Quantitative Models and Construction Methods 1. **Model Name**: Securities Firms' Golden Stock Performance Enhancement Portfolio - **Model Construction Idea**: The model aims to optimize the selection of stocks from the securities firms' golden stock pool to outperform the benchmark, which is the median of equity-biased hybrid fund indices. The model leverages a multi-factor approach to select stocks with high alpha potential while controlling for deviations in individual stocks and style factors from the golden stock pool [39][43]. - **Model Construction Process**: - The securities firms' golden stock pool is used as the stock selection universe and constraint benchmark. - A multi-factor model is applied to further optimize the selection of stocks from the pool. - The portfolio is constructed by controlling the deviation of individual stocks and style factors from the golden stock pool. - The industry allocation is based on the distribution of all public funds [43]. - **Model Evaluation**: The model demonstrates strong alpha generation potential and consistently outperforms the equity-biased hybrid fund index. It reflects the research strength of securities firms and their ability to capture market trends effectively [43]. --- Model Backtesting Results 1. **Securities Firms' Golden Stock Performance Enhancement Portfolio** - **Absolute Return (Monthly)**: -0.55% (2025/09/01 - 2025/09/30) [42] - **Excess Return (Monthly)**: -3.50% relative to equity-biased hybrid fund index (2025/09/01 - 2025/09/30) [42] - **Absolute Return (Year-to-Date)**: 33.26% (2025/01/02 - 2025/09/30) [42] - **Excess Return (Year-to-Date)**: 1.19% relative to equity-biased hybrid fund index (2025/01/02 - 2025/09/30) [42] - **Ranking in Active Equity Funds (Year-to-Date)**: 43.07% percentile (1494/3469) [42] - **Historical Performance (2018-2025)**: - Annualized Return: 19.34% - Annualized Excess Return: 14.38% relative to equity-biased hybrid fund index - Consistently ranked in the top 30% of active equity funds each year [44][47] --- Quantitative Factors and Construction Methods 1. **Factor Name**: Intraday Return - **Factor Construction Idea**: Measures the return generated within a single trading day to capture short-term price movements [27][28]. - **Factor Evaluation**: Demonstrated strong performance in the most recent month [27][28]. 2. **Factor Name**: BP (Book-to-Price Ratio) - **Factor Construction Idea**: Reflects the valuation of a stock by comparing its book value to its market price [27][28]. - **Factor Evaluation**: Performed well in the most recent month but underperformed year-to-date [27][28]. 3. **Factor Name**: Volatility - **Factor Construction Idea**: Measures the degree of variation in a stock's price over a specific period, capturing risk and uncertainty [27][28]. - **Factor Evaluation**: Showed strong performance in the most recent month but underperformed year-to-date [27][28]. 4. **Factor Name**: Total Market Capitalization - **Factor Construction Idea**: Represents the total market value of a company's outstanding shares, often used to gauge company size [27][28]. - **Factor Evaluation**: Underperformed in the most recent month but performed well year-to-date [27][28]. 5. **Factor Name**: SUE (Standardized Unexpected Earnings) - **Factor Construction Idea**: Measures the deviation of actual earnings from expected earnings, standardized by the standard deviation of past earnings surprises [27][28]. - **Factor Evaluation**: Underperformed in the most recent month [27][28]. 6. **Factor Name**: Single-Quarter Earnings Surprise - **Factor Construction Idea**: Captures the magnitude of earnings surprises in a single quarter [27][28]. - **Factor Evaluation**: Underperformed in the most recent month but performed well year-to-date [27][28]. 7. **Factor Name**: Single-Quarter Revenue Growth - **Factor Construction Idea**: Measures the growth in revenue over a single quarter, reflecting a company's sales performance [27][28]. - **Factor Evaluation**: Performed well year-to-date [27][28]. 8. **Factor Name**: Analyst Net Upward Revision - **Factor Construction Idea**: Tracks the net number of upward revisions in analysts' earnings estimates for a stock [27][28]. - **Factor Evaluation**: Performed well year-to-date [27][28]. 9. **Factor Name**: Expected Dividend Yield - **Factor Construction Idea**: Represents the expected annual dividend payments as a percentage of the stock price [27][28]. - **Factor Evaluation**: Underperformed year-to-date [27][28]. --- Factors' Backtesting Results 1. **Intraday Return Factor** - **Recent Month Performance**: Strong [27][28] - **Year-to-Date Performance**: Not specified [27][28] 2. **BP Factor** - **Recent Month Performance**: Strong [27][28] - **Year-to-Date Performance**: Weak [27][28] 3. **Volatility Factor** - **Recent Month Performance**: Strong [27][28] - **Year-to-Date Performance**: Weak [27][28] 4. **Total Market Capitalization Factor** - **Recent Month Performance**: Weak [27][28] - **Year-to-Date Performance**: Strong [27][28] 5. **SUE Factor** - **Recent Month Performance**: Weak [27][28] - **Year-to-Date Performance**: Not specified [27][28] 6. **Single-Quarter Earnings Surprise Factor** - **Recent Month Performance**: Weak [27][28] - **Year-to-Date Performance**: Strong [27][28] 7. **Single-Quarter Revenue Growth Factor** - **Recent Month Performance**: Not specified [27][28] - **Year-to-Date Performance**: Strong [27][28] 8. **Analyst Net Upward Revision Factor** - **Recent Month Performance**: Not specified [27][28] - **Year-to-Date Performance**: Strong [27][28] 9. **Expected Dividend Yield Factor** - **Recent Month Performance**: Not specified [27][28] - **Year-to-Date Performance**: Weak [27][28]

