[Table_Page] 投资策略|专题报告 2026 年 3 月 1 日 证券研究报告 [Table_Title] 如何看待港股持续疲软？ ——港股市场策略展望 [Table_Summary] 报告摘要： （1）三月即将迎来解禁小高峰： 11 年年中、15 年下半年、19 年 3 月、 21 年二季度、22 年年中，解禁潮与港股下跌出现在相似时间段。26 年 3 月主要是有色金属、茶饮、汽车、医药等行业的解禁潮，中大型规模 公司将解禁 872 亿港元，高于去年年末的解禁高峰。由于港股通账户并 不能够参与打新，享受稀缺性公司的上市红利，却需要承受限售股解禁 带来的风险，这可能确实是南向资金对于 2026 年港股的主要担忧之一。 （2）港股年报业绩披露期限临近，存在业绩利空压力：财年与自然年一 致的港股公司需要在 3 月 31 日前披露年报业绩初步公告。 （3）次要因素—地缘政治摩擦升级：尽管美伊军事冲突升级，但对港股 冲击或有限，因假期全球风险偏好并未恶化，非美资产普遍走强；全球 资金更多通过增配油气对冲风险而非抛售股票。 ⚫ 何时能够利空出尽？ 观察窗口期：3 月中下旬，上述压制因素高峰过去，可能会利空出尽 ...

Core Insights - The AI industry is transitioning from a capital-intensive model training phase to a monetization phase focused on commercial efficiency as of 2026 [1][13] - Lenovo is emerging as a key player in this new phase, leveraging its extensive installed base and end-to-end delivery capabilities to capitalize on the hybrid AI architecture [1][12] Financial Performance - Lenovo reported revenue of 157.5 billion RMB for the third quarter of the 2025/26 fiscal year, representing an over 18% year-on-year growth, significantly exceeding market expectations [2][14] - Adjusted net profit increased by 36% year-on-year, surpassing market expectations by 32%, with profit growth outpacing revenue growth [2][14] Market Challenges and Strategic Response - The industry is facing challenges such as rising prices of key components like storage chips and concerns over the sustainability of the AI cycle [2][14] - Lenovo has successfully fulfilled its commitment to double-digit revenue growth and profitability, demonstrating resilience and operational strength in a complex environment [2][14] Shift in AI Industry Focus - The focus of the AI industry is shifting from training to inference, with a greater emphasis on cost-effective and user-friendly solutions for daily AI functionality [4][16] - The market is moving away from purely cloud-based AI models due to high bandwidth costs and energy bills, necessitating more economical solutions [5][16] Competitive Landscape - Companies that can cover a broad customer base and possess global supply chain resilience are becoming increasingly favored in the AI commercial ecosystem [6][18] - Lenovo's extensive device installation base positions it as a critical player in the AI monetization phase, where hardware serves as the platform for AI interaction [6][18] Hybrid AI Strategy - Lenovo's hybrid AI strategy is not merely a hardware upgrade but a comprehensive approach to AI deployment, transferring inference costs from the cloud to local devices [7][19] - The integration of local NPU and heterogeneous computing in PCs enhances response speed and privacy for users while significantly reducing AI usage costs [7][19] Revenue Growth in AI Segment - In the third quarter, Lenovo's AI-related revenue grew by 72%, accounting for 32% of total revenue, with AI PC revenue increasing by 39% and AI smartphone revenue surging by 202% [11][22] - The average selling price of AI PCs reached 845 USD, significantly above industry averages, providing structural support for profitability [20][22] Long-term Business Model - The hybrid AI model is seen as a commercially optimal solution, addressing cost and trust issues associated with cloud-based AI [21] - Lenovo's business model is evolving from merely selling devices to offering comprehensive AI capabilities, positioning the company as a steady revenue generator in the AI monetization phase [21][22] Future Outlook - Lenovo aims to leverage its hybrid AI strategy and global supply chain to create sustainable, scalable, and profitable AI business opportunities [12][24] - The company is focused on integrating AI into everyday life and enterprise operations, ensuring long-term returns for shareholders [12][24]

