【环球网财经综合报道】近日，伴随着市场对"AI泡沫"的担忧加剧，美国科技股正持续遭到猛烈抛售，投资者信 心受到严峻考验。然而，在此背景下，华尔街的目光正悄然转向东方。一位管理着6.6万亿美元资产的瑞银全球财 富管理高管最新发声，高呼是时候将目光投向中国的人工智能股票了。 瑞银全球财富管理美洲区资产配置主管Jason Draho日前明确表示，中国科技板块提供了一种颇有吸引力的方式来 平衡美国科技股持仓，尤其是在当前美国科技股估值过高、市场疑虑逐渐升温的背景下。他强调了一个关键洞 察："虽然看起来中美科技板块可能会走势同步，但实际上二者相关性较低。"Draho解释说，这种低相关性源于双 方存在的竞争关系以及背后驱动因素的差异。他举例称，今年早些时候中国深度求索（DeepSeek）聊天机器人的 发布，就曾导致美国科技股的显著下跌。"如果中国科技或人工智能模型最终表现出色，就能起到多元化配置的作 用，"Draho在接受采访时表示，"它们背后的驱动因素既有国内政策，也有技术本身的差异，因此不仅可能跑赢各 自市场，走势也可能在一定程度上相互独立。" 在过去几天内，摩根士丹利和摩根大通等顶级投行也纷纷发表看涨观点。摩根士丹 ...

专题：申万宏源2026资本市场投资年会 11月19日，申万宏源证券2026资本市场投资年会在上海举办， 申万宏源研究资产配置资深高级分析师 冯晓宇，发表申万宏源2026年美股投资策略，当前市场对"AI泡沫"的担心主要是新技术带来经济规模效 应的时间vs资本市场对新技术投资回报的乐观期待之间的落差。 自2023年以来，本轮美股AI行情已经持续三年时间。2023~2024年美股AI行情以提估值为主，AI基础设 施以及有业绩的AI应用普涨。2025年AI相关行业（科技、通信）估值贡献相对有限，并且债务融资支 撑投资比例增加，在AI产业投资的ROI（投资回报率）尚不明确的阶段，市场进入对效益审视更严苛的 阶段。 当前市场对"AI泡沫"的担心主要是新技术带来经济规模效应的时间vs资本市场对新技术投资回报的乐观 期待之间的落差。从技术发展周期来看，新技术往往有助于生产效率的提升，不过生产力的提升（释放 新需求）需要技术成本系统性下降、企业组织架构调整、就业结构调整，这需要以年为单位的时间。这 个过程中，资本市场往往以乐观思维推演新技术的终局效用。而当市场看到生产力转化过程中的边界 时，股价泡沫往往出现破裂。对于边界的衡量 ...

Core Viewpoint - The AI industry is experiencing a mix of optimism and concern regarding a potential bubble, with various industry leaders expressing differing opinions on the sustainability and valuation of AI investments [1][2][5][6]. Group 1: Opinions on AI Bubble - Sam Altman, CEO of OpenAI, believes the industry is in an AI bubble, suggesting that current excitement may be excessive despite AI's transformative potential [2]. - Bill Gates acknowledges the existence of a bubble but compares it to the internet bubble, indicating that while there is hype, AI technology holds significant value [2]. - Mark Cuban does not see similarities with the internet bubble, emphasizing the quality of AI companies and the absence of speculative IPOs [3]. - Mark Zuckerberg suggests that a bubble could form if companies fail to make continuous progress, but ongoing advancements may prevent a collapse [4]. - Jensen Huang, CEO of NVIDIA, asserts that there is no bubble, framing the current transition as a natural evolution in computing [4]. - Sundar Pichai, CEO of Alphabet, recognizes irrational elements in the market but believes that no company, including Google, would be immune to a potential bubble burst [4]. - Jeff Bezos describes the current situation as an "industry bubble," highlighting the challenge of distinguishing between quality and poor investments [5]. - Daniel Pinto from JPMorgan anticipates a market correction, citing overvaluation of AI companies [6]. - Bret Taylor, chairman of OpenAI, aligns with Altman, warning of significant losses for many investors [7]. - Eric Schmidt, former CEO of Google, suggests that what appears to be a bubble may actually represent a new industry structure [8]. - Pat Gelsinger, former CEO of Intel, acknowledges a bubble but believes it won't burst in the near future, as companies are just beginning to realize AI's benefits [8]. - Joe Tsai, co-founder of Alibaba, expresses concern over the speculative construction of data centers, indicating signs of a bubble [9]. - Lisa Su, CEO of AMD, argues that bubble concerns are misguided, advocating for a long-term view on AI's transformative potential [10]. - Ray Dalio compares the current cycle to the late 1990s, warning against conflating technological success with investment success [10]. - Thomas Siebel, CEO of C3.ai, claims that a significant bubble exists, particularly criticizing the overvaluation of OpenAI [10].

