证券研究报告|行业研究周报 [Table_Date] 2026 年 02 月 07 日 马斯克构建超级科技矩阵：太空算力时代即 将到来 [Table_Title2] 计算机行业周报 ► 二、太空算力成本迎降维打击，2-3 年成全球算力 最低区域 ► 一、SpaceX 全资并购 xAI，万亿级太空+AI 巨头诞 生 2026 年 2 月，马斯克旗下商业航天龙头 SpaceX 以全 股票方式，完成对其创立的 AI 公司 xAI 的全资并购， 这是全球科技产业史上规模最大的跨领域整合案例之 一，标志着航天与 AI 技术进入深度融合阶段。交易 后，双方合并新实体估值达 1.25 万亿美元，其中 SpaceX 估值 1 万亿美元、xAI 估值 2500 亿美元，跻身 全球最具价值科技企业行列。此次并购实现了核心资 源全面协同，SpaceX 的火箭发射、星链卫星互联网为 太空算力提供运力与通信支撑，xAI 的 Grok 大模型、 超大规模算力集群为 SpaceX 提供"智能大脑"。这是 一场双向救赎，xAI 摆脱资金链断裂风险并获得太空 算力场景，SpaceX 补齐星际探索的 AI 短板，同时推 动马斯克身家增至 852 ...

Investment Rating - The investment rating for the company is "Outperform" [6] Core Insights - The company's overall performance shows steady growth in e-commerce and accelerated growth in cloud revenue, with Q4 2025 revenue at $213.4 billion, up 14% year-over-year, and operating profit at $25 billion, up 18% year-over-year [1][9] - For Q1 2026, the company expects revenue between $173.5 billion and $178.5 billion, representing a year-over-year growth of 11%-15% [1][9] - The company plans to increase capital expenditures to $200 billion in 2026, a 56% year-over-year increase [3][4] Summary by Sections Overall Performance - Q4 2025 revenue reached $213.4 billion, with a net profit of $21.2 billion, reflecting a 6% year-over-year increase [1][9] - The company included $2.44 billion in litigation, impairment, and severance costs in its operating profit [1][9] Retail and Other Businesses - North American revenue was $127.1 billion, up 10% year-over-year, while international revenue was $50.7 billion, up 17% year-over-year [2][21] - Advertising revenue grew by 23% year-over-year, reaching $21.3 billion in Q4 2025 [2][21] - The AI shopping assistant Rufus has 300 million active users, contributing nearly $12 billion in annualized incremental sales [2][29] Cloud Business - AWS revenue for Q4 2025 was $35.6 billion, a 24% year-over-year increase, with an operating profit margin of 35% [3][20] - The company plans significant capital expenditures for AWS, with $38.5 billion spent in Q4 2025, a 48% increase year-over-year [3][20] - The company has delivered over 1.4 million Trainium2 chips, with a 30%-40% improvement in cost-effectiveness compared to similar GPUs [3][27] Financial Forecasts - Revenue projections for 2026 and 2027 are adjusted to $808.5 billion and $904.8 billion, respectively, with net profit forecasts of $86.7 billion and $99.7 billion [4][5] - The company expects an EPS of $8.19 for 2026 and $9.41 for 2027 [5][33]