Core Insights - OpenAI's primary challenge is to transform its AI technology into a profitable business model, aiming to quadruple its revenue within a year while facing significant financial pressures [1][3]. Financial Performance - OpenAI's revenue for the previous year was approximately $13 billion, but it anticipates an additional $100 billion in expenses over the next four years [3]. - Currently, about 60% of OpenAI's revenue comes from consumer products, while 40% is derived from enterprise technology [5]. Advertising Strategy - OpenAI has begun to implement advertising within ChatGPT, despite CEO Sam Altman's previous reluctance to do so, viewing it as a last resort for revenue generation [2][6]. - The company is in the early stages of building an advertising sales team and requires significant infrastructure to support this new business model [6][8]. Competition and Market Position - OpenAI faces stiff competition from established companies like Google and Microsoft, as well as emerging competitors like Anthropic, which is focusing on enterprise tools [4][8]. - Anthropic has recently criticized OpenAI's advertising strategy, highlighting the competitive landscape in the AI sector [8]. Enterprise Market Focus - OpenAI aims to increase the proportion of its revenue from enterprise products to 50% by the end of the year, with tools like Codex and ChatGPT Enterprise being key offerings [8]. - The company is encountering challenges in convincing ordinary businesses to pay high fees for its software solutions [8]. Revenue Generation Models - OpenAI is exploring a "value-sharing" model, where it could take a portion of profits from scientific discoveries facilitated by its technology [9][11]. - Concerns have arisen among researchers regarding the potential for OpenAI to claim a share of their discoveries, prompting the company to clarify its stance on this issue [11].

Core Insights - The recent financial reports from major US tech companies highlight the critical focus on the profitability and revenue contributions of their AI businesses, with investors increasingly concerned about the conversion of high investments into actual profits [1][2] Group 1: Financial Performance - Alphabet's cloud business reported a 48% year-over-year growth, achieving $17.664 billion in quarterly revenue, indicating a promising outlook for AI monetization [1] - Despite strong revenue growth, Alphabet's stock fell over 6% following the announcement of a capital expenditure plan for 2026 projected between $175 billion and $185 billion, nearly doubling from 2025 [1] - Amazon's free cash flow decreased by over 70%, yet it plans to increase capital expenditures to $200 billion, exceeding Wall Street expectations by approximately 37%, leading to an 8% drop in its stock price [1] Group 2: Market Sentiment and Investment Concerns - There is growing anxiety in the US stock market regarding the return on AI investments, with any company unable to demonstrate a viable business model facing potential valuation corrections [2] - The tech sector has experienced significant stock declines, with major software companies like Microsoft and SAP seeing drops exceeding 8% due to concerns over cloud business order growth not meeting expectations [2] - The focus has shifted from technical metrics to the necessity of "business validation," with companies like Microsoft, Google, and Meta adopting different strategies to achieve monetization [2][3] Group 3: Challenges in AI Monetization - The projected capital expenditures of over $600 billion for the seven major US tech companies in 2026 raise concerns about whether these investments can yield returns without eroding free cash flow [3] - The integration of AI technology into actual business operations remains challenging, with enterprise clients demanding tangible business value rather than mere functionality [3] - High costs and reliability issues of generative AI in complex scenarios have led some companies to reassess their AI investment priorities, as initial expectations for efficiency gains have not been met [3] Group 4: Strategic Responses - Companies are adopting varied strategies in response to the current market pressures, with some focusing on deep integration of AI with core business functions while others continue to increase investments to achieve economies of scale [4] - The capital market has signaled that future competition will center around cash flow rather than narrative, indicating a shift in investor sentiment [4] - Companies that fail to demonstrate AI monetization capabilities may face significant challenges in the upcoming industry reshuffling, potentially leading to their exit from the market [4]