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]

据追风交易台消息，美银18日发布的全球基金经理调查显示，基金经理们的股票配置升至2025年2月以来的最高水平， 但现金持仓降至仅3.7%，触发该行"卖 出信号"， 市场担忧过度看涨的仓位可能成为风险资产的阻力。 然而，乐观情绪的背后是日益增长的风险警报。 45%的受访者将"AI泡沫"视为当前市场的最大尾部风险， 这一比例较上月显著上升。同时，"做多七巨头"以 54%的得票率重回"最拥挤交易"的榜首，凸显了市场资金在大型科技股上的高度集中。 市场情绪正处于一个微妙的十字路口。美银的调查显示，尽管投资者对经济前景的乐观情绪升温，但对人工智能驱动的投资狂热及其潜在泡沫的担忧也在同步 加剧，同时，一个长达20年未见的警示信号——企业过度投资——已经出现。 尽管宏观情绪本月有所改善，全球增长预期年内首次转正，53%的投资者预测经济将软着陆，但创纪录的63%受访者认为股市当前估值过高。 美银警告，这一转变的背后，是对AI领域资本支出热潮的规模及其融资方式的普遍担忧。 若无美联储降息等新的催化剂，风险资产可能面临回调压力。 AI泡沫成头号风险，20年来首现企业"过度投资"警告 美银调查显示，在经历了数月的市场狂热后，基金经 ...

Group 1: Market Overview - The US stock market experienced a significant adjustment on November 18, with all three major indices closing lower, led by a decline in technology stocks. The Dow Jones Industrial Average fell by 498.5 points, a decrease of 1.07%, closing at 46091.74 points; the S&P 500 dropped by 0.83% to 6617.32 points; and the Nasdaq Composite Index decreased by 1.21%, ending at 22432.85 points. Concerns over technology stock valuation bubbles and a cooling expectation for Federal Reserve interest rate cuts exerted dual pressure on the market [1][2]. Group 2: Technology Sector Performance - The technology sector faced severe losses, with major tech stocks generally declining. Amazon's stock plummeted by 4.43%, resulting in a market value loss of approximately 784 billion RMB, making it one of the most significant losers among tech giants. Additionally, semiconductor leader Nvidia fell by 2.81%, and software giant Microsoft dropped by 2.70%. The Philadelphia Semiconductor Index also declined by 2.31% [2]. - The market showed a mixed performance among popular Chinese concept stocks, with the Nasdaq Golden Dragon China Index rising by 0.42%. iQIYI saw a notable increase of over 6%, while Pinduoduo's stock fell by more than 7% due to slowing growth, making it one of the largest decliners among Chinese concept stocks. Analysts noted that 45% of fund managers view the "AI bubble" as the biggest tail risk in the current market, with ongoing concerns about excessive investments in AI projects by tech companies [2]. Group 3: AI Industry Developments - Amid the downturn in tech stocks, significant collaboration emerged in the AI sector. Microsoft, Nvidia, and Anthropic, a major competitor of OpenAI, formed a three-party alliance, with Anthropic planning to invest $30 billion in cloud computing power from Microsoft Azure over the next decade, all supported by Nvidia hardware. In return, Nvidia committed to investing up to $10 billion in Anthropic, while Microsoft pledged up to $5 billion. This collaboration boosted Anthropic's valuation to approximately $350 billion, nearly doubling from $183 billion in September [5]. - This $45 billion partnership creates a unique "symbiotic cycle": Microsoft's investment will flow back through Anthropic's cloud power purchases, while Nvidia solidifies its hardware dominance by binding with an AI leader. Anthropic aims to implement a "dual cloud strategy" to reduce reliance on a single cloud service provider, further diluting OpenAI's industry influence. Market attention is now focused on Nvidia's upcoming Q4 fiscal year 2025 earnings report, which is expected to reshape market expectations for the AI industry, with projected revenues of $39.3 billion (up 78% year-on-year) and data center revenues of $35.6 billion (up 93% year-on-year) [5].

