证券研究报告·策略报告·策略周评 策略周评 20250831 牛市二阶段：聚焦 AI 应用的赔率交易 2025 年 08 月 31 日 [Table_Tag] [Table_Summary] 观点 6-8 月市场主线围绕人工智能展开，目前行情主要集中于上游基础设施类 硬件，其中景气确定性最高的海外链光模块、PCB 等在 6 月率先启动，8 月中旬以寒武纪为核心的国产算力正式开始补涨。在大盘量能充沛且产业 逻辑没有出现明显瑕疵的情况下，我们并不认为算力行情即将结束，但对 于场外持币者，上游硬件的强赚钱效应令人躁动不安，而客观上获利盘的 不断增厚会带来一定止盈压力，主观层面持续累积的涨幅又难免使风险偏 好较低的资金焦虑。"畏高"资金该如何选择？基于产业趋势确定性和赔 率思维，在下游应用端"另开一桌"亦是不错的思路。 本轮 AI 行情中，下游应用滞涨的核心原因在于短期确定性不足——具有 破圈效应的爆款单品和顺畅的商业模式均未出现，对应到上市公司层面则 体现为业绩能见度不足，本轮行情中没有成为资金首选。但从科技浪潮演 进角度，终局看 AI 赋能万物一定是通过应用端实现，亦即应用端的爆发 具有中期确定性，且空间较上游硬 ...

Group 1 - The event "Yang 'Digital' Pu Digital Salon" was held with the theme "Policy Protection and Compliance Progress: New Ecology of Quality Internet Content Creation" [1][3] - The venue "V Gathering Place" aims to create an ecosystem for quality internet content creators, providing convenient services since its opening in July [3][15] - The event featured discussions on the transformation of mainstream media in short video formats and the challenges faced in this transition [10][11] Group 2 - AIGC (AI-Generated Content) was highlighted for its revolutionary impact on content production, emphasizing its efficiency and potential across various applications [11][12] - Legal compliance issues in internet content creation were discussed, focusing on AI tool usage, intellectual property rights, and advertising regulations [12][15] - The salon encourages interaction among participants, fostering discussions on the development paths for quality internet content creation [14][15]

Core Viewpoint - AI technology has fully penetrated the e-commerce business chain, significantly enhancing efficiency and experience across key nodes such as content generation, intelligent recommendations, customer service response, supply chain management, and logistics fulfillment [1][2] Group 1: AI's Role in E-commerce Transformation - The e-commerce industry is transitioning from a traffic-driven model to an intelligence-driven model, with AI technology becoming a core variable for new growth [2] - Consumption has become the main driver of economic growth, with e-commerce's penetration rate rising from approximately 24% in 2013 to 78% in 2023, making it a crucial channel for stable consumption growth [2] - AI is driving efficiency upgrades in product production, supply chain, and customer service, creating a closed-loop capability for intelligent operations [2] Group 2: Policy and New Economic Models - Policy support is focused on expanding domestic demand and encouraging the integration of e-commerce with AI technology [2] - New models such as the "first-release economy" and scenario-based consumption are accelerating, supported by AI-enabled content production and scenario innovation [2] Group 3: Efficiency, Cost Reduction, and Revenue Growth - On the supply side, AI assists e-commerce in upgrading through efficiency improvements and cost reductions, with platforms like Alibaba, Pinduoduo, and Amazon leveraging AI for product selection, content generation, and operational optimization [3] - The integration of AI with SaaS is enhancing merchant management efficiency and user engagement, thereby strengthening platform service capabilities [3] - AI applications in cross-border e-commerce, including product selection, translation, and logistics forecasting, are helping Chinese merchants efficiently expand into overseas markets [3] Group 4: User Experience and Consumer Decision-making - AI is reshaping user experience and consumer decision-making through personalized recommendations, virtual fitting, and voice interaction, enhancing the shopping experience [3] - The upgrade of customer service and shopping guidance through AI-driven customer service and product selection is improving user satisfaction and conversion rates, establishing a new path for "goods finding people" [3]

