Core Insights - Kuaishou, founded 14 years ago, has consistently positioned itself as a technology company driven by AI and innovation, exploring the integration of technology and industry to enhance efficiency and upgrade industries [1][2] - The launch of Kuaishou's AI model, Keling AI, has established it as a leading player in global video generation technology, with over 45 million users and partnerships with more than 20,000 enterprises across over 80 countries [2][4] - Kuaishou emphasizes the importance of embedding AI deeply into business operations and workflows to create a sustainable competitive advantage, developing a comprehensive intelligent content and commercial ecosystem [3][5] Group 1 - Kuaishou's commitment to AI technology development aligns with national strategies, showcasing its ability to innovate and adapt to new technological waves [1][2] - The company has successfully transitioned Keling AI from a content creation tool to a new productivity engine that reshapes the video production chain [2][3] - Kuaishou's AI capabilities extend across various applications, including e-commerce and advertising, enhancing its operational efficiency and market reach [3][5] Group 2 - The establishment of the first overseas AI cooperation center in collaboration with BRICS nations marks Kuaishou's efforts to support global SMEs and promote innovation on an international scale [4][5] - Keling AI has gained recognition in the global entertainment industry, participating in major film festivals and connecting creators with broader audiences [4] - Kuaishou believes that its innovative capabilities will enable Chinese enterprises to serve global users effectively, driving high-quality development in the real economy [5]

快手科技高级副总裁马宏彬 马宏彬称，从记录与分享的生活社区到深耕产业的数字平台，快手14年来持续进化，既受益于数字技术飞速发展的时代与政府鼓励政策，也源于公司创新驱 动的技术底蕴与丰厚AI基因。 随着新一轮AI科技浪潮到来，2023年初快手启动了新一代AI战略，从基础研究到技术突破、从工具革新到应用创新，多领域、多层次地加大研发投入、布 局与应用探索。在马宏彬看来，AI技术浪潮正在从单纯的技术突破走向大规模应用创新，推动科技与产业深度融合与高质量发展。 这一轮生成式AI浪潮中，快手凭借硬核技术突破与应用创新，推出了全球首个用户可用的真实影像级视频生成大模型可灵AI，并迅速跻身全球第一梯队。 经过一年多高速迭代，可灵2.5 Turbo版本生成能力已逼近影视级标准，正在全球范围成为有望重塑视频生产工业的新质生产力引擎。今年下半年，可灵AI接 连亮相釜山电影节、东京电影节和戛纳秋季电视节，引发关注并连接起更广阔市场。 "新一代AI技术，正在从程序员手中的'技术秀'向产业一线的'效率革命'演进。"马宏彬表示，可灵AI可使影视制作成本降低至传统方法的十分之一，制作时 间缩短约60%，将推动影视产业生态与创作逻辑发生深度 ...

11月8日，2025世界互联网大会乌镇峰会"人工智能技术创新与治理论坛"上，快手科技(01024)高级副总裁马宏彬发表主题演讲，分享了快手在AI时代的创新 实践与底层思考。 马宏彬表示，从记录与分享的生活社区到深耕产业的数字平台，快手14年来持续进化，既受益于数字技术飞速发展的时代与政府鼓励政策，也源于公司创新 驱动的技术底蕴与丰厚AI基因。 随着新一轮AI科技浪潮到来，2023年初快手启动了新一代AI战略，从基础研究到技术突破、从工具革新到应用创新，多领域、多层次地加大研发投入、布 局与应用探索。在马宏彬看来，AI技术浪潮正在从单纯的技术突破走向大规模应用创新，推动科技与产业深度融合与高质量发展。 从"技术秀"到"效率革命" 人工智能在赋能各行各业的同时，也带来了安全、伦理等方面的全新挑战。马宏彬表示，快手积极探索"将治理嵌入技术"，实现从"治理技术"到"技术治 理"的范式改变。 新的治理理念下，很多老问题有了更高效解决方法，快手还积极响应《人工智能生成合成内容标识办法》，已上线用户可见的显式标识及嵌入元数据中的隐 式标识双重体系。"当治理真正嵌入技术内核，合规不再是创新的约束条件，而是创新的方向指引。"马 ...

