行 业 研 究 传媒 2025 年 12 月 31 日 可灵 AI：商业化空间进一步提升 投资要点： 可灵"全能灵感周"，多个产品功能迭代。 行 业 动 态 跟 踪 可灵"全能灵感周"，第一天上线了"视频 O1 模型"，可以实现 仅需简单对话，即可使用图片、视频等各种素材；同时提高了视频的 稳定性，主题建构能够在复杂场景中保持一致。 第二天上线了"图片 O1 模型"，可以实现让元素高度统一，响应 专业级创作需求。 第三天上线了音画同出视频"可灵 2.6"。实现单次可以生成完整 的音视频。 第四天上线了"可灵数字人 2.0"，表演力全面进化，情绪传达更 生动细节，支持单次上限至 5 分钟。 第五天上线了"可灵 O1 主体库和对比模板"，支持多个主体，多 素材随心搭配。通过智能 AI 补图，可以创建专属形象。 可灵 AI 功能迭代带来商业化空间进一步提升。 以可灵 2.6 模型为例，同样生成 5 秒的视频，快手 1.6 模型"标准 模式"收费 20 个灵感值，高品质模式收费 35 个灵感值；而可灵 2.6 模型，仅支持高品质模型，收费 50 个灵感值。价格提升，进一步打开 商业化空间。 投资建议 建议关注快手及其 ...

点 击 蓝 字 关 注 我 们 本文全文共 1662 字，阅读全文约 5 分钟 前11个月全国一般公共预算收入同比增长0.8% 前11个月全国一般公共预算收入同比增长0.8% 2025年1-11月我国电子商务发展情况 2025年1-11月全国吸收外资6931.8亿元人民币 A 0.8% 其中,前11个月 全国税收收入 全国非税收入 164814 亿元 35702 亿元 ▼ 3.7% A 1.8% 分中央和地方看,前11个月 中央一般公共 地方一般公寓 预算收入 预算本级收入 88464 亿元 1.12052 亿元 A 2.2% V 1% 财政支出方面,前11个月 全国一般公共 预算或出 248538 亿元 A 1.4% 分中央和地方看 网络零售提品质释放内需潜力。 深入实施《关于大力发展数字消费共创数字时代美好生活的指导意 见》，数字消费、线上服务成为增长牵引力。据国家统计局数据，1-11月全国网上零售额增长9.1%。商务 部重点监测平台数字产品增长8.2%，其中智能穿戴、智能机器人分别增长22.1%和19.4%，线上服务消费 增长21.7%。网络零售带动快递业务量快速增长，据国家邮政局数据，截至11月30日 ...

丝路电商促合作深化互利共赢。举办全国"丝路云品"电商月活动，打造澳大利亚牛排、乌兹别克斯坦饼 干、印尼燕窝等爆款产品，销量增长超10倍，带动1-11月重点监测平台进口商品零售额增长5.6%，放大 电商联通国内国际消费市场的渠道效应。能力建设创新赋能伙伴国，连续在泰国、老挝举办云上大讲堂 海外实训专场，来自政府、行业和中小企业的150余名代表参与，助力周边命运共同体建设取得新成 果。 商务部：1-11月全国网上零售额增长9.1% 中新网12月19日电 据商务部网站消息，商务部电子商务司负责人介绍2025年1-11月我国电子商务发展 情况。 商务部贯彻落实党的二十届四中全会和中央经济工作会议精神，发挥我国电子商务在提振消费、驱动创 新、对外开放等方面的积极作用，助力建设强大国内市场、培育壮大新动能、推动国际经贸合作共赢。 网络零售提品质释放内需潜力。深入实施《关于大力发展数字消费共创数字时代美好生活的指导意 见》，数字消费、线上服务成为增长牵引力。据国家统计局数据，1-11月全国网上零售额增长9.1%。商 务部重点监测平台数字产品增长8.2%，其中智能穿戴、智能机器人分别增长22.1%和19.4%，线上服务 消费 ...

