. Group 2: Industry Impact - American directors and industry professionals have expressed concerns that Seedance 2.0 could disrupt Hollywood, with one director stating he created a trailer for a never-released film in just 20 minutes for $60, achieving cinematic quality [4]. - The sentiment among tech workers and AI researchers echoes this concern, with repeated assertions that "Hollywood is cooked" [19][20]. Group 3: User Accessibility - Seedance 2.0 is set to be integrated into CapCut, making advanced video creation accessible to a younger audience, specifically mentioning that every 12-year-old in America will soon have this capability on their iPhones [14][15]. Group 4: Platform Adjustments - In response to concerns about potential misuse, the platform's operators announced that during the beta testing phase, they would not support the input of real human images or videos as reference subjects, emphasizing the importance of respecting creative boundaries [26].

Group 1 - ByteDance officially launched the latest video generation model Seedance 2.0, with its AI products Doubao and Jimeng integrating this model for user experience [1] - Users can access Seedance 2.0 through Doubao APP, computer, web version, and Jimeng APP, as well as Jimeng web version [1] - The platform currently does not support uploading real human images or videos as reference material, which has been a point of feedback from users [2] Group 2 - ByteDance clarified that Doubao APP and Jimeng APP allow for real human appearances, requiring users to complete a verification process through recording before generating a digital avatar for AI video creation [2] - The computer and web versions of Doubao and Jimeng explicitly state that uploading real human face materials is not supported [2]

Core Insights - The launch of the Seedance 2.0 video model by ByteDance marks a significant technological breakthrough in the film and media industry, referred to as the "DS moment" [2] - The AI video generation technology is transitioning from "technically feasible" to "commercially viable," with expectations for a surge in applications by 2026 [2][9] Group 1: Market Reaction - Following the introduction of Seedance 2.0, the media and entertainment sector experienced a strong rally, with multiple stocks hitting their daily limit up [3] - On February 10, the cultural media sector saw a broad surge, with stocks like Reading Culture and Light Media achieving 20% gains, while several others reached 10% limits [3] - The East Finance Excellence Growth Fund led the market with a 12.25% return over three days, highlighting the strong performance of media and entertainment-themed funds [3] Group 2: Fund Performance - Various ETFs and actively managed funds in the media sector reported significant returns, with the Guotai CSI Media Theme ETF gaining 11.67% and the Yinghua CSI Media Theme ETF increasing by 11.53% [6] - The East Finance Excellence Growth Fund, with a small scale of 0.15 billion, is noted for its high elasticity and focus on technology sectors, particularly AI [5] Group 3: Industry Outlook - Analysts believe that AI video models are at an early stage, with a golden window for rapid development, suggesting a significant market opportunity as creation barriers are lowered [8] - The cost and efficiency advantages of AI-generated content are expected to grow, with production costs for episodes likely to decrease significantly [9] - The industry is anticipated to experience high growth, particularly in low-positioned film sectors benefiting from AI technology [9]

Core Viewpoint - ByteDance has launched its latest video generation model, Seedance 2.0, which enhances user experience in creating AI-generated videos through its products Doubao and Jimeng [1] Group 1: Product Features - Seedance 2.0 supports four modalities of input: image, video, audio, and text, allowing for a richer expression and more controlled generation of content [1] - Users can create digital avatars by completing a real-person verification process through recording, enabling them to generate AI videos featuring their likeness [1] - The model allows for a more natural and efficient creative process, resembling that of a true "director," as it expands beyond text-based prompts [1] Group 2: User Experience - Doubao App and Jimeng App support real-person appearances, while the desktop and web versions do not currently allow for the upload of real human face materials [1] - The model has gained global attention due to its multi-modal references and precise controllability, following a limited internal testing phase [1]

Core Insights - The launch of ByteDance's Seedance 2.0 has created significant buzz in the tech industry, even catching the attention of Elon Musk, who remarked on the rapid development of the model [1][4] Group 1: Technology Breakthroughs - Seedance 2.0 is defined as an AI video generation model capable of creating movie-quality videos from text or images, marking a shift from simple video generation tools to a creative engine with "director-like" capabilities [3] - Key technological advancements include: - Self-storyboarding and camera movement planning based on user descriptions, achieving professional-level cinematography [3] - Support for multi-modal inputs, allowing up to 9 images, 3 videos, and 3 audio clips to create precise replicas of actions, effects, styles, and sounds [3] - Native audio-visual synchronization, generating matching sound effects and music while ensuring lip-sync and emotional alignment [3] - Multi-shot narrative consistency, enabling the generation of complete narrative segments with multiple camera angles without character inconsistencies [3] Group 2: Market Impact - The excitement surrounding Seedance 2.0 has led to a surge in the media sector on the A-share market, with companies like Dook Culture and Rongxin Culture experiencing a 20% increase, while others like Light Media and Chinese Online saw gains exceeding 10% [6] - The model is expected to significantly reduce production costs and timelines for AI-generated dramas and short films, potentially expanding the related industry chain [6] Group 3: Ethical and Operational Challenges - The ability of Seedance 2.0 to replicate creator voices and imagine unseen scene details has sparked discussions about data sourcing and copyright issues, highlighting a common challenge in the global AI industry regarding the pace of technological advancement versus legal and ethical frameworks [6] - ByteDance has implemented risk mitigation measures, restricting the use of real human images/videos during the internal testing phase and requiring live authentication for generating real human videos [6] Group 4: User Experience and Membership - The popularity of Seedance 2.0 has led to system congestion, with users reporting wait times of up to 3 hours for video generation requests, even with paid "accelerated" features [8] - Currently, the model's features are available only to paid members, with membership tiers priced at 659 yuan, 1899 yuan, and 5199 yuan, and video generation operates on a points consumption system [9] Group 5: Competitive Landscape - The release of Seedance 2.0 signifies a heated competition in the AI video generation sector, with analysts noting that the landscape is evolving similarly to the competitive state of large language models in 2025, where differentiation will likely emerge in specific application scenarios [9]