2026.02.24 本文字数：2349，阅读时长大约4分钟 作者 |第一财经 樊志菁 眼下，美股投资者正被迫直面并压力测试一个曾被认为不可能发生的情景：若科技巨头与AI头部云服 务商持续走弱，是否会同时拖垮美国股市与美国经济？ 今年以来，美股整体走势已经落后于全球主要市场，纳指和标普500指数已经完全回吐年内涨幅。投资 者对 "七大科技巨头"（Magnificent Seven）前景感到越发不安，微软年内跌幅超17%，亚马逊下跌超 10%。周一，Meta成为第三家跌入技术性熊市的科技巨头。 与此同时，行业板块波动剧烈，软件、财富管理、房地产服务等部门的股票，因担忧受AI颠覆性冲击 而遭遇重挫。 不只是简单的风格切换 新年伊始，一批此前落后的板块领涨美股，而 "七大科技巨头"则因市场担忧其巨额AI投入承压。 本月早些时候，因AI创新可能冲击其核心业务。软件股的抛售蔓延至整个科技板块，投资者抛售所有 可能被AI技术替代的公司股票。FactSet数据显示，标普500能源板块今年以来大涨超22%，在11个板块 中领跑；必需消费板块上涨12.7%，工业板块上涨13.1%。与之相对，科技权重极高的信息技术板块下 跌4. ...

事实上，这种从单纯的"比特币开采"向"全球算力基建"转型的逻辑已成为行业共识。ZFX山海证券表 示，另一矿业巨头 MARA Holdings 也在近期完成了对法国计算基础设施公司 Exaion 的控股权收购，深 度切入人工智能领域。目前，包括 HIVE、Hut 8 和 TeraWulf 在内的多家知名矿企都在积极改造原有的 比特币矿场能源设施，以承接利润率更高的 AI 高性能计算（HPC）业务。ZFX山海证券认为，随着挖 矿难度在 2026 年再度反弹及哈希价格承压，矿工群体正在经历一场从单一币本位收益向多元化算力收 入的结构性演变。 事实上，这种从单纯的"比特币开采"向"全球算力基建"转型的逻辑已成为行业共识。ZFX山海证券表 示，另一矿业巨头 MARA Holdings 也在近期完成了对法国计算基础设施公司 Exaion 的控股权收购，深 度切入人工智能领域。目前，包括 HIVE、Hut 8 和 TeraWulf 在内的多家知名矿企都在积极改造原有的 比特币矿场能源设施，以承接利润率更高的 AI 高性能计算（HPC）业务。ZFX山海证券认为，随着挖 矿难度在 2026 年再度反弹及哈希价格承压，矿工群体正 ...

| In USD $ millions | | Three months ended Dec 31 | | | Twelve months ended Dec 31 | | | --- | --- | --- | --- | --- | --- | --- | | | 2024 | 2025 | Change | 2024 | 2025 | Change | | Revenues | 35.2 | 227.7 | 547% | 91.5 | 529.8 | 479% | | Adjusted EBITDA / (loss) | (63.9) | 15.0 | n/m | (226.3) | (64.9) | -71% | | Net income / (loss) from continuing operations | (122.9) | (249.6) | 103% | (352.0) | 29.0 | n/m | | Adjusted net loss | (69.0) | (173.0) | 151% | (238.5) | (446.7) | 87% | Nebius如今已成为"新云服务"提供商中的 ...

智通财经APP获悉，AI云计算服务提供商Nebius(NBIS.US)公布了其2025年第四季度业绩。财报显示，Nebius四季度营收同比暴增547%至2.277亿美元，但不 及市场预期的2.475亿美元;调整后的息税折旧摊销前利润(BEBITDA)为1500万美元;调整后的净亏损为1.73亿美元，较上年同期的净亏损0.69亿美元扩大 151%。 Nebius如今已成为"新云服务"提供商中的佼佼者。其核心业务包括供应英伟达(NVDA.US)GPU及AI云服务，助力企业扩展AI基础设施。全球AI算力需求的持 续旺盛——即便是微软、亚马逊(AMZN.US)等头部云厂商也面临算力短缺——让Nebius及其主要竞争对手CoreWeave(CRWV.US)迎来需求爆发。Nebius因与 美国超大规模企业达成多项备受瞩目的人工智能基础设施供应协议而声名鹊起，其中包括与微软的170亿美元合作及与Meta(META.US)的30亿美元订单。 值得一提的是，Nebius周四宣布，计划在法国Béthune新建一座240兆瓦的数据中心，建成后将成为欧洲最大的数据中心之一。该公司首席传播官汤姆·布莱克 韦尔称，该项目将对原倍耐力轮胎 ...

