大摩闭门会：邢自强、Laura Wang：全球震荡，何去何从 摘要 美国 AI 投资采取重资产豪赌策略，追求 AGI，而中国则采取轻量化路线， 关注产业生态铺开，利用基础设施、人才和数据成本优势降低 AI 投资成 本，从而减少泡沫风险，对经济产生更稳定和可持续的影响。 当前科技股估值接近 23 倍市盈率，结构脆弱，依赖少数大盘股上涨， 长期看好科技股，受益于行业普及应用，多数标普 500 公司预期 AI 带 来盈利增长，短期有波动风险，长期增长潜力良好。 预计美联储将在 2026 年 1 月、4 月和 6 月各降息一次，总体降息幅度 不变，但 12 月份降息预期取消，美股整体行情将由普惠型辐射型行情 带动，而非仅依赖人工智能相关公司。 中国银行业未来几年将逐渐提高贷款利率，以覆盖长期风险并消化不良 资产，监管支持合理定价，民生银行将贷款利率回升作为考核重点，有 助于金融行业持续反弹。 高线城市房地产市场企稳可能要等到 2027 年，因消化前期过剩房产过 程复杂，房地产市场的传导机制广泛而复杂，单靠政策进行单点突破往 往效果不及预期。 预计 2026 年消费平稳中放缓，投资比 2025 年有所改善但全年基本持 ...

Investment Rating - The report maintains a "Buy" rating for the company, with a target price raised to HKD 790, indicating a potential upside of 26.7% from the last closing price [4][55]. Core Insights - The company's Q3 2025 performance showed robust growth driven by AI efficiency and ecosystem expansion, with revenue reaching RMB 193 billion, a year-on-year increase of 15% and a quarter-on-quarter increase of 5% [1][6]. - Non-IFRS net profit for Q3 2025 was RMB 70.6 billion, reflecting an 18% year-on-year growth and a 12% quarter-on-quarter growth, indicating improved operational efficiency and profit quality [1][6]. - The gaming segment showed strong performance, with domestic game revenue of RMB 42.8 billion, up 15% year-on-year, and international game revenue of RMB 20.8 billion, up 43% year-on-year [2][8]. Summary by Sections Financial Performance - Q3 2025 revenue was RMB 193 billion, with a gross profit of RMB 108.8 billion, resulting in a gross margin of 56%, up from the previous year [1][6][10]. - The increase in gross profit was attributed to high-quality advertising inventory and improved content structure in gaming [10]. Gaming Business - The value-added services revenue reached RMB 95.9 billion, a 16% year-on-year increase, with strong contributions from social networking and gaming [2][8]. - Key titles like "Delta Force" and "Honor of Kings" maintained strong daily active users (DAU), with "Honor of Kings" achieving a historic high of 139 million DAU [2][8]. Marketing Services - Marketing services revenue grew by 21% year-on-year to RMB 35.8 billion, driven by the expansion of core ecosystem scenarios like video accounts and mini-programs [3][9]. - AI-driven improvements in eCPM and automated ad placement through AIM+ contributed significantly to revenue growth [27][29]. Financial Technology and Enterprise Services - Financial technology and enterprise services revenue reached RMB 58.2 billion, a 10% year-on-year increase, supported by strong online and offline payment activities [3][44]. - The company maintained a low bad debt ratio in its consumer credit business, reflecting its strong risk management capabilities [44][46]. Future Outlook - The report anticipates continued high-quality growth in advertising and gaming, driven by the recovery of advertising budgets and the ongoing expansion of the ecosystem [40][55]. - The integration of AI in advertising is expected to enhance efficiency and contribute to a new phase of sustainable growth [40][55].