Investment Rating - The industry is rated positively, with expectations for significant growth in capital expenditures from major companies like Amazon and Google, indicating a bullish outlook for the sector [1][26]. Core Insights - Amazon raised its 2026 capital expenditure forecast to $200 billion, a more than 50% increase from $131 billion in 2025, driven by strong demand signals in AI hardware and cloud services [1]. - Google's parent company, Alphabet, expects its 2026 capital expenditure to be between $175 billion and $185 billion, nearly double its 2025 spending, reflecting robust growth in its cloud business [1]. - The demand for AI infrastructure is expected to lead to a significant increase in ASIC (Application-Specific Integrated Circuit) production from companies like Google, Amazon, Meta, OpenAI, and Microsoft, with a projected explosive growth in 2026-2027 [1][4][26]. Summary by Sections 1. Capital Expenditure and Demand - Amazon's AWS cloud division saw a 24% year-over-year growth, with a backlog of $244 billion, indicating strong enterprise investment in AI and core cloud services [1]. - Google's cloud revenue reached $17.7 billion in Q4, a 48% year-over-year increase, driven by strong demand for enterprise-level AI products [1]. 2. AI Hardware and ASIC Demand - The report highlights the strong demand for AI-related hardware, with Amazon's self-developed Trainium chip showing a 30%-40% cost advantage over similar GPUs, leading to significant revenue potential [1]. - Companies are expected to ramp up production of new generation ASIC chips, with Amazon and Google entering a phase of increased demand and production [1][4]. 3. PCB and Semiconductor Industry Outlook - The PCB industry is experiencing high demand, with companies expanding production capacity due to strong orders driven by AI applications [4][26]. - The semiconductor equipment sector is also expected to benefit from increased capital expenditures, with a focus on domestic production capabilities in light of global supply chain challenges [23][25]. 4. Specific Company Insights - Companies like North Huachuang and Zhongwei are positioned to benefit from the growing demand for semiconductor equipment and materials, with a focus on domestic production and technological advancements [28][25]. - The report emphasizes the growth potential for companies involved in AI-related PCB manufacturing, with expectations for sustained high growth in performance and revenue [4][26].

来源：中央纪委国家监委网站 关注" 新华网财经 "视频号 更多财经资讯等你来看 往期推荐 腾讯"元宝分10亿"活动2月1日0点正式开启，用户最高可领万元现金红 包，官方攻略发布 据中央纪委国家监委驻交通运输部纪检监察组、北京市纪委监委消息：原首都机场集团公司党委委员、常务副总经理王星涉嫌严重违纪违法，目前正接受 中央纪委国家监委驻交通运输部纪检监察组和北京市东城区监委纪律审查和监察调查。 （中央纪委国家监委驻交通运输部纪检监察组 北京市纪委监委） 亚马逊官宣裁撤1.6万个职位 ...

Core Viewpoint - The narrative around AI is shifting from "software eating the world" to "hardware being constrained by the world," highlighting the political and physical limitations facing the expansion of data centers in the U.S. [2] Group 1: Political Environment - New York has proposed a bill to pause the construction and operation of new data centers for at least three years, marking it as the sixth state to consider such a measure [3] - There is a rare bipartisan agreement between Senator Bernie Sanders and Governor Ron DeSantis on the need to slow down the rapid increase of data centers due to public concerns over their impact [3][4] - DeSantis has shifted his stance from supporting tax incentives for data centers to advocating for legislation that requires these centers to fully pay for their water and electricity costs [4] Group 2: Financial Implications - The anticipated capital expenditure of approximately $600 billion by 2026 is now under scrutiny due to political and physical constraints [6] - Major tech companies, including Microsoft, Meta, Amazon, and Google, plan to spend $670 billion on AI infrastructure this year, which is a significant investment [8] - Amazon alone is expected to increase its capital expenditure by nearly 60% to $200 billion this year [9] Group 3: Energy Demand and Infrastructure - Data centers' energy demand is projected to double by 2035, increasing from 34.7 GW in 2024 to 106 GW, equivalent to the electricity consumption of 80 million households [11] - In Texas, the ERCOT has proposed a review of approximately 8.2 GW of power consumption projects, which could significantly impact previously approved projects [12] - The uncertainty surrounding energy supply is jeopardizing the expansion plans of tech giants, as the inability to connect to the grid could prevent the realization of the $670 billion budget [12] Group 4: Market Reactions - The financial market has reacted sharply to the risk of capital expenditures not being realized, leading to significant sell-offs in tech stocks [13] - Independent power producers (IPPs) and nuclear power stocks have also seen declines, as the market realizes that without grid expansion, new power demands cannot be met [14] - There is a growing trend of funds moving from high-beta tech stocks to defensive sectors like chemicals and regional banks, indicating a shift in market sentiment [14][15]

