美国总统特朗普在公开场合升级了对美联储主席鲍威尔的批评，直言"我很想炒掉他"，并称其"极度无能"。 周三特朗普在美国华盛顿举行的美国-沙特投资论坛上，敦促美国财政部长贝森特加快寻找鲍威尔继任者的进程 。特朗普对坐在观众席的贝森特说： 你得努力点，贝森特。贝森特唯一搞砸的就是美联储这件事。 鲍威尔的美联储主席任期将于明年5月到期，美联储理事任期则到2028年。特朗普还半开玩笑地对贝森特施压： 利率太高了，贝森特，如果你不快点搞定，我就要炒掉你。 正在牵头物色美联储新主席人选的贝森特近日称， 特朗普将在感恩节11月27日后与三位最终候选人会面，新的人选可能在12月25日圣诞节前宣布。 继任者搜寻进入冲刺阶段 周二据媒体报道，特朗普称美国政府对美联储最高职位已有一些"很好的人选"。 特朗普此前曾多次表示，贝森特将是他对美联储主席职位的首选，但贝森特告诉他更愿意留任美国财政部和国税局负责人。 目前贝森特已确定五位最终候选人名单： 美国白宫经济顾问哈塞特、前美联储理事沃什、现任美联储理事沃勒、负责监管的美联储副主席鲍曼，以及贝莱德高 管Rick Rieder。 特朗普周三的讲话首次暗示了白宫内部对鲍威尔去留问题的分歧。 ...

截至三季度末，H&H International Investment共持有11家公司股票，总市值约1044亿元人民币。苹果依然占据第一大重仓股地位，持仓市值88.69亿美元，占 比60.42%。伯克希尔·哈撒韦以26.1亿美元的持仓市值位列第二，占比17.78%。 | Stock | History | | Shares Held | Market Value | % of | Previous % of | Rank | Change in | 8 | 1/2 | Qtr 1st | Est. Avg | Qtr End | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | | | Sector | or Principal | → | Portfolio | Portfolio | | Shares | Change | Ownership | Owned | Price | Price | Price 1D | | | | | Amt | | | | | | | ...

当地时间周三， 美国劳工统计局（Bureau of Labor Statistics，BLS）表示，将不会发布10月非农就 业报告，而是把相关的就业数据纳入11月报告。 11月非农就业报告将于12月16日发布，比原计划晚了一个多星期，并且是在美联储今年最后一次会议 之后公布。 9月非农就业报告将于本周四发布。 美国非农就业报告由两项调查组成，一项面向家庭，另一项面向企业机构，后者用来统计非农就业人 数。 由于美国创纪录的政府关门，BLS无法收集10月的住户调查数据——这些数据用于计算包括失业率在内 的关键指标 。BLS表示， 这些数据无法事后补采 。BLS还表示， 将延长11月家庭调查和机构调查的收 集周期。 经济学家此前就指出，由于数据收集方式极为依赖人工，家庭调查数据存在被跳过的风险，但他们仍预 期10月的非农就业数据会如期发布。白宫国家经济委员会主任哈赛特上周曾表示，10月就业报告将不会 包含失业率。 虽然许多企业保存记录并以电子方式报告薪酬数据，但要事后通过电话联系居民、并让受访者回忆10月 某一周的就业状况，难度要大得多。家庭调查由BLS与人口普查局联合进行，每月调查约6万户家庭。 BLS还将跳过发 ...

Core Viewpoint - Nvidia's recent financial results indicate strong demand for AI infrastructure, with revenue growth exceeding 60%, alleviating concerns about an AI bubble [1][3][15]. Financial Performance - Revenue for Q3 reached $57.01 billion, a year-on-year increase of approximately 62%, surpassing analyst expectations of $55.19 billion [4]. - Non-GAAP adjusted EPS for Q3 was $1.30, up 60% year-on-year, exceeding the expected $1.26 [5]. - Adjusted gross margin for Q3 was 73.6%, slightly below the expected 74.0% [6]. - Adjusted operating expenses for Q3 were $4.215 billion, a 38% increase year-on-year [7]. Segment Performance - Data center revenue for Q3 was $51.2 billion, a 66% year-on-year increase, exceeding analyst expectations [8]. - Gaming and AI PC revenue for Q3 was $4.3 billion, a 30% year-on-year increase [9]. - Professional visualization revenue for Q3 was $760 million, a 56% year-on-year increase [10]. - Automotive and robotics revenue for Q3 was $592 million, a 32% year-on-year increase [10]. Guidance and Future Outlook - Q4 revenue is projected to be $65 billion, with a range of $63.7 billion to $66.3 billion, exceeding analyst expectations [12]. - Q4 gross margin is expected to be 75.0%, indicating a potential year-on-year increase [13]. - Q4 adjusted operating expenses are projected to be $5 billion, higher than analyst expectations [14]. Market Position and Strategy - Nvidia holds $60.6 billion in cash and equivalents, indicating strong financial support for AI applications [2]. - The company has secured $500 billion in chip orders for 2025 and 2026, highlighting significant future revenue potential [23][24]. - Major clients like Microsoft, Amazon, Alphabet, and Meta account for over 40% of Nvidia's sales, with expected AI spending growth of 34% to $440 billion in the next 12 months [22]. Competitive Landscape - Nvidia's CFO noted that the revenue guidance does not include data center computing revenue from China, indicating potential market challenges [19]. - The company is exploring ways to offer more competitive data center products for the Chinese market amid geopolitical tensions [26].