中兴通讯（000063.SZ） 优于大市 证券研究报告 | 2025年10月09日 致力于"成为网络连接和智能算力的领导者"，有望受益 AI 发展。（1）算 力方面：大模型持续迭代，AI 需求持续增长，全球各大 CSP 云厂均加大 AI 算力基础设施投入，预计 2025 年，海外亚马逊、谷歌 、微软 、Meta、甲 骨文等厂商合计 Capex 超过 3700 亿美元，同比增幅超 60%；国内字节、腾讯、 阿里、百度等 Capex 有望超过 3600 亿元，中国移动预计 2025 年在算力领域 投入 373 亿元，电信和联通在算力领域投入同比分别增长 22%和 28%。根据 通信产业网数据，中兴通讯服务器在电信运营商市场份额占比 21%，名列第 一。此外公司已与字节、阿里、百度等云厂展开深度定制的服务器合作。（2） 连接方面：运营商持续布局 5G-A 网络，近期工信部给三大运营商发放卫星 通信牌照，未来面向 6G 通信网络蓄势待发。根据 Delloro 和 Omdia 咨询机 构数据，中兴通讯无线通信市场竞争力仍然位列全球头部（仅次于华为/爱 立信/诺基亚）。此外，5G 网络拉动的 FWA(Fixed Wire ...

========= - The "券商金股业绩增强组合" (Broker Gold Stock Performance Enhanced Portfolio) aims to outperform the median of public funds by optimizing the broker gold stock pool[39] - The portfolio uses the偏股混合型基金指数 (Equity-biased Hybrid Fund Index) as its benchmark, with a 90% position last month[39] - The absolute return of the portfolio for the month (2050901-20250930) was -0.55%, with an excess return of -3.50% relative to the偏股混合型基金指数[42] - The absolute return for the year (20250102-20250930) was 33.26%, with an excess return of 1.19% relative to the偏股混合型基金指数[42] - The portfolio ranked in the 43.07% percentile among active equity funds this year (1494/3469)[42] - The annualized return of the portfolio from 2018.1.2 to 2025.6.30 was 19.34%, with an annualized excess return of 14.38% relative to the偏股混合型基金指数[44] - The portfolio consistently ranked in the top 30% of active equity funds each year from 2018 to 2025[44] - The portfolio's performance statistics for each year from 2018 to 2025 are detailed in Table 6[47] - The portfolio's historical performance is illustrated in Figure 12[46] =========

Macro and Strategy - The fixed income investment strategy indicates that while the fundamentals are improving, verification is still needed, and there are phase-specific opportunities in the bond market [12][19] - The September PMI data shows a slight recovery in manufacturing PMI to 49.8% and a decrease in non-manufacturing PMI to 50.0% [13] - The report highlights a significant increase in gold prices, nearing $4000 per ounce, driven by long-term logic and short-term triggers [13] Industry and Company - The manufacturing growth report notes that OpenAI plans to invest up to $100 billion in collaboration with Nvidia, while Kepler has secured several billion-dollar humanoid robot orders [23] - The petrochemical industry investment strategy emphasizes a stable growth plan, with a target of over 5% annual growth in industry value added from 2025 to 2026 [28][29] - The report on the agricultural products sector indicates a rapid decline in egg prices and an expected acceleration in beef prices in Q4 [10] - The AI infrastructure segment is highlighted, with Nvidia and OpenAI planning to deploy at least 10GW of Nvidia systems, reflecting strong demand for AI computing power [24] - The report discusses the ongoing development of humanoid robots, with multiple companies receiving large orders, indicating a shift towards commercialization [23][25] Investment Strategies - The investment strategy for the computer industry suggests a focus on AI applications, with significant capital expenditures from major companies [10] - The report recommends monitoring the supply chain for humanoid robots and AI infrastructure, emphasizing companies with strong positions in these sectors [26][27] - The chemical industry strategy highlights the importance of optimizing supply and demand, with a focus on high-quality development and the elimination of outdated capacities [29][30]

