每周 一 景：湖南衡阳衡山 点击右上角菜单，收听朗读版 【 海外策略】"秋日胜春朝"——四季度港股市场展望 上半年港股明显占优， 6 月后表现偏弱， 9 月以来已经重拾升势。 今年上半年，港股表现优异，涨幅在全球市场中居于前列，恒生指数、恒生科技指数最大 涨幅分别达到 33% 、 49% 。 6 月中旬至 8 月底港股股指阶段性表现弱于 A 股，港股科技板块跑输现象更为明显，这一现象或与港股权重科技稀缺资产表 现不佳存在一定关联。一方面，海内外科技巨头持续加码算力资本开支，对算力硬件端提振较大。另一方面阿里、京东和美团在外卖业务加大补贴，科技巨头 盈利增速一致预期持续下修，叠加近期 AI 应用软件端的催化相对较少，使得港股权重科技龙头股表现不佳。进入 9 月以来，伴随阿里巴巴、百度等互联网大 厂在 AI 领域的乐观情绪发酵，互联网叙事逐步转向" AI 赋能"，港股指数表现明显回暖，其中科技板块弹性更大。 当前港股估值其实并不高，科技板块性价比更加突出。 港股自 2024 年初后已处于牛市氛围中，今年以来也维持了震荡上行的格局。尽管港股市场前期积累 的涨幅已经较为可观，但目前估值仍不算高。尤其港股科技板块低估优 ...

Core Insights - The computer industry is experiencing accelerated revenue growth in the first half of 2025, with significant profit improvements and stable growth in Q2 2025 [1][2] - The industry is focusing on cost reduction and efficiency enhancement, with improvements in cash flow management despite a year-on-year decline in operating cash flow [3] - The AI sector is witnessing substantial growth, particularly in computing chips and AI application software, suggesting potential investment opportunities in related core companies [1][5] Growth Perspective - In the first half of 2025, the overall revenue of the computer industry increased by 12.55% year-on-year, with a median revenue growth rate of 3.21% for individual companies [1][2] - The total net profit attributable to shareholders reached 3.522 billion yuan, marking a significant turnaround from losses, with a median net profit growth rate of 7.49% for individual companies [1] - The industry reported a non-recurring net profit of -1.397 billion yuan, indicating a substantial narrowing of losses, with a median growth rate of 6.25% for individual companies [1] Distribution Analysis - 56.59% of companies achieved year-on-year revenue growth, an increase from 53.89% in the previous year, with 54.79% of companies reporting net profit growth [2] - In Q2 2025, the industry revenue grew by 8.10% year-on-year, continuing the recovery trend from Q1, with a median growth rate of 3.20% for individual companies [2] - The total net profit for Q2 2025 increased by 73.31% year-on-year, with a median growth rate of 10.78% for individual companies [2] Quality of Growth - The overall gross margin for the industry in the first half of 2025 was 20.98%, a decline of 2.49 percentage points year-on-year, while the expense ratio decreased to 20.74% [3] - In Q2 2025, the gross margin was 22.01%, with a total operating cash flow of -39.56 billion yuan, reflecting a year-on-year increase of 25.53% [3] - The management of accounts receivable has improved, with the growth rate of accounts receivable lower than that of revenue [3] Sector Performance - Eight sectors reported positive revenue growth in the first half of 2025, with significant growth in the information technology sector, AI, and financial IT [4] - The AI sector's revenue growth median was 10.64% in the first half of 2025, with explosive growth in chip sub-sectors [5][6] - The performance of the AI sector is expected to continue improving, with substantial growth in both chip and application software segments [5][6]

Core Investment Trends - Morgan Stanley identifies three core investment themes for the stock market over the next 12 months: AI computing power leaders and software giants benefiting from the AI wave, semiconductor equipment themes benefiting from favorable policies under the Trump administration, and essential consumer goods leaders amid a fragmented consumption chain [1][2]. AI Computing Power and Software Giants - The demand for AI computing power is experiencing explosive growth, driven by significant investments in AI infrastructure by the U.S. government and tech giants, indicating a bullish outlook for companies like Nvidia, TSMC, and Broadcom [6][10]. - Analysts predict that major tech companies, including Google, Microsoft, Meta, and Amazon, will spend over $350 billion on AI computing infrastructure in 2023, representing a nearly 50% year-over-year growth [6][10]. - By 2025, AI-related capital expenditures in tech companies are expected to reach 28%, up from 12% in 2023, with AI applications driving significant increases in efficiency and productivity [7][11]. Semiconductor Equipment Beneficiaries - The "One Big Beautiful Bill Act" (OBBBA) is expected to significantly boost free cash flow for U.S. manufacturing companies, particularly in the semiconductor equipment sector, as companies shift production back to the U.S. [13][14]. - Semiconductor equipment leaders are positioned to benefit from the unprecedented demand for AI chips, with companies like ASML and Applied Materials playing crucial roles in the manufacturing process [15][16]. Essential Consumer Goods Amid Consumption Fragmentation - The market is witnessing a divergence in performance, with essential consumer goods companies expected to show resilient growth while discretionary spending is under pressure [17][18]. - High-income consumers are less affected by inflation and continue to spend on non-essential items, while low-income consumers are shifting towards cheaper alternatives, leading to a stark contrast in consumption patterns [18].