Core Viewpoint - The article discusses the debate surrounding the economic lifespan of GPUs, which is crucial for understanding the profitability of major tech companies and the validity of current AI valuations. Bernstein's report suggests a depreciation period of 6 years for GPUs, arguing that this is economically reasonable, while critics like Michael Burry claim the actual lifespan is only 2-3 years, warning of potential accounting manipulation to inflate profits [1][11]. Group 1: Economic Viability of GPU Depreciation - Bernstein analysts argue that a 6-year depreciation period for GPUs is justified, as the cash costs of operating older GPUs are significantly lower than their rental prices [2][4]. - The report highlights that even 5-year-old NVIDIA A100 chips can still yield "comfortable profits," indicating that the depreciation policies of major cloud service providers are fair and not merely for financial embellishment [2][4]. - The analysis shows that the contribution profit margin for A100 chips can reach up to 70%, with operational costs being substantially lower than rental income, providing strong economic incentives for extending GPU usage [4][5]. Group 2: Market Demand and Old GPUs - The current market environment supports the value of older GPUs, as there is overwhelming demand for computing power, with AI labs willing to pay for any available capacity, even for outdated models [6][7]. - Industry analysts note that the A100's computing capacity remains nearly fully booked, suggesting that as long as demand stays strong, older hardware will continue to hold value [8]. Group 3: Depreciation Policies of Tech Giants - Google has a depreciation period of six years for its servers and network equipment, while Microsoft ranges from two to six years, and Meta plans to extend some assets to 5.5 years starting January 2025 [9][10]. - Notably, Amazon has reduced the expected lifespan of some servers and network equipment from six years to five years, reflecting differing views within the industry on hardware iteration speed [10]. Group 4: Criticism and Concerns - Michael Burry warns that tech giants are artificially inflating profits by extending the effective lifespan of assets, predicting that this accounting practice could lead to a profit inflation of $176 billion from 2026 to 2028 [11][12]. - Burry specifically points out that companies like Oracle and Meta could see their profits overstated by 26.9% and 20.8%, respectively, due to these practices [12]. - Previous warnings from Bank of America and Morgan Stanley indicate that the market may be underestimating the true scale of AI investments and the potential surge in future depreciation costs, which could reveal a lower actual profitability for tech giants than expected [14][15].

会议纪要显示，上月末的最近一次货币政策会议上，美联储决策者对12月是否降息存在严重分歧，支持降息的一方并未在人数上占绝对优势，对于缩减资产负 债表（缩表）的量化紧缩（QT）行动，则几乎完全一致认为应该停止。在金融稳定的风险方面，一些人担心股市无序下跌。 美东时间11月19日周三公布的美联储会议纪要写道： "在讨论货币政策的近期走向时， 与会者对 （货币政策）委员会（FOMC） 12月会议最有可能采取的政策决定表达了截然不同的看法 。大多数（Most）与会者 认为，随着委员会逐步转向更为中性的政策立场，可能适合"进一步降息， "然而，其中一些（several）人暗示，他们未必认为12月会议适合再降息25个基点。 一些（Several） 与会者评估 认为 ，如果在接下来的两次会议之间，经济 发展符合他们的预期， 可能较为适合12月"进一步降息 。" 许多（Many） 与会者表示，根据他们的经济展望，在今年剩余时间内， 可能适合"维持利率不变 。 所有与会者一致认为，货币政策并非一成不变，而是会受到各种最新数据、不断变化的经济前景以及风险平衡的影响。 媒体指出，在美联储会议纪要中常用的所谓计数术语中，"许多"（Ma ...

