图片来源： Y Combinator Z Highlights 本期节目于 2025 年 12 月 22 日在 Y Combinator 播出 ，由 Jared Friedman 、 Diana Hu 、 Harj Taggar 、 Garry Tan 共同主持，长期围绕 AI 创业、模型演 进与 YC 内部一线观察展开讨论。在 YC Winter26 批次结束后，四位合伙人围绕 AI 经济是否进入 " 泡沫期 " 、 LLM 偏好变化以及应用层机会展开系统复 盘。 Anthropic 是最受欢迎的 Garry Tan ： 欢迎回到新一期的《 The Light Cone 》。今天我们来聊聊在 2025 年我们看到的最令人惊讶的事情。 Diana ，你发现了一个非常疯狂的事 情。几乎可以算是在 YC 中， YC 批次里，谁是首选 LLM 的一次换岗。 Diana Hu ： 是的，事实上，我们刚刚结束了针对公司的冬季 26 期筛选周期。我们向所有申请 YC 的创始人提出的问题之一是，你们的技术栈和选择的模 型是什么？ 令人震惊的是，在很长一段时间里， OpenAI 都是明显的赢家。但在去年的所有批次中，这个数字 ...

Core Insights - The article discusses the rising prominence of Agent companies in the AI sector, highlighting their challenges and opportunities as they navigate the market landscape leading up to 2025 [6]. Group 1: Agent Companies and Market Trends - The acquisition of Manus by Meta for over $2 billion is seen as a milestone event in the Agent space, sparking diverse interpretations within the industry [7]. - The concept of "Situated Agency" is introduced, emphasizing that an Agent's capabilities are deeply intertwined with its environment, tools, and memory [7]. - The market acceptance of Agents has surged, with 52% of companies using generative AI deploying Agents in production environments [8]. Group 2: Investment and Capital Flow - Over 20 U.S. Agent startups raised over $100 million in funding in the past year, covering various sectors such as programming, B2B customer service, healthcare, and legal [11]. - Harvey, a legal-focused Agent company, completed a $160 million Series E funding round, achieving a valuation of $48 billion [11]. - ElevenLabs raised $100 million to transition towards a conversational Agent platform, indicating a shift in focus towards dialogue-driven applications [12]. Group 3: Sector-Specific Developments - In the legal sector, EvenUp raised $150 million to automate routine legal tasks, showcasing the growing interest in legal technology [11]. - In the search domain, Parallel and You.com both secured over $100 million in funding, reflecting the demand for Agent capabilities in search infrastructure [12]. - The healthcare sector is also seeing significant investment, with companies like Abridge raising $300 million to develop clinical dialogue Agents [15].

电影《大空头》原型、知名投资者Michael Burry在周五收盘后通过Substack发文披露，他持有甲骨文的看跌期权，并在 过去六个月内直接做空了该公司。Burry此前已公开做空英伟达和Palantir，此次对甲骨文的押注进一步强化了他对当前 AI市场估值过高、泡沫严重的判断。 Burry直言不讳地批评了甲骨文目前的战略定位及其激进的投资行为。他认为，这家老牌数据库软件公司为了追逐AI热 点，正在进行不必要的重资产扩张，试图通过昂贵的数据中心建设与亚马逊、微软等云巨头竞争。 Burry在回复读者提问时尖锐地指出："我不喜欢它的定位，也不喜欢它正在进行的投资。它不需要做现在正在做的事 情，我不知道它为什么要这样做。也许是出于虚荣心（Ego）。" "纯粹的AI泡沫载体" Burry的做空逻辑核心在于，他认为甲骨文缺乏其他科技巨头那样的安全垫，是一个脆弱的"纯粹AI泡沫载体"。 在解释为何选择做空甲骨文而非其他科技巨头时，Burry详细阐述了他的筛选标准。他指出，微软（Microsoft）、谷歌 母公司Alphabet和Meta等公司拥有强大的核心业务护城河，即使AI投资失败，它们依然能生存并保持主导地位。 " ...

