策 略 研 究 投资要点： 风险提示：增长政策落地进度不及预期，国内经济修复不及预期。 年内继续看好港股的三大理由 [Table_Authors] 本报告导读： ①近期 AH 溢价创近六年新低反映港股市场流动性无虞，近期股指走弱主要受互联 网权重板块结构性拖累。②展望未来，我们认为年内需重视港股三大因素：AI 领域 技术突破催化科技成长、美联储降息背景下外资可能超预期，南下增配力量仍有较 大空间。③受益于资产稀缺性优势，港股市场望持续吸引增量资金入市助推行情向 上，结构上重视本轮产业周期中弹性更大的恒生科技。 [Table_Report] 策略研究 /[Table_Date] 2025.08.23 2025-08-24 请务必阅读正文之后的免责条款部分 策 略 研 究 海 外 证 券 研 究 报 告 [Table_Summary] 6 月以来港股指数表现弱于 A 股，但 AH 溢价指数创近六年新低。6 月中旬以来在 A 股指数不断向上创新高之际，港股行情表现却偏震 荡。尽管港股指数表现弱于 A 股，但我们观察到恒生 AH 股溢价指 数较 6/19 的 131.54 进一步下行至 8/15 的低点 122.6， ...

Group 1: US Stock Market Strategy - Recent hawkish comments from the Federal Reserve Chairman and unexpectedly weak non-farm employment data have interrupted the upward trend of the US stock market[1] - Approximately 75% of manufacturers and service providers plan to raise product prices within three months due to increased tariff costs[1] - The upcoming July inflation data is a critical observation point; a rebound could exacerbate market volatility[1] Group 2: Economic Outlook - The US stock market is expected to remain in a consolidation phase in the short term, but medium to long-term earnings growth is anticipated to support upward movement[1] - The Federal Reserve's future policy path is uncertain, with significant fluctuations in expectations for a rate cut in September[1] Group 3: Vietnam Economic Performance - Vietnam's GDP growth rate for Q2 reached 7.96%, supported by strong performance in the industrial and construction sectors, which grew by 8.97%[3] - Foreign Direct Investment (FDI) registered a total of approximately $21.5 billion by the end of Q2, marking a 41.7% increase compared to the same period in 2024[3] Group 4: Company Reports - Tech company Unity Software's advertising network revenue increased by 15% due to the Unity Vector AI platform, prompting an upward revision of revenue forecasts for 2025-2027[6] - Tech company Techtronic Industries' mid-term performance met expectations, with revenue forecasts for 2025-2027 set at $15.637 billion, $16.992 billion, and $18.422 billion respectively[5]

Core Insights - The US stock market has shifted from a continuous upward trend to a wait-and-see phase, influenced by recent trade agreements and mixed economic signals [3][4][5] - The technology sector has shown strong performance, particularly driven by major companies like Apple and Amazon, while traditional sectors like energy and finance have struggled [5][12] Market Performance - As of August 7, the S&P 500 index has seen a cumulative increase of approximately 2.6% over the last 10 trading days, while the Nasdaq 100 index has performed even better with a rise of about 4.3% [4][6] - The recent performance of the Nasdaq is largely attributed to positive developments from tech giants, particularly in AI and manufacturing initiatives [5][12] Sector Analysis - The technology sector has outperformed, with significant contributions from companies like Apple, which announced a $100 billion investment in US manufacturing, leading to a 5% increase in its stock price [5][12] - Other AI-related companies such as Nvidia, Microsoft, and Meta have also reported strong earnings, contributing to their stock price increases [5][12] Economic Indicators - The July non-farm payroll data revealed an increase of only 73,000 jobs, significantly below the expected 104,000, leading to concerns about a cooling labor market [12] - The downward revision of previous months' job growth by a total of 258,000 jobs has heightened fears of recession or stagflation [12] Federal Reserve Policy - The Federal Reserve's recent meetings have left interest rates unchanged, but there is speculation about potential rate cuts in September, influenced by inflation and tariff uncertainties [9][15] - The market's expectation for a rate cut has fluctuated, with the probability dropping to 42% following comments from Fed Chair Powell [9][15]