12月19日，商务部电子商务司负责人介绍2025年1-11月我国电子商务发展情况。网络零售提品质释放内 需潜力。据国家统计局数据，1-11月全国网上零售额增长9.1%。商务部重点监测平台数字产品增长 8.2%，其中智能穿戴、智能机器人分别增长22.1%和19.4%，线上服务消费增长21.7%。网络零售带动快 递业务量快速增长，据国家邮政局数据，截至11月30日我国快递业务量突破1800亿件，创历史新高。 丝路电商促合作深化互利共赢。举办全国"丝路云品"电商月活动，打造澳大利亚牛排、乌兹别克斯坦饼 干、印尼燕窝等爆款产品，销量增长超10倍，带动1-11月重点监测平台进口商品零售额增长5.6%，放大 电商联通国内国际消费市场的渠道效应。能力建设创新赋能伙伴国，连续在泰国、老挝举办云上大讲堂 海外实训专场，来自政府、行业和中小企业的150余名代表参与，助力周边命运共同体建设取得新成 果。 财经频道更多独家策划、专家专栏，免费查阅>> 责任编辑：栎树 电商平台助转型增强创新驱动。产业电商惠企对接近450场，推进农业、工业中小企业加快数智化转 型，1-11月商务大数据监测农村网零额增长9.8%，重点监测产业电商平台纺织 ...

Core Viewpoint - ByteDance has released the Seedance 1.5 Pro video generation model, which significantly enhances audio-visual synchronization and local dialect support, improving the realism and emotional expression in generated videos [1][36]. Group 1: Key Features of Seedance 1.5 Pro - The model supports audio-visual synchronization generation, with improved lip-sync and tone alignment capabilities, particularly effective for various dialects [3][4]. - Enhanced semantic understanding allows the model to better interpret narrative contexts, improving emotional control and professional performance [3][12]. - The model offers precise and rich camera control, enabling complex shots such as long takes and zooms [3][26]. - It can generate videos of varying lengths, with a maximum of 12 seconds in a single output [3]. Group 2: Dialect and Cultural Relevance - The ability to generate dialect content is crucial for adding authenticity and regional characteristics to characters in film and television [5][12]. - The model has shown impressive results in generating dialects like Shaanxi and Sichuan, maintaining the unique phonetic qualities and emotional tones [7][9][11]. Group 3: Emotional and Performance Capabilities - The model demonstrates strong emotional expression, effectively conveying complex feelings such as fear and desperation through facial expressions and voice modulation [20][21]. - It can generate realistic animal sounds and expressions, enhancing the appeal of pet-related content [15][17]. Group 4: Technical Advancements - The model has improved its ability to handle complex camera movements, including advanced techniques like the Hitchcock zoom, achieving smooth transitions and maintaining visual consistency [29][30][32]. - The integration of audio capabilities with high-quality text-to-video generation has significantly reduced the complexity of video production [36][37]. Group 5: Market Implications - The advancements in Seedance 1.5 Pro are expected to lead to a surge in video generation products and video agent applications, making it easier for users to create high-quality content [37].

Core Insights - Yuanbao has officially launched a new feature called "one-sentence video creation," enabling users to easily create videos without any prior editing skills [1] - The underlying technology is based on Tencent's latest open-source HunyuanVideo 1.5 model, which supports both text-to-video and image-to-video functionalities [1] Group 1: Feature Overview - Users can transform a single sentence or a static photo into a dynamic video using two methods: "text to video" and "image to video" [1] - The "text to video" feature allows users to input a descriptive prompt, such as "a cat walking in a cyberpunk city," to generate a video [1] - The "image to video" feature enables users to upload a photo and apply simple commands to animate the static image, making it easy to add motion to landscapes or pet photos [1] Group 2: Technical Specifications - The HunyuanVideo 1.5 model supports both Chinese and English for video generation and ensures high consistency in color tones and details between images and videos [1] - The model operates efficiently with a lightweight size of only 8.3 billion parameters and can run smoothly on consumer-grade graphics cards with 14GB of memory [1]

Core Viewpoint - The company, Chuling Information (300250.SZ), clarifies that it does not engage in the fields of trusted computing products and generative video, focusing instead on its integrated enterprise intelligent computing network product [1] Group 1: Product Offering - The company has developed an integrated computing network product that combines all-optical connectivity, computing power scheduling, and partial intelligent agent deployment [1] - This product provides various practical business intelligent agents, including network coverage and operation maintenance, business insight, security control, video analysis, intelligent switchboard, document assistant, and knowledge base assistant [1] - The integrated solution is designed to meet the intelligent needs of enterprises across different business scenarios [1]