截至 10 月 8 日 18 时，国庆档票房 18.08 亿，同比预计略有下滑；其中，观影人次/平均票价分别为 4933.5 万次/36.6 元，分别同比-5.3%/-9.4%，票价下滑为本次档期票房下滑的主要原因。 国信传媒观点：1）国庆档票房同比略有下滑。截至 10 月 8 日 18 时，国庆档票房为 18.08 亿元，相比 2024 年国庆档下滑 14.1%。其中，观影人次/平均票价分别为 4933.5 万次/36.6 元，分别同比-5.3%/-9.4%；2） 从单日票房来看，本次国庆档票房呈现先抑后扬的趋势。10 月 1 日至 10 月 6 日单日票房分别同比下滑 35.0%/22.5%/18.2%/11.5%/12.5%/14.5%至3.21/2.89/2.56/2.39/2.09亿元，10月7日票房同比增长20.8% 至 1.86 亿元，票房表现先抑后扬可能与国庆假期多出一天有关；3）《志愿军》《731》等影片表现良好。 国庆档期票房前三的影片分别为《志愿军：浴血和平》《731》《刺杀小说家 2》，档期票房分别为 4.43、 3.41、2.91 亿元，占档期整体票房比重分别为 24.5%、18.9 ...

Investment Rating - The investment rating for Kuaishou-W (01024.HK) is "Outperform the Market" (maintained) [2][7]. Core Insights - OpenAI's release of the Sora 2 video generation model and the Sora App is expected to accelerate the demand for AI-generated content (AIGC) [3][6]. - Sora 2 features significant advancements in image resolution, style control, and instruction adherence, integrating audio-video synchronization technology to enhance physical simulation accuracy [4]. - The Sora App, designed similarly to TikTok, allows users to create AI-generated videos up to 10 seconds long and includes a "Cameo" feature for digital avatars, which could create new monetization opportunities for IP holders [5][6]. Financial Forecast and Valuation - The projected financials for Kuaishou include: - Revenue growth from CNY 660.26 billion in 2024 to CNY 865.93 billion in 2027, reflecting an annual growth rate of approximately 8% to 11% [19]. - Net profit is expected to increase from CNY 194.07 billion in 2024 to CNY 268.42 billion in 2027, with a net profit growth rate of 8% to 18% [19]. - The adjusted EPS is forecasted to rise from CNY 24.15 in 2024 to CNY 34.12 in 2027 [19]. - Key financial ratios include: - ROE is projected to decrease from 20% in 2024 to 17% in 2027 [19]. - The gross margin is expected to improve from 53% in 2024 to 58% in 2027 [19].

Investment Rating - The report maintains an "Outperform" rating for the artificial intelligence industry [1] Core Insights - OpenAI's hardware and computing ecosystem is recognized as a key narrative in AI, with significant investments aimed at building a robust infrastructure, including a $500 billion plan to develop 10GW of AI capacity over four years [2][8] - OpenAI's revenue is projected to grow significantly, with estimates of $13 billion in 2025 and $200 billion by 2030, driven by explosive growth in token consumption and user engagement [2][22] - The transition of ChatGPT from a tool-based application to an agent platform reflects OpenAI's ambition to capture various market segments, including information retrieval, e-commerce, and social media [2][30] Summary by Sections 01 OpenAI Hardware and Computing Ecosystem - OpenAI's "Stargate" project aims to invest $500 billion in AI infrastructure, with major contributions from SoftBank, Oracle, and others [8] - NVIDIA has committed $100 billion to secure chip supply for future data centers, while AMD has exchanged stock options for GPU deployment rights [8][12] - The collaboration with hardware partners has led to significant stock price increases for these companies, indicating strong market confidence [15][16] 02 OpenAI Software Ecosystem Ambitions and Progress - ChatGPT is evolving into an agent platform, enhancing user interaction and decision-making capabilities [2][30] - The introduction of features like ChatGPT Pulse and instant checkout demonstrates OpenAI's focus on integrating AI into everyday user experiences [30][41] - User engagement metrics show a rapid increase, with over 700 million weekly active users projected by mid-2025 [63] 03 ChatGPT Functionality, Users, and Commercialization Analysis - ChatGPT's user base is expected to reach nearly 1 billion by the end of 2025, with a high retention rate for paid subscriptions [22][63] - The platform's primary functions include practical guidance, information retrieval, and writing assistance, with a notable shift towards non-work-related applications [64][66] - The commercial model is evolving from subscription-based to transaction-based revenue, indicating a significant potential for monetization [41][64] 04 Model Performance Comparison - OpenAI's models, particularly the GPT series, maintain a competitive edge in performance, with GPT-5 expected to further enhance capabilities [2][19] - The computational requirements for training new models have increased significantly, reflecting the growing complexity of AI tasks [19][21]