Investment Rating - The report does not explicitly state an investment rating for the industry [1]. Core Insights - The TMT sector experienced a decline in the week of November 10-14, with the overall TMT sector underperforming the CSI 300 index. The media sector fell by 1.36%, computer by 3.03%, electronics by 4.77%, and communications by 4.77%. The top-performing sub-industries included portal websites (2.39%), communication engineering and services (1.57%), and panels (0.69%). Conversely, passive components (-9.72%), printed circuit boards (-9.15%), and communication network devices (-6.98%) saw significant declines [3][5]. - Recent updates in large models include OpenAI's release of the GPT-5.1 series, which focuses on enhancing interaction experience and reasoning efficiency. The domestic model Kimi K2 Thinking, released by Moonlight Dark Side, supports continuous multi-round tool calls and autonomous reasoning, achieving state-of-the-art performance in various benchmarks [3][5]. - The report highlights a significant acceleration in the iteration of domestic open-source models compared to overseas closed-source competitors, suggesting a narrowing performance gap and an optimistic outlook for domestic models to capture market share due to their cost-effectiveness [3][5]. Weekly Market Review - The TMT sector's performance for the week was as follows: Media (-1.36%), Computer (-3.03%), Electronics (-4.77%), and Communication (-4.77%), with the overall TMT sector underperforming the CSI 300 index, which fell by 1.08% [3][5]. - The top three gainers in the electronics sector were ShenGong Co. (38.03%), BaiBang Technology (20.89%), and Victory Precision (19.24%). In the computer sector, the top gainers were Haixia Innovation (33.10%), Yingfang Software (19.30%), and Hengyin Technology (17.94%). The media sector saw gains from ZhiDeMai (16.31%), MingChen Health (14.91%), and TianXiaXiu (10.47%). The communication sector's top gainers included Online and Offline (46.55%), ZhongFuTong (30.69%), and MeiLiXin (11.87%) [3][5][12][13][6]. Industry News and Key Company Announcements - Significant events in the industry include the release of Baidu's Wenxin large model 5.0, which utilizes a multimodal ultra-sparse mixed expert architecture with over 2.4 trillion parameters. Additionally, the global AI model ranking system LMArena has ranked models from Anthropic, GPT-5, and China's ZhiPu GLM-4.6 as the top three in programming and code generation [15][18]. - Baidu's new Kunlun chips M100 and M300 were announced, targeting large-scale AI inference and multimodal model training, expected to launch in early 2026 and 2027, respectively [21][22]. - The report notes that Tencent's president stated the company does not perceive itself as lagging in the AI race, emphasizing the competitive landscape among domestic models [15][18].

Core Viewpoint - The article discusses Tencent's recent financial performance and its strategic focus on AI, highlighting the company's strong position in the gaming, social media, and music sectors, while also addressing the implications of its agreement with Apple regarding in-app payment fees [2][3][4][5]. Financial Performance - Tencent reported Q3 2025 revenue of 192.87 billion yuan, a 15% year-on-year increase, and a non-IFRS operating profit of 72.57 billion yuan, up 18% year-on-year [2]. - The market reacted coolly to the financial results, with Tencent's stock price at 646.5 HKD, down 1.45% [2]. Gaming Sector Insights - The game "Delta Action" achieved over 30 million monthly active users in September, peaking at 35 million during the National Day holiday, indicating Tencent's dominance in the domestic DAU shooting game market [3]. - The game represents a successful attempt to adapt a AAA game to mobile, enhancing Tencent's industrial pipeline and potentially positioning it as the second-largest game in China in the coming years [3]. Agreement with Apple - Tencent and Apple reached an agreement on a 15% revenue share for iOS mini-games, resolving previous uncertainties regarding payment processing [3][4]. - This agreement simplifies payment processes, which is expected to improve user retention and conversion rates for paid services [4]. AI Strategy - AI has been a focal point in Tencent's recent financial reports, with the company stating it has sufficient GPU resources and is not facing a decisive model in the Chinese market [5]. - Tencent's approach to AI is pragmatic, focusing on efficiency rather than cost-cutting, and the company continues to develop the "intelligent agent" capabilities within WeChat, aiming to transform it into a comprehensive smart ecosystem [5].