Core Viewpoint - Goldman Sachs predicts a significant acceleration in AI cloud business growth in Q2, driven by surging demand for AI inference and rapid development of AI applications, while competition in the food delivery sector is expected to lead to a substantial decline in platform profits [1][3][4]. Group 1: AI Cloud Business Growth - The AI cloud business of Chinese internet giants is expected to see a notable acceleration in growth during Q2, with Alibaba Cloud's revenue growth forecasted to increase from 18% in Q1 to 23% in Q2 [5][6]. - Despite being lower than Google Cloud's 32% and Microsoft Azure's 39% growth rates, this increase indicates a positive trend driven by the surge in AI inference demand [6]. - In June, ByteDance's daily token processing reached 15 trillion, while Alibaba's was around 4-5 trillion, highlighting the rapid growth in AI inference demand [6][7]. Group 2: Impact of Food Delivery Competition - Intense competition in the food delivery market is significantly impacting the performance of transaction platforms [8]. - As a result, Alibaba's EBITA is expected to decline by 16% year-on-year, while Meituan and JD's EBIT are projected to drop by 58% and 70%, respectively [9]. - Goldman Sachs estimates that Alibaba's food delivery business incurred losses of 11 billion yuan in Q2, which is expected to expand to 19 billion yuan in Q3 [9]. Group 3: Performance Disparities Among Companies - The report anticipates a divergence in performance among major platforms, with gaming and ride-hailing sectors expected to show resilience [11]. - Tencent is projected to report a revenue growth of 11% in Q2, with adjusted EBIT growth of 15%, driven by strong performance in its gaming segment [11][12]. - Meituan is expected to see a revenue increase of 16%, but its adjusted EBIT is forecasted to plummet by 58% due to intensified competition in food delivery [12]. - JD is also expected to grow revenue by 16%, but its adjusted EBIT is anticipated to decline by 70% [12].

Investment Rating - The report does not explicitly provide an investment rating for the cloud computing industry Core Insights - Cloud computing is a critical infrastructure in the AI era, facilitating the deep integration of AI into various industries and driving the digital economy forward [7][8] - The global cloud computing market is projected to reach nearly $2 trillion by 2030, with a significant contribution from AI applications [8][20] - China's cloud computing market is expected to exceed 3 trillion yuan by 2030, driven by the integration of quantum computing, blockchain, and AI [9][45] Summary by Sections Global Cloud Computing Development Overview - Countries are accelerating cloud computing strategies to enhance competitiveness in the AI era, with the US investing over $100 billion in cloud infrastructure since 2018 [14][15] - The global cloud computing market is experiencing stable growth, with a market size of $692.9 billion in 2024, reflecting a year-on-year growth of 20.3% [19][20] - Cloud computing technologies are continuously integrating, becoming the engine for innovative development across various sectors [27] China's Cloud Computing Development Overview - China's cloud computing market reached 828.8 billion yuan in 2024, growing by 34.4% year-on-year, with public cloud services accounting for 621.6 billion yuan [45][46] - The integration of AI with cloud computing is becoming a key driver for market growth, with significant advancements in IaaS and SaaS sectors [48][51] - The dual drive of "cloud + AI" is accelerating the adoption of intelligent applications across industries, with a focus on government and transportation sectors [52][56] Cloud Computing Driving Service Paradigm Innovation - New service models such as AIIaaS, AIPaaS, and AIMSP are emerging, reflecting the evolution of cloud services towards AI integration [9][24] - The concept of "one cloud, multiple calculations" is gaining traction, allowing for efficient resource management and data integration across various computing needs [29][30] - The cloud excellence architecture is helping enterprises optimize their cloud usage, focusing on security, stability, and operational efficiency [34][35] Development Outlook - The report emphasizes the importance of cloud computing in supporting the digital transformation of traditional industries, with a focus on enhancing operational efficiency and service quality [57][58] - The integration of cloud computing with AI technologies is expected to create new business models and enhance the overall productivity of various sectors [36][38]

Summary of Key Points from the Conference Call Company and Industry - **Company**: Alibaba Cloud - **Industry**: Cloud Computing and AI Services Core Insights and Arguments - **Revenue Growth**: Alibaba Cloud expects a revenue growth of 20%-25% year-on-year for FY2025, with total revenue projected at approximately 150 billion RMB [1][5] - **Profit Margin**: The profit margin is anticipated to be between 8%-8.5%, although depreciation and amortization may lead to a slight decline in profit margins [1][5][16] - **AI Cloud Revenue**: In Q2, AI cloud revenue is expected to nearly double year-on-year, reaching between 4 billion to 5 billion RMB, with GPU leasing accounting for 45%-50% of this revenue [1][8] - **Capital Expenditure**: Total capital expenditure for 2025 is expected to exceed 100 billion RMB, with 20-30 billion RMB allocated for overseas operations [4][31] - **Customer Composition**: Approximately 70% of Alibaba Cloud's revenue comes from external customers, with the remaining 30%-32% from internal customers [3] Additional Important Content - **AI Business Growth**: The AI business is expected to see stable growth due to the long project implementation cycles of enterprise clients, with a gradual increase in usage anticipated in the second half of the year [10][11] - **GPU Leasing Margins**: The gross margin for GPU leasing is projected to be between 26%-29%, while AI PaaS and SaaS products have higher margins of 42%-45% and 45%-48%, respectively [1][12] - **Supply Chain Challenges**: Alibaba Cloud plans to procure 150,000 to 200,000 computing cards in 2025, but U.S. restrictions may hinder achieving this target [17][24] - **Market Competition**: The competitive landscape is expected to intensify, with ongoing price wars and a focus on AI services among cloud providers [19][18] - **Stock Levels**: Alibaba Cloud currently has around 140,000 NVIDIA cards in stock, with additional domestic chip inventory [25] Financial Performance Expectations - **Q2 Performance**: Revenue growth for Q2 is projected to be between 19.5%-20.5%, with a quarter-on-quarter growth of 6.5%-7.5% [2] - **Future Orders**: The company is optimistic about maintaining stable growth in the coming quarters, although explosive growth is not anticipated [6][10] Strategic Plans - **Product Mix**: The company plans to increase the proportion of domestic chips in its product mix while continuing to rely on NVIDIA and AMD cards [33][26] - **Cost Control**: Alibaba Cloud is focusing on domestic alternatives and technological iterations to manage costs amid external competition [18] This summary encapsulates the key points discussed in the conference call, highlighting Alibaba Cloud's growth expectations, financial performance, competitive landscape, and strategic initiatives.