Core Viewpoint - The emergence of the "Qijing" brand, a collaboration between GAC Group and Huawei, signifies a strategic shift in the Chinese smart automotive market, emphasizing youthfulness and technology as key differentiators in a competitive landscape [2][4][5]. Group 1: Brand Development - The "Qijing" brand was born from a strategic meeting between GAC Group's chairman and Huawei's founder, where the need for a younger and more tech-oriented brand was emphasized [2]. - The brand aims to leverage Huawei's technological expertise and GAC's manufacturing capabilities to create a new category of high-end smart electric vehicles [4][30]. - The first model of "Qijing" is expected to launch in mid-2026, with significant anticipation surrounding its design and technological integration [20][21]. Group 2: Market Insights - The target demographic for "Qijing" includes young consumers, with 41.5% of internet users aged 20 to 29 already utilizing generative AI products, indicating a strong inclination towards technology [9]. - By 2025, the penetration rate of new energy vehicles in China is projected to exceed 52%, highlighting a significant shift towards smart and electric vehicles among younger consumers [10]. - The increasing popularity of smart vehicles among younger generations is evident, with 48.2% of new users of smart car apps being born in the 1990s and 2000s [10]. Group 3: Strategic Collaboration - The partnership between GAC and Huawei is characterized by a deep integration of their operations, moving beyond traditional client-supplier relationships to a collaborative model that enhances product development and marketing [12][13]. - Huawei's extensive experience in smart technology, combined with GAC's 28 years of automotive manufacturing expertise, positions "Qijing" to potentially disrupt the high-end electric vehicle market [17][30]. - The collaboration includes the implementation of Huawei's integrated product development and marketing services, ensuring that user needs are central to the product development process [13][14]. Group 4: Leadership and Management - Liu Jiaming has been appointed as the CEO of "Qijing," bringing over 25 years of automotive industry experience to lead the brand's strategic direction and market expansion [22][25]. - The leadership team is focused on creating a product that embodies technology, fashion, and sports elements, aiming for a strong market impact upon launch [21][29]. Group 5: Future Outlook - The upcoming launch of "Qijing" is anticipated to set a new standard for traditional manufacturing companies transitioning to smart technology, with expectations for it to become a model for future collaborations in the automotive industry [29][31]. - The competitive landscape for smart vehicles is expected to intensify by the time "Qijing" enters the market, but the combination of Huawei's technology and GAC's manufacturing experience may provide a unique advantage [30][31].

Core Insights - The market is shifting focus from "what can be discussed" to "what can be accomplished," indicating a clear trend towards AI as a productivity tool rather than just a conversational agent [1][5] - Major industry events in November, including the launch of Alibaba's "Qianwen" app and Ant Group's "Lingguang" AI assistant, signify a pivotal moment in the AI application landscape [2][4] - The competition is evolving from measuring daily active users to understanding who can meet core user needs effectively [5][8] Group 1: AI Application Landscape - The new battleground for AI applications is emerging as the "productivity entry point," with Ant Group's "Lingguang" exemplifying this shift by focusing on task resolution rather than general interaction [6][18] - Users are transitioning from seeking simple Q&A interactions to wanting one-click solutions that accomplish tasks [7][19] - The AI landscape by 2025 is expected to be more defined, with leading players adopting distinct strategic paths [8][61] Group 2: Competitive Layers - The AI entry point has diversified into three layers: 1. The first layer focuses on traffic "moats," represented by Tencent and ByteDance, which aim to enhance consumer engagement through AI [10][11] 2. The second layer serves as a B2B "arsenal," with companies like Baidu and Alibaba providing AI solutions to enhance business operations [12][13][14] 3. The third layer is the C-end "personal assistant," which addresses proactive user tasks rather than passive entertainment [15][16][17] Group 3: Ant Group's "Lingguang" - "Lingguang" is positioned as the first all-code-generating multi-modal AI assistant, capable of creating interactive tools in response to user requests [21][24] - The shift from providing information to generating tools represents a fundamental change in AI's role, moving towards "scene productivity" [25][36] - The value of the "productivity entry point" lies in creating closed loops for high-intent scenarios, enhancing user engagement and satisfaction [26][27] Group 4: User Engagement and Tool Creation - The "Lingguang" assistant allows users to generate interactive applications, transforming them from passive consumers to active tool creators [41][44] - This model addresses long-tail demands that traditional app development cannot meet due to high costs and long timelines [42][43] - The focus on problem-solving is crucial for user retention, as it emphasizes the practical utility of AI [45] Group 5: Strategic Framework - Ant Group's AI strategy is characterized by a systematic approach, with "Lingguang" and "AQ" (AI health manager) serving as key components of its broader vision [52][60] - The underlying technology, including the Ant Bailing Large Model, ensures strategic autonomy in capability, safety, and cost [54][55] - The expansion into physical applications through "Lingbo Technology" reflects the ambition to extend AI's service capabilities beyond the digital realm [58][59] Conclusion - The AI entry battle is becoming increasingly nuanced, with Ant Group's "Lingguang" introducing a significant variable by transforming AI from a knowledge provider to a capability enhancer [63][64] - The ultimate success in the AI landscape will depend on the ability to solve real user problems rather than merely attracting attention [65]