Core Insights - The core message of the article emphasizes that AI has become the primary driver of cloud computing development, transforming cloud infrastructure from merely supporting IT to being essential for the deployment and scalability of AI applications [1] Group 1: Cloud Infrastructure Modernization - By 2027, over 85% of organizations in China will transition traditional cloud environments to new platforms that accommodate AI workloads, indicating that traditional IaaS/PaaS models are insufficient for scaling AI applications [2] - The modernization of cloud infrastructure is a prerequisite for enterprises to develop intelligent business operations [2] Group 2: AI-Driven Cloud Operations - By 2027, 80% of China's top 500 enterprises will deploy agent-based AI platforms to automate IT cloud operations, providing large-scale, continuous monitoring, analysis, and fault recovery capabilities with minimal human intervention [3][4] Group 3: AI Cloud Service Providers - By 2029, at least 30% of high-grade GPU resources will be provided by AI cloud service providers that offer cloud characteristics, flexible billing, APIs, and software services, distinguishing them from traditional GPU resource providers [5][6] Group 4: Edge AI and Private Cloud Platforms - By 2028, embodied intelligence will experience explosive growth, with cloud service providers deploying AI infrastructure and agents at the enterprise edge to support 60% of business scenarios [7] - To meet data privacy needs and reduce risks associated with public large language models, 60% of organizations in China will adopt private cloud platforms that offer greater control over data governance by 2028 [8][9] Group 5: AI Cost Governance and Heterogeneous Cloud Infrastructure - By 2028, companies that do not integrate AI investments into their cost governance will face a 30% increase in costs and lower overall returns on AI-related projects [10] - Over 80% of organizations in China will adopt heterogeneous cloud infrastructure by 2028 to balance mixed CPU, GPU, and storage technologies, optimizing the cost-effectiveness of AI workloads [11] Group 6: Cloud Risk Management and AI Workload Automation - By 2029, 50% of organizations implementing digital autonomy in China will migrate sensitive workloads to new cloud platforms to mitigate risks and enhance autonomy due to geopolitical uncertainties [12] - By 2029, 60% of organizations will use cloud-based AI integration tools to assess cost and performance metrics, deploying AI agents to automate workload collaboration and optimize workload substitution [13] Group 7: Intelligent SaaS Platforms - By 2029, 50% of Chinese enterprises will adopt SaaS platforms for real-time workflows, integrating predefined app functions with AI agents to create modular and interactive solutions [14] - The SaaS model is evolving towards a combination of "applications + agents," reflecting a shift in the cloud computing market from a growth phase to a capability reconstruction phase [15]

Group 1 - The core viewpoint of the report is that the domestic cloud computing market in China is expected to grow significantly by 20-30% to reach a scale of 450-500 billion RMB by 2026, with AI cloud contributing the largest incremental growth [1] - The AI cloud market is projected to double in size by 2026, reaching approximately 100 billion RMB [1] - Alibaba is anticipated to capture 80% of the incremental growth in the AI cloud market due to its comprehensive AI capabilities and existing customer base [1] Group 2 - Alibaba's cloud revenue for the calendar year 2026 is expected to exceed a year-on-year growth of 33%, reaching around 200 billion RMB [1] - The growth rate of external and overseas cloud revenue is expected to surpass that of internal and domestic cloud revenue [1] - The long-term profit margin for cloud business is projected to improve due to the increasing share of high-priced AI cloud services, particularly AIPaaS, and fragmented sales [1] Group 3 - For capital expenditure, Alibaba is estimated to invest between 160-180 billion RMB in 2026, with 70% of the budget allocated for the procurement of AI servers, including AI chips [1]

Group 1 - The core viewpoint of the article is that the Chinese cloud computing market is expected to grow significantly by 2026, with a focus on AI integration and Alibaba's market position [1] - Experts predict that the domestic cloud computing market size will reach between 450 billion to 500 billion RMB by 2026, with an annual growth rate of 20-30% [1] - The AI cloud segment is expected to double in size by 2026, reaching approximately 100 billion RMB, contributing the most to the overall market growth [1] Group 2 - Alibaba is anticipated to capture 80% of the AI cloud market increment due to its comprehensive AI capabilities and existing customer base [1] - The forecasted cloud revenue for Alibaba in 2026 is expected to exceed 200 billion RMB, reflecting a year-on-year growth of over 33% [1] - External and overseas cloud revenue growth is projected to outpace that of internal and domestic cloud revenue [1] Group 3 - Long-term profit margins for cloud services are expected to improve due to the rising proportion of high-priced AI cloud services, particularly AIPaaS [1] - Alibaba's capital expenditure for 2026 is estimated to be between 160 billion to 180 billion RMB, with 70% of the budget allocated for the procurement of AI servers, including AI chips [1]