Investment Rating - The report assigns a "Neutral" rating for CoreWeave (CRWV.O) based on its current market performance and outlook [4]. Core Insights - CoreWeave announced the acquisition of Core Scientific for approximately $9 billion in an all-stock transaction, aimed at accelerating its AI data center strategy [1]. - Following the announcement, CoreWeave's stock fell by 3% on July 7 and continued to decline by 5% on July 8, while Core Scientific's stock dropped nearly 18% on the same day [1]. - The acquisition is expected to be completed in Q4 2025, pending regulatory and shareholder approvals [8]. Financial Projections - Revenue is projected to grow significantly from $229 million in 2023 to $16.418 billion by 2027, with operating profit turning positive by 2024 [3]. - Adjusted net profit is expected to improve from a loss of $45 million in 2023 to a profit of $1.08 billion in 2027 [3]. - The price-to-sales ratio is forecasted to decrease from 317.5 in 2023 to 4.4 by 2027, indicating improved valuation metrics over time [3]. Transaction Details - The exchange ratio for the acquisition is set at 1:0.1235, with Core Scientific shareholders expected to hold less than 10% of the combined company post-transaction [8]. - The acquisition price of $20.40 per share represents a 66% premium over Core Scientific's stock price prior to acquisition discussions [8]. - CoreWeave aims to leverage Core Scientific's existing 1.3 GW power capacity data centers, with plans to repurpose mining facilities for high-performance computing [8]. Strategic Implications - The acquisition reflects a broader trend where cloud service giants are moving towards owning core energy assets rather than relying on power leasing [9]. - CoreWeave's integration of Core Scientific is anticipated to enhance operational efficiency and reduce future leasing costs, potentially saving over $5 billion in operational costs by 2027 [8]. - The ownership of critical power resources is expected to unlock approximately $5 billion in financing opportunities, which could lower the company's weighted average cost of capital (WACC) [8]. Market Performance - CoreWeave's stock has shown volatility, with a 1-month return of -7% and a 3-month return of 211% [5]. - The market is concerned about the high valuation premium associated with the acquisition, especially given Core Scientific's significant projected losses in 2024 [12]. - Investors are advised to monitor CoreWeave's contract renewals with key clients like Microsoft and OpenAI, as well as its ability to secure financing at lower rates [10].

Core Viewpoint - PPIO, founded by Yao Xin, is preparing for an IPO on the Hong Kong Stock Exchange, aiming to capitalize on the growing demand for distributed cloud computing and AI services [2][5]. Company Overview - PPIO was established in 2018 as an independent distributed cloud computing service provider, with its parent company PPLabs Technology Limited registered in the Cayman Islands in 2019 [5]. - Yao Xin, the founder and CEO, previously founded PPTV, a video platform that raised over $700 million before being sold to Suning in 2014 [5]. - PPIO has received angel investments from notable figures and institutions, including Wang Xiaochuan and Panlin Fund [5]. Financial Performance - PPIO's revenue for 2022, 2023, and 2024 was approximately RMB 286 million, RMB 358 million, and RMB 558 million, respectively, with a compound annual growth rate (CAGR) of 39.7% [11]. - The company reported losses of RMB 850 million, RMB 1.894 billion, and RMB 2.935 billion for the same years, totaling approximately RMB 5.68 billion in cumulative losses [11]. - Adjusted net losses were RMB 39.13 million, RMB 37.08 million, and RMB 61.62 million, totaling around RMB 138 million [12]. Business Segments - PPIO operates primarily in two segments: edge cloud services and AI cloud services, with partnerships including major telecom operators and Alibaba Cloud [8]. - In the fiscal year 2024, revenue from edge node services accounted for 70%, edge CDN for 28.1%, and AI cloud services for 1.9% [12]. Market Potential - The edge cloud computing market in China is projected to reach RMB 37 billion by 2029, with a CAGR of 22.9% from 2024 to 2029 [10]. - The global AI cloud computing market is expected to reach RMB 427.7 billion by 2029, with a CAGR of 68.5% during the same period [10]. - PPIO claims to be one of the earliest independent providers of AI cloud computing services in China, with significant growth in token consumption from 27.1 billion to 141.9 billion between December 2024 and May 2025 [10]. Future Outlook - Yao Xin anticipates a significant reduction in AI inference costs and believes distributed computing will drive the next wave of development [15]. - The company plans to use IPO proceeds to enhance its technology capabilities, expand its market presence, and support overseas growth [16].