Group 1 - The core point of the article is that Palantir's stock price increased by 2% to $159.13 after experiencing six consecutive days of decline [1]

Core Viewpoint - A research personnel from a scientific institution misused AI software, leading to the leakage of confidential information, as disclosed by the National Security Department [1] Group 1 - The individual, referred to as Xiao Li, uploaded core data and experimental results to an AI application while preparing a research report, which resulted in a breach of sensitive information in the research field [1] - The National Security Department has taken serious action against Xiao Li following the incident [1]

Core Insights - The latest corporate software spending survey by Bank of America indicates a slight downward adjustment in expected growth for enterprise software spending to approximately 9.9% for 2025, influenced by global tariff policies, yet AI-related software spending remains a top priority in corporate budgets [1][8][14] - The demand for AI applications is expected to drive significant growth in enterprise AI software budgets, with projections showing that AI-related spending will account for 27.7% of software budgets in 2025 and increase to 31.6% in 2026 [14][18] - Companies like Palantir, Nvidia, and AMD are positioned to benefit from the robust demand for AI infrastructure, with Palantir's stock surging over 65% since April, reflecting strong performance in the AI and data analytics sector [3][5][26] Group 1: AI Software Spending Trends - The survey reveals that AI software spending is becoming the fastest-growing investment direction for enterprises, with a focus on enhancing operational efficiency and reducing costs [7][18] - Companies are increasingly prioritizing AI investments in cloud infrastructure and back-office operations, with 60% of respondents indicating plans to invest in cloud AI software [23][25] - The shift in AI spending focus from front-end applications to back-office operations highlights a growing trend towards improving internal efficiencies [23][26] Group 2: Market Performance and Projections - The stock performance of AI-focused companies has been robust, with Nvidia and Broadcom seeing significant price increases, indicating strong market confidence in AI infrastructure [3][5] - The strong earnings reports and optimistic outlooks from AI application software providers like C3.ai and Palantir are driving investor interest and stock price increases [5][6] - Anthropic, a notable player in the AI space, has reported a substantial increase in annual revenue, showcasing the growing demand for generative AI applications in the business world [5][6]

Core Viewpoint - The AI revolution is expected to significantly boost macroeconomic demand through capital expenditure, with explosive growth in AI-related capital spending in the US and China projected for the coming years [1][2][3]. Group 1: AI Capital Expenditure in the US - In the second half of 2023, AI-related capital expenditure in the US has seen explosive growth, with an estimated increase of $861 million by 2025, accounting for approximately 0.3% of US GDP [2]. - The North American cloud service providers are projected to have a total capital expenditure of $3,385 million in 2025, reflecting a 34% increase from 2024 [2]. Group 2: AI Capital Expenditure in China - Following the emergence of DeepSeek, major Chinese companies have ramped up their AI capital expenditures, with Alibaba announcing an investment exceeding 3,800 billion RMB over the next three years, surpassing the total of the past decade [3]. - By 2025, the total capital expenditure increase for representative Chinese firms is expected to reach 1,498 billion RMB, approximately 0.11% of GDP, with optimistic and pessimistic scenarios predicting increases of 2,144 billion and 822 billion RMB, respectively [3]. Group 3: Comparison with Previous Investment Cycles - The current AI capital expenditure cycle is anticipated to resemble the scale of capital expenditures seen during the 2021-2022 new energy investment boom, serving as an additional driving force for traditional economic cycles and gradually enhancing productivity across various industries [4][5]. - Historically, capital expenditure in China has been a lagging variable, typically following improvements in real estate or exports, but the current AI investment cycle is driven by technological iteration, allowing for demand creation ahead of revenue and profit realization [5]. Group 4: Long-term Economic Impact of AI - The AI revolution is expected to increase economic growth by 1-2 percentage points through enhanced total factor productivity, although it may reduce the number of jobs created per unit of GDP growth [62]. - The impact of AI on employment is complex, with historical trends indicating that technological advancements often lead to both job displacement and creation, but recent studies suggest that the displacement effect may be more pronounced in the current AI context [62][69]. Group 5: Industry Beneficiaries - Industries such as information technology, education, finance, manufacturing, and healthcare are projected to benefit the most from AI integration, characterized by high AI penetration and elasticity [80][82]. - The highest AI penetration is observed in information services, professional technology, and educational services, while agriculture and utilities show the lowest penetration [82].