Core Insights - Microsoft, Nvidia, and Anthropic have formed a strategic partnership, creating a tightly-knit "AI Alliance" that binds capital, computing power, and models together [1][4] - The partnership involves significant investments, with Microsoft committing up to $5 billion and Nvidia up to $10 billion in Anthropic [2] - Anthropic's valuation has surged to $350 billion, marking an expansion of "closed-loop" investments in the AI sector [4] Investment and Collaboration Details - Anthropic will purchase $30 billion worth of Azure computing power from Microsoft and has signed contracts for up to 1 gigawatt of additional computing power, all running on Nvidia's AI systems [3][7] - This collaboration represents Nvidia's first deep technical partnership with Anthropic, aimed at optimizing Anthropic's products for performance and efficiency [5][6] - Anthropic's initial commitment includes acquiring up to 1 gigawatt of computing power, utilizing Nvidia's advanced architectures [7][8] Product and Market Strategy - Microsoft and Anthropic are expanding their collaboration to provide broader access to Anthropic's Claude models for enterprise users [9] - Azure AI Foundry customers will have access to Anthropic's Claude models, making Claude the only cutting-edge LLM model available on the three major cloud platforms [10] - Microsoft will continue to integrate Claude into its Copilot product suite, enhancing its offerings alongside OpenAI's models [11][12] Infrastructure and Growth Plans - Founded by former OpenAI employees, Anthropic is accelerating its infrastructure development, planning to invest $50 billion in custom data centers across the U.S. [13] - Anthropic has also secured a deal with Google to supply up to 1 million AI chips, significantly boosting its computing capabilities [13] Market Concerns and Risks - The announcement of this partnership comes amid rising skepticism about the AI investment boom, with Nvidia and Microsoft's stock prices dropping nearly 3% on the day of the announcement [14] - Concerns about a potential "AI bubble" are prevalent, with 45% of fund managers viewing it as a major risk, as the closed-loop investment model raises questions about the sustainability of AI products generating sufficient revenue [17]

Core Insights - Google has officially launched its most powerful AI model, Gemini 3, which is now integrated into various products and platforms, marking a significant advancement in AI capabilities [1][2][3] - Gemini 3 is positioned as a leading model in several performance benchmarks, showcasing its advanced multi-modal understanding and code generation abilities [2][3] - The evolution from AI as a conversational partner to a digital colleague signifies a major paradigm shift in human-AI collaboration, where humans will provide strategic guidance rather than correcting basic errors [4][28] Group 1: Model Performance and Capabilities - Gemini 3 is described as the best multi-modal understanding model and the most powerful agent and code generation model developed by Google [3] - The model demonstrates remarkable "agent" capabilities, allowing it to perform complex tasks such as writing code, building interactive applications, and executing multi-step tasks [3][4] - The transition from "description" to "action" indicates that AI is evolving into a general-purpose tool capable of completing actual work rather than just generating text [4][28] Group 2: Practical Applications and Tools - Alongside Gemini 3, Google introduced Antigravity, a tool that enables AI to autonomously write programs and perform tasks on computers, fundamentally changing the nature of what it means to "code" [11][12] - Users can interact with Gemini 3 in natural language, allowing it to manage tasks such as creating web pages and summarizing information, which enhances productivity [12][14] - The AI's ability to handle complex data analysis and generate original research papers demonstrates its advanced cognitive capabilities, comparable to a graduate-level researcher [18][25] Group 3: Future Implications and Trends - The emergence of Gemini 3 signifies a shift from the "chatbot era" to the "digital colleague era," where AI will play a more integral role in professional environments [28] - The model's imperfections reflect a transition to more nuanced human-like errors, indicating that AI is becoming more sophisticated in its decision-making processes [17][25] - The need for human oversight is emphasized, as users will need to guide and verify AI actions, marking a significant change in the relationship between humans and AI [28][31]