Group 1 - The rapid rise of artificial intelligence (AI) is reshaping the global business landscape, becoming a strategic issue that no company can ignore [2][3] - Companies must actively embrace AI to gain a competitive edge and transition from passive adaptation to leading transformation [2] - Chinese companies have shown early advantages in AI practices across various sectors, including smart manufacturing and digital marketing [3] Group 2 - The leading position of Chinese companies is rooted in a complex market environment and a highly digitalized user behavior [3] - As AI moves to the core of business operations, companies face structural challenges that require strategic responses rather than tactical fixes [3][4] - A systematic and structured information perception mechanism is essential for companies to respond to AI transformations [4] Group 3 - Companies should establish an open external knowledge network that includes academia, industry, and venture capital to better capture trends and understand emerging technologies [5] - Successful companies are leveraging corporate venture capital (CVC) to observe innovation mechanisms and adapt to the AI ecosystem [5][6] - The return on investment (ROI) for AI initiatives is crucial, and companies must evaluate the reasonableness of their AI investments [6][7] Group 4 - AI is not a one-time project but a strategic asset that evolves and adds value over time [7][8] - Companies need to adopt a multi-dimensional ROI assessment framework that allows for continuous validation and dynamic adjustments [9] - AI-related strategic decisions should involve cross-departmental collaboration rather than relying on a few individuals [10][11] Group 5 - High-level management must actively participate in AI strategy and foster a culture of shared understanding and capability [11][12] - Companies should create a culture that allows for "smart trial and error," accepting small, manageable failures as part of the learning process [12][13] - The future organization will require a talent pool that integrates data understanding, AI application, and business innovation [13][14] Group 6 - A new breed of AI-native companies will emerge, redefining industry logic and challenging traditional enterprises [14][15] - Companies need the courage and capability for self-disruption to adapt to the changing landscape [15][16] - The core of AI transformation lies in the synchronization of strategic recognition and organizational capability [16][17] Group 7 - The competition in the AI era will focus on systematic learning and adaptability rather than just tool usage [18] - Companies must embrace "AI thinking," which involves a comprehensive restructuring of strategy, organization, processes, and culture [18]

Core Insights - The global AI infrastructure sector has reached a milestone with the rebranding of Junee to SuperX, reflecting a strategic shift towards becoming a one-stop AI infrastructure solution provider focused on enhancing lives with AI [1] Industry Challenges - The demand for AI computing power is experiencing exponential growth, with large-scale clusters facing challenges of high costs, long delivery times, and high energy consumption [2] - For instance, xAI's Colossus cluster, despite achieving rapid delivery in 214 days, consumes 1.3 million gallons of water daily and has an energy load equivalent to that of 100,000 households [2] SuperX's Solutions - SuperX aims to address these challenges through an end-to-end infrastructure solution that integrates cold plate liquid cooling technology (D2C) and high-voltage direct current power supply, reducing total energy consumption by over 10.2% compared to traditional air cooling, and optimizing PUE to industry-leading levels [2] - The construction efficiency has improved significantly, with modular blind plug design allowing for minute-level server hot-swapping, reducing data center construction time by 60% compared to traditional methods [2] - SuperX also emphasizes sovereign AI empowerment by establishing distributed AI factories that comply with data security regulations, mitigating cross-border data flow risks [2] Paradigm Shift in AI - The company is driving a paradigm shift from "computing power accumulation" to "value creation" in the AI industry, promoting the democratization of computing power through a trillion-parameter model fine-tuning and inference applications that are affordable for small and medium enterprises [2] - SuperX's unique "AI Factory-as-a-Result" model aims to deliver AI outcomes, providing microsecond-level ultra-low latency inference for real-time decision-making in manufacturing, healthcare, and other sectors [3] - The company is forming an ecosystem in collaboration with industry partners, encompassing AI chips, servers, AIDC, AI cloud, and final AI applications to promote AI inclusivity [3]