Core Viewpoint - Microsoft reported strong Q2 FY2026 earnings with revenue of $81.3 billion and EPS of $4.14, exceeding Wall Street expectations, yet the stock price fell over 6% post-announcement due to concerns over high capital expenditures and slower growth in cloud services [2][3]. Group 1: Financial Performance - Microsoft’s capital expenditures surged approximately 66% year-over-year to a record $37.5 billion, raising investor concerns about the return on investment (ROI) in relation to Azure's revenue growth of 39% [3][5]. - The company emphasized that the current growth limitation is not demand but supply, indicating a tight supply-demand relationship in the market [3][6]. Group 2: Strategic Insights - CFO Amy Hood stated that if all newly launched GPUs were allocated to Azure, growth rates would exceed 40%, highlighting a resource allocation challenge between external customer demands and internal AI product needs [5][6]. - CEO Satya Nadella stressed the importance of customer lifetime value (LTV) over short-term growth in individual business units, advocating for a balanced approach to resource allocation [7][8]. Group 3: AI and Product Development - Microsoft reported a 160% year-over-year increase in paid seats for Microsoft 365 Copilot, reaching 15 million users, and a tenfold increase in daily active users, countering concerns about declining AI tool usage [9][10]. - GitHub Copilot also saw a 75% increase in paid subscribers, indicating strong growth in AI productivity tools across both consumer and business sectors [11]. Group 4: Cost Management and Infrastructure - Microsoft introduced its self-developed Maya 200 accelerator, claiming a 30% reduction in total cost of ownership (TCO) compared to existing hardware, as part of its strategy to control AI infrastructure costs [12][13]. - The company highlighted the critical role of data storage and management in AI, with Microsoft Fabric achieving an annual revenue run rate exceeding $2 billion and a 60% year-over-year growth [15]. Group 5: Market Outlook - Microsoft expressed strong confidence in long-term AI demand, framing the current landscape as a "arms race" for computing power, where efficiency in deployment will determine market leaders [16][17].

Core Viewpoint - Jefferies reports that Microsoft's stock price has dropped 18% since the first fiscal quarter, despite announcing investment commitments of $250 billion in OpenAI and $30 billion in Anthropic, with a 23% compression in valuation multiples as investors shift towards semiconductor stocks [1] Group 1: Financial Performance - Microsoft is expected to significantly expand its capacity this year due to a large backlog of orders, indicating potential for upward movement [1] - The estimated price-to-earnings ratio for Microsoft's fiscal year 2027 is 23 times, which is attractive compared to Amazon and Google's ratios of 24 and 25 times respectively [1] Group 2: Competitive Positioning - Microsoft's operational visibility, remaining performance obligations (RPO) value, and AI monetization pathways are clearer compared to Amazon and Google [1] Group 3: Investment Recommendation - Jefferies maintains a "Buy" rating for Microsoft with a target price of $675 [1]

Core Viewpoint - Bank of America Securities has issued a report indicating that Kingdee International (00268) is expected to achieve an operating cash flow of 1.1 billion RMB in the fiscal year 2025, representing an 18% year-on-year growth. The firm maintains a "Buy" rating on Kingdee with a target price of 20.2 HKD [1] Group 1: Financial Performance - For the fiscal year 2025, Kingdee International anticipates revenue growth of 12% to 7 billion RMB, with a range of 6.95 billion to 7.05 billion RMB [1] - The adjusted net profit is expected to be 215 million RMB, with a range of 190 million to 240 million RMB, aligning with market consensus and Bank of America's expectations [1] - The adjusted net profit excludes items such as equity incentive expenses, goodwill impairment, and changes in the fair value of investments [1] Group 2: Growth Drivers - The revenue growth for fiscal year 2025 is primarily driven by strong expansion in large enterprise business [1] - Revenue growth is expected to accelerate in the second half of 2025, with a year-on-year increase of 12.5%, compared to an 11% increase in the first half [1] - The improvement in downstream demand is partly attributed to an increase in new orders from small and medium-sized enterprises in the fourth quarter of 2025 [1]