Group 1 - The core view of the report emphasizes the importance of "rebalancing" in the investment strategy for 2026, highlighting the dual focus on AI technology, overseas equipment, and global pricing resources as the main consensus among institutional investors [1][2] - The report indicates that the share of technology and overseas sectors in A-share profits (excluding finance) is approaching 40% by Q4 2025, suggesting a significant shift in the profit structure towards high-end technology and manufacturing, which is expected to reshape the A-share profit landscape and drive a new upward cycle in 2026-2027 [1][2] - The report outlines a transition from "new triumphing over old" in 2025 to "new and old dancing together" in 2026, where "new" refers to AI technology moving downstream and "old" refers to traditional industries stabilizing and growing through overseas business [2][3] Group 2 - The report highlights that global pricing resources, particularly gold, are experiencing a shift in asset allocation due to narratives of de-globalization and financialization, with a notable increase in trading sentiment driven by interest rate cuts and a weak dollar [2][3] - It is noted that the pricing of resource commodities is becoming increasingly differentiated, with financial attributes of resource pricing outperforming those based on commodity attributes [2][3] - The report stresses the need to be cautious of the assumption that the dollar will remain weak throughout 2026, as there may be a return to commodity attributes and a decline in financial attributes, making supply-demand fundamentals more critical for resource price increases [3] Group 3 - Observations from Q4 2025 indicate a significant increase in institutional holdings in sectors such as non-ferrous metals, communications, basic chemicals, non-bank financials, and machinery, while reductions were noted in pharmaceuticals, computing, electronics, media, and power equipment [9][10] - The report identifies a divergence in institutional investment in the AI industry chain, with a decrease in holdings in sectors with weaker earnings visibility, while sectors with strong earnings visibility, such as optical modules, saw increases [10][11] - The report also notes that institutional investors are increasingly favoring resource commodities that benefit from price increases, particularly in the non-ferrous and chemical sectors, indicating a strategic shift towards these areas [10][11]

Summary of Key Points from Conference Call Records Industry Overview - The focus is on the AI and cloud services industry, particularly the transition to cloud-based computing and cost reduction strategies in 2026 [1][3][4]. Core Insights and Arguments - AI cloud services are expected to enter a growth phase in 2025, with model vendors converting capital expenditures into Kubernetes (K8S) to achieve cloud computing and cost reduction, similar to the mobile internet era [1][3]. - Data center computing power is projected to see significant growth with the large-scale adoption of 1.6T and 224G single-channel technologies, which will profoundly impact China's optical communication industry [1][4]. - LightCounting forecasts that by 2030, the shipment of 1.6T will accelerate, surpassing 800G and 400G around 2028, with Chinese upstream optical chip companies likely achieving global replacement [1][5]. - The transition from copper connections to performance realization narratives is highlighted, with expectations that copper connections will remain dominant for the next two years [1][6]. - The commercial space sector is anticipated to have significant developments in 2026, including reusable liquid rockets and the gradual implementation of commercial satellite constellations [1][12]. Additional Important Insights - The infrastructure for computing power is experiencing price increases, with cloud computing seeing its first price hike in over 20 years, expected to continue until 2026 [2][22][23]. - The CDN market may follow suit with price increases, potentially impacting smaller clients who may seek cheaper alternatives [24][25]. - AI inference in edge computing is projected to have vast applications, with edge nodes becoming ideal for reducing latency and improving cost-effectiveness [26]. - Liquid cooling systems are becoming crucial in data center construction due to increased power demands from GPUs, with significant market opportunities arising from this trend [27]. - Power technology is also evolving, with new market spaces emerging from innovations in power supply systems within data centers [28]. Investment Opportunities - Key investment opportunities in the communication industry include companies like Xuzhuang and Yizhongtian, which are gaining market share in the optical module sector [5][6]. - The potential for growth in AI modules from IoT modules is emphasized, with companies that can provide integrated intelligent solutions expected to see rapid growth [13]. - The commercial rocket sector is highlighted as a critical area for investment, with expectations of successful validation of new reusable rocket models in 2026 [18][19]. - Investors are advised to focus on companies with strong backing and significant satellite planning capabilities in the space computing sector [19][21].

Core Insights - The report highlights the overall positive trend in the A-share market, with various sectors such as finance, semiconductors, and aerospace showing strong performance, while some sectors like coal and pharmaceuticals lag behind [5][8][12] - The report emphasizes the importance of macroeconomic data and policy changes in guiding investment strategies, suggesting a balanced approach to portfolio allocation [7][11][33] Domestic Market Performance - The Shanghai Composite Index closed at 4,139.90 with a slight increase of 0.18%, while the Shenzhen Component Index closed at 14,329.91, up by 0.09% [3] - The average price-to-earnings ratio for the Shanghai Composite and ChiNext Index stands at 16.93 and 53.33 respectively, indicating a favorable long-term investment environment [7][12] International Market Performance - Major international indices such as the Dow Jones and S&P 500 experienced declines of 0.67% and 0.45% respectively, while the Nikkei 225 saw a modest increase of 0.62% [4] Industry Analysis - The report discusses the media sector, noting a decrease in fund holdings in Q4 2025, with a significant focus on the gaming sub-sector, which remains highly favored among institutional investors [13][16] - The automotive industry is projected to achieve record production and sales in 2025, driven by policies encouraging vehicle upgrades and a strong demand for electric vehicles [31][33] Investment Recommendations - The report suggests focusing on sectors with strong fundamentals and high dividend yields, such as traditional engineering machinery and shipbuilding, while also highlighting opportunities in emerging technologies like humanoid robots and AI applications [17][18][22] - Specific companies to watch include major players in the gaming sector and those involved in AI applications, as they are expected to benefit from favorable market conditions and technological advancements [16][27]