经营业绩： 重大事项： 上海实业控股(00363)：上海华氏及上实城开上海大健康管理拟7359.8万元出售上海上实医疗美容医院 49%股权 恒瑞医药(01276)：瑞维鲁胺片获得药物临床试验批准通知书 康希诺生物(06185)：吸入用结核病加强疫苗于印度尼西亚启动I期临床试验并完成首例受试者入组 中慧生物-B(02627)：四价流感病毒亚单位疫苗(佐剂)及三价流感病毒亚单位疫苗(佐剂)的I期临床试验启 动 智云健康(09955)与京东健康(06618)订立服务合作协议 拓宽莉芙敏片®及其他P2M管线产品的线上销售 与服务渠道 华润置地(01109)拟配售华润万象生活(01209)股份 推动股东结构多元化 首程控股(00697)：北京机器人基金投资星动纪元 京东健康(06618)发布第三季度业绩 经营盈利12.43亿元 同比增加125.3% 中芯国际(00981)公布第三季度业绩 公司拥有人应占盈利约1.92亿美元 环比增长44.7% 同比增长28.9% 哔哩哔哩-W(09626)第三季度净利润4.69亿元 同比扭亏为盈 各项核心用户指标均创历史新高 中国太保(02601)：太平洋人寿保险前10个月原保险保费收 ...

Core Viewpoint - Tencent's President Liu Chiping stated that the company does not consider itself to be lagging in the AI competition, emphasizing that all companies are closely matched in this race [1] Group 1 - Liu highlighted that the next generation of Tencent's models will demonstrate significant improvements in mixed capabilities [1] - He mentioned that no single model in the domestic market has a decisive advantage, indicating a competitive landscape where different models may excel in various scenarios [1]

Core Insights - Tencent Holdings reported a solid growth in revenue and profit for Q3 2025, with total revenue reaching 1928.69 billion RMB, a year-on-year increase of 15% and a quarter-on-quarter increase of 5% [1] - The company's net profit attributable to equity holders was 631.33 billion RMB, reflecting a year-on-year growth of 19% and a quarter-on-quarter growth of 13% [1] Revenue Breakdown - The value-added services segment saw a revenue increase of 16% year-on-year, totaling 959 billion RMB, driven by new game releases and growth in established titles [2] - Domestic game revenue was 428 billion RMB, up 15% year-on-year, while international game revenue surged by 43% to 208 billion RMB, primarily due to contributions from Supercell games and new studio acquisitions [2] - Social network revenue grew by 5% to 323 billion RMB, supported by increased income from live streaming services and music subscriptions [2] Marketing and Financial Services - Marketing services revenue increased by 21% year-on-year to 362 billion RMB, attributed to higher ad exposure and improved user engagement [2] - Financial technology and enterprise services revenue rose by 10% to 582 billion RMB, with growth driven by commercial payment activities and cloud service demand [3] Overall Performance - For the first nine months of 2025, Tencent achieved a total revenue of 5573.95 billion RMB, a 14% increase year-on-year, with a net profit of 1665.82 billion RMB, reflecting a 17% growth [3] - As of September 30, 2025, WeChat and QQ reported active user numbers of 1.414 billion and 517 million respectively, with Tencent Video and Tencent Music maintaining leading positions in their respective markets [3] Strategic Investments - The company’s strategic investments in AI have enhanced precision in advertising and user engagement in gaming, while also improving efficiency in programming and video production [4] - Tencent is continuously upgrading its foundational models and technology architecture, with advancements in image and 3D generation models [4]

Core Viewpoint - Tencent is leveraging AI tools like Hunyuan and VISVISE to revolutionize game development, significantly reducing production time and driving strong revenue growth, positioning itself as a leader in the smart entertainment era [3][4][10]. AI Leadership: Rapid Innovation in Development and Distribution - Tencent's AI capabilities are not theoretical but are reshaping game development on an industrial scale, with tools like VISVISE automating the generation of textures and animations, compressing development time to minutes [6][7]. - The Hunyuan Game AI can generate hyper-realistic characters in seconds, democratizing AAA game development for independent developers and Tencent's studios [6][8]. Global Expansion: From Shenzhen to the World - Tencent is expanding its global footprint as the second-largest game publisher, with international game revenue growing by 35% in Q2 2025, driven by AI innovations [8][9]. - Strategic investments, such as in Antix, are enhancing AI digital human technology, significantly reducing production costs in film and supporting Tencent's metaverse strategy [9]. Long-term Value: Core Asset in the AI-Game Era - Tencent is projected to achieve a revenue of approximately 755 billion RMB by 2025, with a compound annual growth rate (CAGR) of 13%, driven by AI and global expansion [12]. - The company is expected to maintain a robust return on equity (ROE) of 15%+, potentially increasing to 20%-30%+ by 2030, supported by its data-driven business model [12]. Conclusion - Tencent is not just adapting to the AI revolution; it is a key architect of this transformation in the gaming industry, providing developers with liberation, players with immersion, and investors with a solid investment opportunity [13].