Core Viewpoint - OpenAI has announced the introduction of advertisements in ChatGPT for free and entry-level subscription users, signaling a need to diversify revenue streams amid operational pressures and upcoming financing rounds [1] Group 1: Company Actions - OpenAI will begin testing advertisements in ChatGPT within the next few weeks, specifically targeting free and entry-level subscription users, while Plus, Pro, Business, and Enterprise subscriptions will remain ad-free [1] - The decision to implement ads is seen as a direct response to the significant development and operational costs faced by OpenAI, indicating a need for sustainable growth to reassure investors [1] Group 2: Industry Insights - A domestic large model vendor's advertising business head noted that the introduction of ads in AI dialogue applications is a pioneering move, with expectations that other companies will follow suit after observing initial results [1] - Advertising professionals anticipate that while AI platforms are in a competitive landscape, there will be a balance between commercialization and user experience, predicting only a limited number of clearly marked ads in AI responses rather than a high volume akin to search engine ads [1] Group 3: User Reactions - Some users have expressed concerns regarding privacy, comparing the potential for ads in ChatGPT to intrusive advertising in personal spaces, such as a therapist's office [2] - Criticism has emerged regarding the potential for ChatGPT to replicate the advertising success of platforms like Instagram, which relies heavily on user data collection and privacy disregard [2]

Core Viewpoint - Meitu Inc. (1357.HK) stock price surged over 12%, reaching a peak of 8.89 HKD, with a total market capitalization exceeding 40 billion HKD, driven by the overall boom in AI concept stocks in the Hong Kong market [1] Group 1: Company Performance - Meitu's productivity tools business is rapidly expanding, and the company is expected to see a significant increase in its paid user rate driven by AI [1] - The stock price increase reflects a broader trend in the AI sector, with other AI-related companies also experiencing substantial gains [1] Group 2: Industry Outlook - The AI industry is currently in a phase of rapid expansion, with the recent IPOs of domestic AI large model companies expected to boost the overall market, including both model and application sectors [1] - Bank of America Securities expressed a cautiously optimistic outlook for software demand by 2026, highlighting the acceleration of AI monetization and the trend of software import substitution [1] - The report identified Meitu, along with Kingdee International and Kingsoft Office, as top picks in the Chinese software sector despite market concerns regarding competition among large models [1]

Core Viewpoint - The AI industry is increasingly dominated by major tech giants like Google, Microsoft, and ByteDance, making it difficult for smaller companies to compete effectively [6][9][10]. Group 1: Competitive Landscape - Major players such as Google and Microsoft are leveraging their vast resources to enhance their AI offerings, with Google's Gemini surpassing OpenAI's ChatGPT in various evaluations [10][12]. - Smaller AI companies like Manus and Kimi are struggling to maintain their market positions as they face overwhelming competition from these tech giants [10][11]. - The integration of AI into widely used applications, such as Google's embedding of Gemini into Android and Microsoft's integration of AI into Office, creates a significant barrier for smaller firms [10][12]. Group 2: Resource Dependency - The success of AI applications is heavily reliant on the backing of large corporations, as smaller companies lack the necessary resources and ecosystem integration to thrive [11][12]. - AI startups often find it challenging to monetize their technologies compared to larger firms that can bundle services and leverage existing customer bases [15][18]. - The financial struggles of AI startups are evident, with many facing increasing losses and limited paths to profitability [24][25]. Group 3: Monetization Strategies - Larger companies can implement diverse monetization strategies, such as bundling AI services with existing products, which enhances their revenue potential [15][18]. - Smaller companies often lack the ability to create similar attractive packages, limiting their monetization options to straightforward subscription models [21][20]. - The competitive pricing landscape for AI subscriptions is constrained, making it difficult for startups to charge premium prices [21][23]. Group 4: Acquisition Trends - The trend of larger companies acquiring smaller AI firms is becoming more prevalent, as seen with Meta's acquisitions of Scale and Manus, which can provide these startups with enhanced capabilities and market access [27][28]. - Acquired companies can benefit from the resources and infrastructure of their parent companies, allowing them to operate more effectively within the market [27][28]. - The desire for independence among some AI firms, like OpenAI, complicates the landscape, as they aim to establish themselves as major players rather than being absorbed by larger entities [28].