智通财经APP获悉，美国科技巨头谷歌(GOOGL.US)在推出Gemini 2.5近八个月之后，在周二重磅宣布 推出升级版人工智能大模型Gemini 3，在与ChatGPT开发商OpenAI的激烈竞逐中加快产品更新迭代步 伐，并且于发布首日立即在谷歌搜索、Gemini应用程序生态及多个开发者平台同步上线，在谷歌旗下的 多个盈利产品中投入使用。谷歌表示，在衡量人工智能模型性能的多个热门行业排行榜上，Gemini 3处 于绝对领先地位，该模型还在数学、多模态理解和事实准确性等多个维度刷新了大模型业内纪录。 谷歌表示，其最新一套基于AI大模型的AI应用产品组合将让用户们在获得理想结果时需要进行的"提示 (prompt)"变得更少。谷歌还发布了一款名为"Google Antigravity"的全新AI智能体平台，让全球开发者们 可以在"更高层级且面向系统性任务"的层面进行AI编码。 谷歌发布于周二的一系列博客文章之一中，谷歌母公司Alphabet的首席执行官桑达尔·皮查伊(Sundar Pichai)表示，这一新的AI大模型将允许用户们对更复杂的问题获得更好的答案，"这样你就可以在更少 提示的情况下获得所需内容"。 ...

Core Viewpoint - The strategic partnership between Microsoft, NVIDIA, and Anthropic represents a significant investment in AI, with concerns about the sustainability of such "closed-loop" financing models in the industry [2][12]. Group 1: Partnership Details - Microsoft commits to invest up to $5 billion in Anthropic, while NVIDIA pledges up to $10 billion, leading to Anthropic's valuation soaring to $350 billion [3][5]. - Anthropic will purchase $30 billion worth of Azure computing power from Microsoft, fully utilizing NVIDIA's AI systems, including the current Grace Blackwell architecture and the upcoming Vera Rubin architecture [5][11]. Group 2: Strategic Implications - The collaboration expands Microsoft's AI offerings, allowing Azure customers access to Anthropic's Claude models, which will be available on major cloud platforms [8][9]. - Despite this new partnership, Microsoft maintains its core relationship with OpenAI, having invested heavily in it since 2019, with OpenAI's valuation reaching $135 billion [9][10]. Group 3: Market Concerns - 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 [13]. - Concerns about a potential AI bubble are prevalent, with 45% of fund managers identifying it as a significant market risk, questioning whether AI products can generate sufficient revenue to justify the massive investments [12][15].

（原标题：?AI浪潮仍然汹涌! 从巴菲特"最后一舞"到Gemini 3震撼问世 谷歌托起"AI牛市叙事"） 智通财经APP获悉，美国科技巨头谷歌(GOOGL.US)在推出Gemini 2.5近八个月之后，在周二重磅宣布推出升级版人工智能大模型Gemini 3，在与 ChatGPT开发商OpenAI的激烈竞逐中加快产品更新迭代步伐，并且于发布首日立即在谷歌搜索、Gemini应用程序生态及多个开发者平台同步上 线，在谷歌旗下的多个盈利产品中投入使用。谷歌表示，在衡量人工智能模型性能的多个热门行业排行榜上，Gemini 3处于绝对领先地位，该模 型还在数学、多模态理解和事实准确性等多个维度刷新了大模型业内纪录。 谷歌表示，其最新一套基于AI大模型的AI应用产品组合将让用户们在获得理想结果时需要进行的"提示(prompt)"变得更少。谷歌还发布了一款名 为"Google Antigravity"的全新AI智能体平台，让全球开发者们可以在"更高层级且面向系统性任务"的层面进行AI编码。 谷歌发布于周二的一系列博客文章之一中，谷歌母公司Alphabet的首席执行官桑达尔·皮查伊(Sundar Pichai)表示，这一新的 ...

热门中概股多数上涨，纳斯达克中国金龙指数涨0.42%，阿里巴巴涨1.27%，京东跌0.30%，拼多多跌 7.33%，蔚来汽车跌1.64%，小鹏汽车涨0.85%，理想汽车跌0.80%，哔哩哔哩涨0.34%，百度涨2.66%， 网易涨1.06%，腾讯音乐跌1.06%，小马智行涨2.30%。 高企的估值以及为建设数据中心而不断累积的债务，使投资者开始质疑人工智能（AI）相关股票的定 价。 一项最新调查显示，45%的基金经理认为"AI泡沫"是当前市场最大的"尾部风险"，并出现了对企业过度 支出（尤其是AI项目投入）的担忧激增。 英伟达最新财报，隐含3200亿美元市值波幅 AI芯片龙头英伟达下跌近3%，英伟达本月已下跌超10%，市场正等待其将在周三收盘后公布的三季度 业绩。 美东时间周二，投资者对科技股估值过高的担忧情绪继续发酵，三大指数集体收跌，纳指和道指跌幅均 超过1%。 截至收盘，道琼斯指数跌1.07%，报46,091.74点；标普500指数跌0.83%，报6,617.32点；纳斯达克指数 跌1.21%，报22,432.85点。 科技股全线大跌 大型科技股普跌，英伟达跌2.81%，苹果跌0.01%，微软跌2.70 ...