Summary of Conference Call Records Industry Overview - The conference call focuses on the AI model landscape in China, highlighting the challenges and advancements in the domestic AI industry compared to international counterparts [1][2][4][5]. Key Points and Arguments 1. **Architecture and Innovation** - Domestic AI models heavily rely on overseas architectures like Transformer and MoE, leading to difficulties in surpassing foreign models [1][2]. - There is a lack of self-developed, breakthrough architectural innovations in China, which hampers competitiveness [2]. 2. **Computational Power** - Chinese AI companies have significantly lower GPU computational power compared to international giants like Microsoft, Google, and Meta, often by an order of magnitude [2]. - The ongoing US-China trade war has restricted resource availability, further impacting computational capabilities [1][2]. 3. **Cost and Performance Focus** - Domestic models prioritize inference cost and cost-effectiveness, aligning with local consumer habits, while international models like GPT focus on top-tier performance [1][2]. - The commercial model differences create a substantial gap in model capabilities [2]. 4. **Data Acquisition** - The relatively lenient data laws in China provide an advantage in data acquisition for training models, unlike the stringent regulations in Europe and the US [3]. 5. **Open Source Strategies** - Alibaba adopts a nearly fully open-source strategy, including model weights, code, and training data, to enhance influence and integrate its cloud services [4]. - Other companies like ByteDance and Kuaishou are more selective in their open-source approaches due to their reliance on proprietary technology [4]. 6. **Multimodal Model Developments** - Domestic companies are making strides in multimodal models, focusing on applications in e-commerce and short videos, which cater to local needs [5][6][7]. - Companies like Alibaba, Kuaishou, Tencent, and ByteDance are developing models that integrate text, image, audio, and video generation [7][8]. 7. **MoE Architecture Adoption** - The MoE architecture is becoming standard among major companies, allowing for reduced computational costs and inference times [10]. - Future optimization directions include precise input allocation, differentiated expert system structures, and improved training stability [10][11]. 8. **Economic Viability of Large Models** - Starting mid-2024, pricing for APIs and consumer services is expected to decrease due to the release of previously constrained GPU resources [13]. - The overall cost conversion rate in the large model industry is increasing, despite initial low profit margins [13][14]. 9. **Competitive Differentiation** - Key competitive differences among leading domestic firms will emerge from their unique strategies in technology iteration, data accumulation, and business models [15]. 10. **Future Trends and Innovations** - The focus will shift towards agent systems that integrate user understanding and tool invocation, enhancing overall efficiency [16]. - The MCP concept will gain traction, addressing data input-output connections and reducing integration costs [22]. Additional Important Insights - The acceptance of paid services among domestic users is low, with conversion rates around 3% to 5%, indicating a need for improved user experience to enhance willingness to pay [20][21]. - Successful AI product cases include interactive systems that combine companionship with professional analysis, indicating a potential path for monetization [22]. This summary encapsulates the critical insights from the conference call, providing a comprehensive overview of the current state and future directions of the AI industry in China.

Core Viewpoint - Goldman Sachs expresses increased confidence in Tencent Holdings' (00700) outlook, believing that Tencent's investments are deepening its competitive moat, and with its unique WeChat ecosystem and global gaming assets, it could become one of the most benefitted companies in the Chinese internet sector from AI application developments, assigning a "Buy" rating with a target price of HKD 701 [1] Group 1 - Tencent recently held the 2025 Tencent Global Digital Ecosystem Summit in Shenzhen, where management emphasized that efficiency improvements driven by intelligence and scale expansion driven by globalization are the two main growth engines for Tencent Cloud and its ecosystem partners [1] - Over the past year, Tencent has iterated its foundational models, releasing more than 30 mixed models, and anticipates that AI applications in gaming, marketing services, native applications, and internal productivity enhancements could drive geometric business growth [1] - Management indicated that Tencent Cloud will strengthen its expansion in overseas markets, with an average annual compound growth rate of high double digits in overseas market revenue over the past three years, and the number of international clients has doubled year-on-year [1]