QuestMobile最新发布的《2025中国移动（600941）互联网半年大报告》显示，截至2025年6月，中国移 动互联网月活跃用户规模已达到创纪录的12.67亿，用户黏性也显著增强，全网用户月人均单日使用时 长逼近8小时。AIGC(人工智能生成内容)行业总使用时长净增量同比增长率高达393.9%，成为新的增长 引擎。同时，以AI(人工智能)搜索为代表的创新应用异军突起，成为上半年当之无愧的"黑马赛道"。 点评：未来，移动互联网发展将更加依赖于技术和用户需求"双轮驱动"。AIGC和AI搜索的崛起，预示 着人工智能将在移动互联网中扮演更加重要的角色。同时，用户对内容质量、个性化体验和隐私保护的 需求，也将推动行业不断创新和优化。 NO.3具身智能公司RoboScience完成近2亿元融资 7月30日，RoboScience宣布完成近2亿元天使轮融资，由京东领投，招商局创投、商汤科技旗下国香资 本跟投，老股东零一创投继续追投，慕石资本担任本轮独家财务顾问。RoboScience注册成立于2024年 12月底，从今年3月开始正式运营，在本轮融资前已完成了数千万元种子轮融资。 点评：随着技术不断成熟和应用场景不断 ...

Core Viewpoint - Chinese enterprises are transitioning from "Made in China" to "Created in China," and further to "Belief in China," emphasizing the importance of craftsmanship in this evolution [2][6][56]. Group 1: Event Overview - The "East Forward" conference, held on July 25, 2025, in Zhejiang, focused on the theme of "craftsmanship" as a key to global expansion for Chinese brands [2][6]. - The event was supported by various governmental bodies and aimed to build a robust overseas service system for enterprises [2][3]. Group 2: Challenges and Strategies - The journey of brands like Yuanqi Forest illustrates the challenges of capturing user needs, innovating cultural symbols, and overcoming compliance hurdles in global markets [10]. - The renewable energy sector is witnessing a shift from participation to leadership, with companies like Zhejiang Chint New Energy focusing on building a global sales network and enhancing brand credibility [12]. Group 3: Cultural and Technological Integration - The integration of AI technology is reshaping industry logic, allowing Chinese stories to resonate globally, with companies becoming co-creators of value rather than mere exporters [16][19]. - The importance of localizing operations and understanding cultural nuances is highlighted as essential for successful market penetration [19][31]. Group 4: Compliance and Risk Management - Compliance with data regulations is critical, as over 80% of countries have implemented data laws, necessitating a global control mechanism for enterprises [22][25]. - The financial inefficiencies faced by outbound enterprises can be mitigated through services that ensure competitive pricing and comprehensive support [27]. Group 5: Market Opportunities - Dubai is emerging as a prime location for Chinese enterprises due to its strategic advantages, including a robust financial system and favorable tax policies [34][41]. - The conference also facilitated discussions on investment opportunities in the UAE, providing a platform for Chinese companies to connect with local stakeholders [41][52]. Group 6: Future Outlook - The concept of "craftsmanship" is seen as a guiding principle for Chinese brands to navigate global markets and establish a lasting presence [38][56]. - The conference concluded with the release of the "East Forward 2025 Global Innovation Directory," aimed at guiding Chinese brands in their international endeavors [38].

Group 1 - The event "Exploring AI Frontiers: Building a Transformation Ecosystem" aims to accelerate the transformation and application of AI technology in the Guangdong-Hong Kong-Macao Greater Bay Area, promoting deep integration of technological innovation and industry [1][2] - Over 70 entrepreneurs, investors, and technical managers from various sectors, including listed companies, state-owned enterprises, and high-tech firms, participated in the event to foster technology transfer and enhance the synergy between academia and industry [2][3] - Professor Xiong Hui from Hong Kong University of Science and Technology (Guangzhou) highlighted that the widespread application of AI large models is changing the global technology competition landscape, with China holding a leading position in open-source development and talent reserves [3][4] Group 2 - The event featured presentations from five research teams at Hong Kong University of Science and Technology (Guangzhou), showcasing breakthroughs in AI, focusing on areas such as multimodal interaction and intelligent system development, which are essential for market demand [3][4] - The university promotes the industrialization of scientific achievements through collaborations with enterprises, including the establishment of joint laboratories, supporting innovation in sectors like education, healthcare, and design [4] - A unique "Five-way Connection" segment was introduced, allowing direct interaction among university research teams, enterprise representatives, technical managers, investment institutions, and university startup project representatives to explore collaboration opportunities [4][5]