以下文章来源于硬AI ，作者专注科技产研的 硬AI . AI时代，快人一步~ 随着AI基础设施建设的军备竞赛进入"深水区"，一个令投资者不安的转折点已然浮现：为了支撑AI算力需求，亚马逊、谷歌和Meta正面临自由现金流被耗尽甚 至透支的风险。 根据摩根大通2026年2月5日发布的研究报告，美国四大云巨头——亚马逊、谷歌、Meta和微软，2026年总资本支出预计将达到 6450亿美元 ，同比激增 56%，新增支出将达到惊人的2300亿美元。 如果说谷歌还只是在"疯狂花钱"，那么 亚马逊 则堪称"透支未来"。 2026年，亚马逊的资本支出指引约为 2000亿美元 （同比增长52%）。但问题的核心在于，亚马逊赚回来的现金已经盖不住支出了——据标普全球市场分析师 预测，亚马逊2026年的运营现金流（OCF）约为 1780亿美元 。 对于投资者而言，2026年，或将是紧盯科技巨头资产负债表的一年。 谷歌的97%增速与亚马逊的"现金赤字" 在这场基建狂潮中， 谷歌 的投入非常激进。 2026年，谷歌的资本支出指引已上调至 1750亿至1850亿美元 ，同比增速高达 97% ，其资金正疯狂涌向服务器和技术基础设施。 这意味 ...

Investment Rating - The industry is rated positively, with expectations of significant growth in capital expenditures from major companies like Amazon and Google, indicating a bullish outlook for the sector [1][26]. Core Insights - Amazon raised its 2026 capital expenditure forecast to $200 billion, a more than 50% increase from $131 billion in 2025, driven by strong demand signals in AI hardware and cloud services [1]. - Google's parent company, Alphabet, expects its 2026 capital expenditure to be between $175 billion and $185 billion, nearly double its 2025 spending, reflecting robust growth in its cloud business [1]. - The demand for AI infrastructure is expected to lead to a significant increase in ASIC (Application-Specific Integrated Circuit) numbers from major tech companies, with a projected explosive growth in 2026-2027 [1][4]. - The semiconductor and PCB (Printed Circuit Board) sectors are anticipated to benefit from the strong demand for AI, with many companies in these areas experiencing high order volumes and expansion plans [4][26]. Summary by Sections Section 1: AI and Cloud Infrastructure - Amazon's AWS cloud division saw a 24% year-over-year growth, with a backlog of $244 billion, indicating strong enterprise investment in AI and cloud services [1]. - Google's cloud revenue reached $17.7 billion in Q4, a 48% increase year-over-year, driven by strong demand for enterprise-level AI products [1]. Section 2: Semiconductor and PCB Industry - The PCB industry is maintaining high demand due to the growth in automotive and industrial applications, with expectations of price increases for copper-clad laminates [6]. - AI-driven demand is expected to boost PCB prices and volumes, with companies actively expanding production to meet this demand [4][26]. Section 3: Consumer Electronics - The consumer electronics sector is seeing a continuous expansion of AI applications, particularly in the Apple supply chain, with innovations in foldable devices and AI glasses [5]. - The demand for AI-enabled devices is expected to drive significant growth in the market, with various manufacturers exploring new product categories [5]. Section 4: Semiconductor Equipment and Materials - The semiconductor equipment sector is experiencing a robust upward trend, with significant capital expenditures expected to support the growth of advanced manufacturing processes [23][25]. - The materials sector is also poised for improvement, with a focus on domestic production capabilities in response to international supply chain challenges [25][34]. Section 5: Company-Specific Insights - Companies like North Huachuang and Zhongwei are positioned to benefit from the increasing demand for semiconductor equipment and materials, with strong growth prospects in their respective markets [28][30]. - Three Ring Group is focusing on high-capacity MLCC (Multi-Layer Ceramic Capacitor) products, which are expected to see increased demand due to AI applications [33].