这一轮生成式AI浪潮中，快手凭借硬核技术突破与应用创新，推出了全球首个用户可用的真实影像级视频生成大模型可灵AI，并迅速跻身全球第一梯队。 智通财经APP获悉，11月8日，2025世界互联网大会乌镇峰会"人工智能技术创新与治理论坛"上，快手科技(01024)高级副总裁马宏彬发表主题演讲，分享了 快手在AI时代的创新实践与底层思考。 马宏彬表示，从记录与分享的生活社区到深耕产业的数字平台，快手14年来持续进化，既受益于数字技术飞速发展的时代与政府鼓励政策，也源于公司创新 驱动的技术底蕴与丰厚AI基因。 随着新一轮AI科技浪潮到来，2023年初快手启动了新一代AI战略，从基础研究到技术突破、从工具革新到应用创新，多领域、多层次地加大研发投入、布 局与应用探索。在马宏彬看来，AI技术浪潮正在从单纯的技术突破走向大规模应用创新，推动科技与产业深度融合与高质量发展。 从"技术秀"到"效率革命" 经过一年多高速迭代，可灵2.5 Turbo版本生成能力已逼近影视级标准，正在全球范围成为有望重塑视频生产工业的新质生产力引擎。今年下半年，可灵AI接 连亮相釜山电影节、东京电影节和戛纳秋季电视节，引发关注并连接起更广阔市场。 " ...

Core Viewpoint - The article emphasizes that 2025 is seen as a pivotal year for AI's deep application, with a focus on the systematic realization of industrial-level value through AI technologies like multimodal generation and agents [1]. Group 1: AI Integration in Business - Companies embracing AI must deeply consider and actively respond to the integration of AI technologies with specific application scenarios [2]. - Kuaishou, a technology-driven company, showcased its plans for integrating AI with business scenarios during its 1024 Programmer's Day event, highlighting the importance of combining AI technology with practical applications [2][4]. - Kuaishou has rapidly integrated AI into various business processes, significantly enhancing content production, recommendation, distribution, and e-commerce search [4]. Group 2: AI Technology Advancements - Since its launch in June 2024, Kuaishou's Keling AI has undergone over 30 updates, becoming an industry benchmark with significant improvements in text response, dynamic effects, and aesthetic quality [7]. - The latest version, Keling 2.5 Turbo, has reduced API prices by 30% compared to its predecessor, achieving the top position on the Artificial Analysis video ranking [7]. Group 3: Innovative AI Systems - Kuaishou's OneRec system, launched in June, aims to address issues in traditional recommendation systems by employing an end-to-end generative architecture, enhancing user preference alignment through reinforcement learning [9]. - The implementation of OneRec has shown measurable improvements in various metrics, such as a 5.09% increase in local life short videos and a 3.25% increase in e-commerce product cards [9]. Group 4: Comprehensive AI Ecosystem - Kuaishou's strategy reflects a broader industry trend where companies are moving from isolated AI breakthroughs to comprehensive application deployments [13]. - The integration of AI technologies is transforming business operations, making AI a crucial engine for growth rather than just a tool for enhancing specific functions [14]. - Kuaishou's early and comprehensive AI strategy has positioned it to capitalize on every wave of AI technology, leading to significant operational efficiency and revenue growth [14][15]. Group 5: Market Impact and Future Potential - Kuaishou's AI applications have created a virtuous cycle from innovation to application and revenue growth, enhancing its competitive edge and market adaptability [15]. - The company's proactive approach to AI has attracted positive reassessment from the market and investors, further solidifying its industry-leading position [16].