Core Viewpoint - The article emphasizes the urgent need for a new framework to understand data protection in the era of generative AI, highlighting the inadequacy of traditional data protection methods in addressing the unique challenges posed by AI technologies [2][3]. Group 1: Data Protection in Generative AI Era - In the generative AI era, data protection extends beyond traditional static data to include various types of data throughout the AI model lifecycle, such as training datasets, AI models, deployment data, user inputs, and AI-generated content [5][10]. - The paper titled "Rethinking Data Protection in the (Generative) Artificial Intelligence Era" aims to provide a novel and systematic perspective on data protection issues in the AI age [3]. Group 2: Types of Data to Protect - Training datasets are crucial as they often contain privacy or copyright data collected from multiple sources, making them a significant asset during model development [7]. - AI models, including their architecture and weights, become important data assets post-training, compressing vast amounts of data and holding substantial application value [7]. - Deployment data, such as system prompts and external databases, are essential for enhancing AI model performance in real-world applications [10]. - User inputs during model inference must be protected due to privacy, security, and ethical concerns, as they may contain sensitive personal information or proprietary business data [10]. - AI-generated content (AIGC) has reached high quality and poses new challenges regarding copyright and data protection, especially when used for training new models [10][17]. Group 3: Proposed Data Protection Framework - The article introduces a new hierarchical data protection framework with four levels: Data Non-usability, Data Privacy-preservation, Data Traceability, and Data Deletability, aiming to balance data utility and control [9][16]. - Level 1, Data Non-usability, prevents data from being used in AI training or inference, providing the highest level of protection [9]. - Level 2, Data Privacy-preservation, focuses on protecting personal privacy within data while maintaining some data usability [16]. - Level 3, Data Traceability, allows for tracking data sources and usage, enabling audits to prevent misuse [16]. - Level 4, Data Deletability, provides the ability to completely delete data or its effects, aligning with regulations like GDPR [16]. Group 4: Global Regulatory Landscape - The article reviews current global data protection laws and regulations, using the proposed hierarchical model to assess existing governance solutions and their strengths and weaknesses [14]. - It highlights the challenges posed by cross-border data flows and differing national standards, which create compliance difficulties for global developers [17]. Group 5: Ethical Considerations - Data protection in the AI era is closely linked to ethical considerations, such as individual autonomy over data, fairness, and the prevention of malicious data use [17]. - The balance between technological innovation and ethical values is a critical consideration for all AI practitioners [17].

Core Viewpoint - The emergence of Generative Artificial Intelligence (AIGC) represents not only a technological revolution but also a reflection of human creativity, values, and trust, emphasizing the need for a humanistic approach to guide technology in serving humanity [2][5]. Group 1: AIGC Definition and Evolution - AIGC is defined as algorithms capable of generating text, images, music, and videos, exemplified by tools like ChatGPT, Midjourney, and DALL·E [2]. - The evolution of AI has progressed through several waves: from symbolic reasoning and rule-based systems to statistical learning, deep learning breakthroughs, and now to AIGC as a collaborative partner rather than just an auxiliary tool [3]. Group 2: Cultural Impact - AIGC is not merely a technical phenomenon; it has become a "cultural software" that reshapes how culture is expressed and defined in the digital age [3]. - The rise of AI-generated content raises questions about originality and the emotional and cultural value of rapidly produced works, echoing concerns raised by philosopher Walter Benjamin regarding mechanical reproduction [3]. Group 3: Applications in Education - AIGC has transformed education by providing personalized, scalable, and adaptive learning experiences, such as AI-assisted tutoring and dynamically generated learning materials [4]. - However, challenges include potential over-reliance on AI by students, which may weaken critical thinking skills, and the risk of exacerbating the digital divide due to uneven technology distribution [4]. Group 4: Applications in Healthcare - In healthcare, AIGC has demonstrated effectiveness through AI-generated diagnostic reports and image analysis tools, enhancing diagnostic efficiency and supporting clinical decision-making [4]. - Notable developments include specialized large language models like Google DeepMind's MedGemma and SenseTime's "Da Yi" model, which assist in diagnosis and patient communication [4]. Group 5: Societal Challenges - AIGC poses significant societal challenges, including information pollution, the ambiguity of copyright in creative industries, and potential job displacement in various sectors [5]. - There is a growing concern about a "crisis of trust" as distinguishing between true and false content becomes increasingly difficult, highlighting the need for responsible guidance in shaping AI's role in society [5].