Market Overview - The market experienced a decline on October 31, 2025, with all three major indices closing lower: Shanghai Composite Index down 0.81%, Shenzhen Component Index down 1.14%, and ChiNext Index down 2.31% [1] Sector Performance - Biopharmaceuticals, chemical pharmaceuticals, and cultural media sectors showed the highest gains, while insurance, small metals, and semiconductors faced the largest declines [1] ETF Performance - Innovation drugs, film and television, and software-related ETFs performed well, likely driven by relevant news [2] Biopharmaceutical Sector - The National Medical Insurance negotiation began on October 30, 2025, introducing a "commercial insurance innovative drug catalog" mechanism, which includes innovative drugs with high clinical value that are not yet part of the basic medical insurance catalog [2] - According to Industrial Securities, the ongoing business development (BD) in the innovative drug sector is expected to sustain the current industry prosperity, with a focus on "innovation + internationalization" [2] Cultural Industry - Cultural enterprises reported a net profit of 909.3 billion yuan in the first three quarters, a year-on-year increase of 14.2%, with an operating income profit margin of 8.30% [2] - The cultural industry is experiencing a transformation driven by AI-generated content (AIGC), which is expected to create new opportunities in the industry [3] Software Industry - The software and information technology service industry in China saw a revenue of 1,111.26 billion yuan in the first three quarters, reflecting a year-on-year growth of 13.0% [4] - Longjiang Securities highlighted the importance of achieving complete autonomy in key areas, particularly in the domestic operating system and chip industries, as a response to global technological competition [4]

Core Insights - The report from China Merchants Securities highlights that generative video technology is revolutionizing traditional film and television content production, activating new dynamics within the industry [1] - AI tools like Sora are capable of handling most production processes for various video formats, leading to innovative content supply and new opportunities across the industry chain [1] Group 1: Technological Advancements - OpenAI's release of the generative video model Sora2 marks significant technological breakthroughs, including realistic physical world simulation, multi-modal integration for simultaneous audio generation, and initial capabilities for narrative logic and editing akin to real filming [2] - The Sora APP allows users to creatively rework popular videos and integrate virtual characters, embedding social attributes into the AI video creation process for rapid dissemination [2] Group 2: Emerging Content Forms - The rise of "AI comic dramas" represents a new content form that merges AI technology with anime short dramas, retaining original storylines and artistic styles while incorporating the quick and diverse characteristics of short videos [2] - This new form is expected to capture a significant share of internet time among younger audiences, creating numerous opportunities across the industry chain [2] Group 3: Commercialization Trends - The next phase of AI video applications is anticipated to focus on deep integration with social interactions, accelerating the commercialization of consumer applications [3] - The evolution of ChatGPT into a comprehensive ecosystem will enhance the platformization of generative video tools, facilitating a broader user reach and transforming from simple tools to a full-cycle platform for generation, distribution, and monetization [3] - The integration of generative video with AIAgent aims to address the challenges of supporting all video production processes, thereby reconstructing the video production paradigm for a one-stop solution [3]

Core Insights - OpenAI has launched a significant upgrade to its video generation model, Sora 2, along with a social application, Sora App, which offers improved physical accuracy, realism, and controllability compared to its predecessor [1][3] - The stock prices of related companies, such as Chuliang Information and Visual China, have seen notable increases following the announcement of Sora 2 [1] Group 1: Sora 2 Features and Market Impact - Sora 2 is defined as a breakthrough in video generation, achieving significant advancements in physical motion and character modeling, with capabilities for multi-modal collaboration, including synchronized audio and dialogue generation [3] - The model has seen rapid adoption, surpassing 1 million downloads within five days of its launch, outpacing the previous popularity of ChatGPT [3] - The competitive landscape includes other models like xAI's Grok Imagine v0.9 and Google's Veo 3.1, which also focus on enhancing video generation quality and capabilities [4] Group 2: Market Growth and Company Strategies - The global market for AI video generation is projected to grow from $615 million in 2024 to $717 million in 2025, with a compound annual growth rate of 20% expected from 2025 to 2032, reaching $2.563 billion by 2032 [6] - Domestic companies are actively positioning themselves in this expanding market, with Hanwang Technology and Visual China making strides in multi-modal recognition and creative content generation [6][5] - Visual China reported a revenue of 399 million yuan in the first half of 2025, reflecting a slight year-on-year growth of 0.05% [6] Group 3: Industry Trends and Investment Opportunities - The development of AI video generation technology is transitioning from auxiliary creation to autonomous generation, with improvements in model capabilities leading to cost reductions and efficiency gains in industries like film, advertising, and gaming [8] - Analysts suggest that the rapid advancement of AI video will stimulate demand for computing power and storage, enhancing investment sentiment in related sectors [8] - Key investment themes include the scaling of AI video applications, the transition of smart terminals, and the monetization potential of AI-driven video content [8]