Investment Rating - The investment rating for Kuaishou-W (01024.HK) is "Outperform the Market" (maintained) [2][7]. Core Insights - OpenAI's release of the Sora 2 video generation model and the Sora App is expected to significantly boost the demand for AI-generated content, creating new opportunities in the social media landscape [3][6]. - Sora 2 features advancements in image resolution, style control, and instruction adherence, integrating audio-visual synchronization to enhance realism in generated videos [4][5]. - The Sora App, designed similarly to TikTok, allows users to create and share AI-generated videos, fostering a new social ecosystem around AI content creation [5][6]. Summary by Sections Company Overview - Kuaishou is positioned as a leader in the AI video sector, with its AI-generated video capabilities showing strong performance and continuous improvement [6]. Market Dynamics - The introduction of Sora 2 is anticipated to reshape the short video ecosystem, potentially increasing the proportion of AI-generated content and narrowing the gap with platforms like Douyin [6]. Financial Projections - The financial outlook includes projected revenues of 660,257 million in 2024, increasing to 865,930 million by 2027, with a net profit growth rate of 68% in 2024 [19].

Investment Rating - The report maintains an "Outperform the Market" rating for the electronic industry [1][10][37]. Core Insights - The report emphasizes the valuation expansion of domestic wafer foundries and predicts a price increase cycle for storage driven by heightened demand [1]. - The NAND market is expected to experience a simultaneous increase in volume and price due to rising AI inference demand, with a forecasted average price increase of approximately 5-10% for NAND Flash products by Q4 2025 [2]. - China's wafer fab capacity is projected to grow at a CAGR of 8.1% from 2024 to 2028, significantly outpacing the global average of 5.3% [3]. - The completion of the shareholding reform of Changjiang Storage Technology Co., Ltd. marks a significant milestone, with plans for expansion expected to boost orders for domestic semiconductor equipment companies [4]. - Alibaba Cloud is accelerating its transformation into a full-stack AI service provider, indicating a strong growth trajectory in AI infrastructure investments [7]. Summary by Sections Market Performance - The Shanghai Composite Index rose by 0.21%, while the electronic sector increased by 3.51%, with semiconductors leading at a 7.64% rise [11]. NAND Market Outlook - The NAND Flash market is experiencing a recovery in prices due to increased enterprise-level demand and supply constraints, with a forecasted price increase of 5-10% [2]. Wafer Fab Capacity Growth - China's wafer fab capacity is expected to grow at a CAGR of 8.1% from 2024 to 2028, with mainstream node capacity projected to increase from 25% to 42% of global capacity [3]. Company Developments - Changjiang Storage's shareholding reform is complete, and its new phase of expansion is anticipated to enhance domestic semiconductor equipment orders [4]. - Alibaba Cloud is investing heavily in AI infrastructure, with a three-year plan involving 380 billion yuan [7]. Key Investment Recommendations - The report recommends focusing on companies such as 德明利, 江波龙, and 中芯国际 due to their strong market positions and growth potential [1][2][3][4][7].

Core Insights - The report highlights resilient growth in travel demand during the National Day holiday, with a projected year-on-year increase of 6.2% in cross-regional mobility from October 1 to 8, 2024, surpassing the pre-holiday forecast of 3.2% [4][24] - The consumer services sector underperformed the broader market in September, with the A-share consumer services sector declining by 2.69%, lagging behind the CSI 300 by 5.89 percentage points [4][5] - The report emphasizes the importance of policy support for service consumption, with various measures aimed at enhancing service sector growth and consumer spending [24][25] Subsector Insights Travel Chain Sector - The travel chain sector is expected to benefit from a shift towards experience-driven consumption, with a focus on hotel and retail experiences as the market enters a low season in Q4 [4][24] - Notable performance was observed in specific regions and segments, such as a 22% increase in visitor numbers to Changbai Mountain from October 1 to 3 and significant sales at duty-free stores in Sanya [4][24] - The report suggests monitoring the recovery of the duty-free sector and the hotel industry as supply and demand dynamics stabilize [4][24] Chain Dining Sector - The dining sector shows signs of stabilization, with leading brands in tea and fast food performing well despite concerns over the expansion capabilities of traditional dining models [4][24] - The report recommends focusing on high-quality dining leaders that have adjusted their business models to enhance growth potential [4][24] Education and Human Services Sector - The education sector is poised for growth as external uncertainties diminish, with increased demand for vocational training and educational services [4][24] - The report highlights the potential of AI in education, with ongoing advancements in educational technology expected to drive growth [4][24] Investment Recommendations - The report suggests a diversified investment strategy, recommending companies such as Trip.com, Atour, Tongcheng Travel, BOSS Zhipin, and others in the consumer services sector [4][24]