Core Insights - The financial industry is experiencing a transformation driven by AI, particularly through the implementation of AI Agents, which are expected to enhance efficiency and innovation by 2025 [1][2] - The integration of multi-agent architecture is seen as a key approach for deploying large models in complex financial scenarios, significantly improving credit operations, risk assessment, and interactive intelligence [1][2] AI Empowerment and Process Transformation - The year 2025 is recognized as the "AI Agent Year" in the global fintech sector, marking a significant shift towards full-chain intelligent revolution in finance [2] - AI is driving a transition from traditional waterfall software development to AI-enhanced agile development, creating a complete closed-loop system from demand analysis to intelligent testing [2][3] Code Quality and Efficiency - AI's integration into coding processes has led to a 20%-30% increase in efficiency, with AI-generated code resolving over 90% of issues, although human quality checks are still necessary [3] - The AI's role in development is expected to evolve from functional enhancement to more significant involvement in decision-making processes [3] Credit Ecosystem Upgrade - AI Agents are set to revolutionize traditional banking operations, customer service, and risk management, with a focus on improving efficiency in credit processes [4][5] - AI can generate customer analysis reports in just 2 hours, a process that previously took weeks or months, significantly enhancing productivity and output quality [4] Risk Assessment and Interaction Intelligence - AI models can conduct real-time monitoring and risk analysis by leveraging large data platforms and unstructured data, providing timely risk alerts [5] - The combination of AI and big data allows for personalized credit solutions, improving the precision of matching customer needs [5] Diverse Application Strategies - Different banks are adopting varied strategies in the AI Agent era, with state-owned banks focusing on integrated smart platforms, while smaller banks may collaborate with vendors to gradually enhance their capabilities [6] Overcoming Data Governance Challenges - AI models are expected to aid in the digital transformation of the financial sector, although challenges such as AI hallucinations and data privacy remain [7] - Knowledge graph technology is proposed to build compliance knowledge bases, enhancing the accuracy and completeness of AI responses [7] Future of AI Service Platforms - AI service platforms are anticipated to become strategic priorities for leading tech companies, simplifying complex operations into user-friendly interfaces [8] - The competitive edge of AI Agents lies in their deep integration with business scenarios, which can reduce trial-and-error costs and accelerate the realization of AI's value in data analysis [8]

Core Insights - The male consumer market is projected to reach 634 million users by April 2025, with a significant increase in monthly usage time to 168 hours, reflecting a year-on-year growth of 6.9% [9][8] - The demographic shift shows an increasing proportion of male users aged 41 and above, expected to reach 48.1%, indicating an aging trend [3][8] - Online consumption capabilities among male users have improved, with 70% of users spending over 1,000 yuan, marking a 1% increase year-on-year [3][8] User Demographics - The male user base is diversifying, with a notable decline in new and second-tier cities, while first-tier cities see a 1% increase in user proportion [3][13] - The online behavior of different age groups varies significantly, with younger males (under 30) being heavy gamers, while middle-aged males (31-45) focus on family-oriented consumption [4][15] Online Behavior and Preferences - Mobile video and social platforms account for over 60% of male users' online time, with video platform usage increasing by 2.2% [11][3] - The top five industries preferred by male users include smart home products, AIGC, traffic violation inquiries, vehicle services, and automotive information, with active user numbers reaching 214 million, 157 million, 146 million, 115 million, and 109 million respectively [3][16] Age Group Insights - Users under 30 are primarily engaged in gaming, with a strong preference for strategy and shooting games, while those aged 46 and above favor traditional card games and news content [4][50] - Males aged 31-45 are characterized as new middle-class consumers, actively participating in financial investments and showing a preference for smart technology and quality consumption [35][36] Consumption Trends - The "male beauty economy" is thriving, with demand shifting from basic skincare to makeup and fragrance products, targeting young male consumers through collaborations with esports and sports IPs [31][20] - Middle-aged males are becoming key decision-makers in family consumption, increasingly favoring smart home products and technology-driven lifestyles [46][44] Media Preferences - The top apps for male users include community socializing, automotive information, and stock trading, with significant user engagement across various platforms [64][66] - For users under 30, popular apps include strategy games and social forums, while those aged 31-45 prefer car services and smart home applications [67][69] - Users aged 46 and above show a strong preference for short video platforms and news apps, indicating a shift towards digital content consumption [71][72]