Core Insights - The core viewpoint of the article emphasizes that the competition in the technology industry is increasingly centered around effectively integrating AI technology with specific application scenarios [3][4]. Group 1: AI Application and Innovation - Kuaishou's "1024 Programmer Festival" highlighted the importance of AI applications in driving innovation and efficiency within the company [3][5]. - The company has made significant breakthroughs in AI technology, including the upgrade of its Keling AI model, which has shown superior performance compared to other models in various dimensions [3][4]. - Kuaishou introduced the OneSearch framework for e-commerce, which has successfully served millions of users daily, marking a significant advancement in industrial search technology [4]. Group 2: AI in Development and Collaboration - Kuaishou launched a comprehensive AI programming suite, including self-developed models and intelligent development tools, aimed at enhancing collaboration in AI development [4][5]. - The CodeFlicker tool has been widely adopted by over 80% of Kuaishou's programmers, showcasing its effectiveness in addressing complex enterprise needs [4]. Group 3: Efficiency and Cost Reduction - AI applications have significantly reduced the development time for new live streaming gifts, allowing for the creation of 160 new gift types in just three weeks, compared to three months previously [6]. - The introduction of AI tools has lowered the costs associated with technical development in live streaming, providing users with more diverse expression options [6]. Group 4: Market Value Reassessment - Kuaishou's advancements in AI technology have prompted brokers to accelerate the reassessment of the company's market value, indicating a positive outlook for its future [8].

AI 重构商业化，电商和广告系统打开提效通道。除此前内容推荐领域的OneRec，快手近期在AI 提升商 业化效率上也进展显著。1）电商方面，OneSearch 是业界首个工业级部署的电商搜索端到端生成式框 架，目前已在快手多个电商搜索场景部署，服务数百万DAU，产生数千万PV，带来CTR 的稳定提升； 2）广告方面，生成式强化学习在广告自动出价场景落地，25 年至今实现超3% 的广告收入提升。 我们预测公司25-27 年经调整归母净利润为196/230/259 亿元。参考可比公司给予公司26 年17xPE 估 值，对应合理价值为3,911 亿CNY，折合4,281 亿HKD（港币兑人民币汇率0.914），目标价99.07 港元/ 股，维持"买入"评级。 机构：东方证券 风险提示 宏观消费恢复不及预期，国内商业化效率不及预期，海外业务亏损加大，可灵技术迭代不及 预期 研究员：项雯倩/李雨琪/金沐阳 1） 可灵2.5 Turbo 延续围绕专业内容创作者需求的思路持续迭代，在稳定性和想象力上再度升级，核心 体现在：①文本理解能力显著提升，提示词响应和逻辑理解更加精细化；②高速动态质量更流畅稳定， 画面运动强度和幅度 ...

Investment Rating - The report maintains a "Buy" rating for the company [5][11]. Core Views - The company is expected to leverage its upgraded 2.5 Turbo model, which combines performance enhancements with a 30% price reduction, to drive user growth and revenue increase [2][3]. - AI is enhancing the core business efficiency, with the commercial system being restructured to provide long-term growth momentum [2]. Financial Forecast and Investment Recommendations - The adjusted net profit forecast for the company from 2025 to 2027 is projected to be CNY 196 billion, CNY 230 billion, and CNY 259 billion respectively [4]. - The target price is set at HKD 99.07 per share, based on a 17x PE valuation for 2026, leading to a reasonable value of CNY 3,911 billion, equivalent to HKD 4,281 billion [4][11]. - The company's revenue is expected to grow from CNY 113.47 billion in 2023 to CNY 166.83 billion in 2027, reflecting a compound annual growth rate [4][15]. Key Financial Metrics - Revenue growth rates are projected at 20.5% for 2023, 11.8% for 2024, and gradually declining to 7.8% by 2027 [4][15]. - The gross margin is expected to improve from 50.6% in 2023 to 56.8% in 2027, indicating enhanced profitability [4][15]. - The net profit margin is projected to increase from 5.6% in 2023 to 14.5% in 2027, showcasing improved efficiency [4][15].