在人工智能浪潮的推动下，美国四大科技巨头正开启一场本世纪以来规模空前的资本竞赛。最新预测显示，Alphabet、亚马逊、Meta Platforms和微软在二零 二六年的合计资本支出有望达到约六千五百亿美元，较前一年大幅增长约百分之六十。这笔令人瞠目的资金，将如洪流般涌入新建数据中心、AI芯片、网 络设备及备用发电机等一长串硬件设施之中。 紧随其后，Alphabet抛出了高达一千八百五十亿美元的年度资本支出预测，让投资者感到不安。亚马逊的预估则更为惊人，其二零二六年资本支出计划可能 触及两千亿美元大关，消息公布后其市值应声下挫。这一系列数字不仅反映了企业对未来的巨额押注，也凸显出行业ai基础设施竞赛的白热化。 这场由少数 科技巨头主导的投资狂潮，正对更广泛的经济层面产生涟漪效应。规模庞大、耗电极高的数据中心在全美各地加速建设，已在局部地区导致能源供应紧张， 并引发了对电价上升的普遍担忧。同时，由于这些科技企业在整体经济活动中的占比日益增大，其庞大的资本开支可能会扭曲宏观经济数据的解读，成为影 响经济决策的新变量。 据凯德北京投资基金管理有限公司分析师分析，全球数据中心建设浪潮正在这股力量的推动下再次提速。从芯 ...

Global Asset Performance - Central banks' monetary policies are diverging, with Australia raising interest rates for the first time in two years, while the European Central Bank and the Bank of England kept rates unchanged [4][5] - Amazon is expected to increase capital expenditures by over 50% this year to build AI infrastructure, impacting its stock price negatively [4] - The U.S. Federal Reserve's stance is seen as favorable for economic stability, with a projected growth of approximately 2.2% in 2026 [4][5] Commodity Market - COMEX gold rose by 5.13%, while COMEX silver fell by 1.28% due to increased margin requirements and selling pressure from leveraged positions [9][10] - The core logic for gold's bull market has shifted from short-term interest rate speculation to hedging against long-term dollar credit risks and global monetary system restructuring [10] - Industrial metals are expected to benefit from global green transitions, indicating a positive long-term demand structure [10] Bond Market - U.S. Treasury yields showed a significant bull steepening, with the 2-year yield declining sharply due to weak labor market data, while long-term yields fell less due to inflation concerns [20][24] - The Chinese bond market is experiencing strong fluctuations, with the yield curve showing complex changes, particularly in the 10-year and 30-year bonds [24][25] - Long-term expectations suggest a potential decline in Chinese bond yields, with the 10-year yield possibly falling to the 1.6%-1.8% range by the end of 2026 [25] Currency Market - The U.S. dollar index increased by 0.51%, driven by a significant sell-off in global tech stocks and traditional safe-haven trading [27] - The euro against the dollar fell by 0.30%, influenced by the ECB's decision to maintain interest rates and easing inflation pressures [34] - The British pound declined by 0.95%, affected by internal policy disagreements and expectations of potential rate cuts [39] Equity Market - The market is shifting from tech growth stocks to value stocks, with traditional sectors like finance and energy leading gains, while tech-heavy indices like NASDAQ faced declines [43][48] - Notable performances include the Indian SENSEX30 rising by 3.54% and the Dow Jones Industrial Average increasing by 2.50% [43] - The outlook suggests that if higher interest rates become the norm, growth stocks may face ongoing valuation pressure, while stable value stocks could outperform [48]

2026.02. 08 本文字数：2470，阅读时长大约4分钟 在理论定义中，AGI被视为AI的终极状态，意味着系统能在会计、法律等多个白领专业领域达到并取 代人类的智力水平。 现代AI"教父"之一本吉奥（Yoshua Bengio）警示称，AGI的研发进程可能会陷入停滞，而这种停滞 对投资者而言无异于灾难。他称，行业很可能在不久的将来"碰壁"，面对目前无法预见且难以快速解 决的技术瓶颈，这可能诱发一场真正的金融崩盘。毕竟，当前涌向AI领域的数万亿资金，大多建立 在技术能以目前速度规律进步的乐观预期之上。 罗伯特·李凭借其在帝国理工学院深造电子工程的学术背景，也提供了基于技术视角的结论。他表 示："从本质上看，AI运行的底层逻辑近年来并未发生颠覆性变革，真正的改变在于算力与存储能力 的爆发式增长，以及模型从处理单一数字数据向多模态（涵盖文本、视频和音频）的进化。尽管如 此，这些模型的核心仍是模式识别，实质上是极其复杂的统计模型。" 罗伯特·李形象地解释道，当用户利用AI模型搜索量子计算时，模型只是通过训练库或互联网搜集海 量资料，并利用模式识别提取共性并予以总结。这种功能虽然高效，却并非真正的智能。 作者 | ...