Core Insights - Kuaishou has introduced OneSearch, an end-to-end generative framework for e-commerce search, addressing challenges in traditional search architectures [1][2][4] Group 1: Innovation and Technology - OneSearch integrates three innovations: Keyword Enhanced Hierarchical Quantization Encoding (KHQE), Multi-Perspective User Behavior Sequence Injection Strategy, and Preference-Aware Reward System (PARS) [2] - KHQE uses RQ-OPQ encoding to model product features, creating a "smart identity" for each product, enhancing retrieval accuracy [2] - The Multi-Perspective User Behavior Sequence Injection Strategy captures both recent preferences and long-term interests, improving user intent understanding [2] - PARS combines multi-stage supervised fine-tuning with adaptive reinforcement learning to capture fine-grained user preference signals, enhancing ranking performance while ensuring diversity [2] Group 2: Performance Metrics - OneSearch has shown significant improvements in various metrics compared to traditional systems, with a 3.22% increase in order volume and a 2.4% growth in the number of buyers [4] - In offline experiments, OneSearch outperformed existing systems in terms of Click-Through Rate (CTR) and Conversion Rate (CVR), with notable improvements in user satisfaction and item quality [5][6] - The system achieved an 8-fold increase in machine computation efficiency and a 75.40% reduction in online inference costs, optimizing resource utilization [5] Group 3: Market Impact and Future Directions - OneSearch's deployment marks a significant breakthrough in replacing traditional search links with generative models in large-scale industrial scenarios [4][6] - The system excels in cold start scenarios, effectively addressing challenges related to long-tail users and newly listed products [6] - Kuaishou plans to explore online real-time encoding solutions and enhance reinforcement learning mechanisms to better match user preferences, aiming for a more intelligent and precise e-commerce search experience [6]

Core Insights - The article discusses the introduction of OneSearch by Kuaishou, an innovative end-to-end generative framework for e-commerce search, aimed at addressing the limitations of traditional cascading search architectures [1][2]. Group 1: Challenges in Traditional E-commerce Search - Traditional e-commerce platforms utilize a "recall, rough ranking, fine ranking" cascading search structure, which, while stable, faces issues such as chaotic product descriptions, relevance problems, bottlenecks in the cascading structure, and cold start challenges [1]. Group 2: OneSearch Framework Innovations - OneSearch integrates three major innovations: - Keyword Hierarchical Quantization Encoding (KHQE) module, which models product features in both vertical and horizontal dimensions, creating a rich semantic "smart ID" for each product, enhancing retrieval accuracy [2]. - Multi-perspective User Behavior Sequence Injection Strategy, which captures both recent preferences and long-term interests, allowing for a deeper understanding of user intent and improving personalized search accuracy [2]. - Preference-Aware Reward System (PARS), which combines multi-stage supervised fine-tuning with adaptive reward reinforcement learning to capture fine-grained user preference signals, enhancing ranking performance while ensuring diversity in generated results [2]. Group 3: Performance Metrics and Results - Offline experiments indicate that OneSearch significantly outperforms existing cascading systems across various metrics. Online deployment results show a 3.22% increase in order volume and a 2.4% growth in the number of buyers, marking a significant achievement in large-scale industrial applications [4]. - In a comparative analysis, OneSearch demonstrated improvements in metrics such as Click-Through Rate (CTR) and Conversion Rate (CVR), with notable enhancements in overall page satisfaction and product quality compared to traditional systems [5]. - OneSearch also excelled in cold start scenarios, effectively addressing the ranking challenges for long-tail users and newly listed products, indicating its robustness in diverse conditions [6]. Group 4: Future Directions - Kuaishou plans to continue exploring online real-time encoding solutions to bridge the gap between predefined encoding and streaming training, while also integrating more powerful reinforcement learning mechanisms to better match user preferences [6]. - The ongoing technological advancements are expected to lead to a more intelligent, precise, and user-centric e-commerce search experience, fulfilling the ideal of "one-step" search for users [6].