Group 1: AI Market Dynamics - The AI economy has established a stable structure with parallel layers of models, applications, and infrastructure, each with considerable profit potential [1] - Investment in AI infrastructure and energy, perceived as a bubble, actually provides affordable computing power and "excess dividends" for the application layer [1] Group 2: LLM Power Shift - By 2025, Anthropic's Claude has surpassed OpenAI's ChatGPT as the most popular large language model (LLM) among Y Combinator projects, indicating a significant shift in market preference [5][6] - The structural change in technology stack and model selection is evident, with OpenAI's market share declining from over 90% [5] Group 3: Developer Relations and Product Philosophy - Anthropic is characterized by a "golden retriever energy," emphasizing a friendly and cooperative approach towards developers, contrasting with OpenAI's more aloof stance [6][7] - This developer-centric design has translated into competitive advantages, particularly in programming assistance, making Anthropic the preferred choice for many founders [8] Group 4: Spillover Effects and Programming Paradigms - Founders' preference for Claude in personal programming contexts leads to a spillover effect, influencing their choice of models for unrelated applications [9] - The concept of "Vibe Coding" has evolved from a qualitative observation to a significant technical domain, demonstrating commercial viability through successful companies like Replit and Emergent [10] Group 5: Team Structure and Efficiency - The measure of company success is shifting from team size to per capita output efficiency, with examples like Gamma achieving $100 million in annual recurring revenue (ARR) with a streamlined team of 50 [12] - The rise of AI has increased productivity but also heightened customer expectations, making talent execution the new bottleneck in a competitive landscape [11] Group 6: Trust Crisis and Specialized Applications - To address complex tasks and build user trust, AI development is shifting focus from general large models to specialized applications capable of executing specific logic [13]

从市场角度来看，当前中美 AI 相关板块并未处于典型的泡沫狂热期，不必过度担忧。同时，技术创新 与产业落地的本质是非线性演进过程，绝非一蹴而就的短期行为。与商业周期相似，科技投资也存在明 显的周期特征。资产的泡沫不会阻断核心技术发展的大趋势，而寻找那些无惧市场泡沫、能够穿越周期 的企业和资产才是关键。 从产业路径来看，中国AI发展路径与美国形成明显分化，根源在于所处的资源约束条件不同。中国所 面临的硬约束在于尖端训练芯片进口受限以及高端 GPU 供给阶段性稀缺，因此建立大规模算力集群在 经济性与可持续性上显然难以与美国比肩。在此情形下，单纯复制美国的算力堆叠模式，并不具备现实 可行性。与其追随美国进行算力军备竞赛，不如选择一条更具中国特色、符合国情的发展道路，即在有 限算力条件下挖掘最优效果，产业界的高性价比模型和开源模型就是这一思路下的产物。例如， DeepSeek 等产品不靠算力堆叠，而是在工程体系优化和精准建模中持续降低成本并提升性能，体现出 用工程能力弥补算力限制的方法论选择。得益于相对较低的 token 使用价格，中国大语言模型如 DeepSeek、通义千问等在全球范围内广受欢迎，过去一年的使用量位 ...

去年年初国产大模型深度求索（DeepSeek）的横空出世，让资本市场掀起了人工智能（AI）投资 热潮。 银河证券认为，当前AI市场呈现出复杂特征。一方面，以"AI+"为核心的科技革命是确定性产业趋 势，巨额资本的持续注入既反映了全球对技术突破与未来价值的共识，也反映出技术迭代真实发生、应 用需求切实存在、产业逻辑清晰可循；另一方面，AI龙头公司估值屡创新高，产业链各环节资本开支 也空前高涨。但整体来看，当前AI资本市场相较于2000年互联网泡沫时期，更偏向于产业革命的早 期。 展望AI投资机会 展望未来，中国经济格局正在迅速转变。2025年10月发布的"十五五"规划建议强调发展新质生产 力，加快先进制造、AI、航空航天、可再生能源和量子科技等领域的发展。 具体投资方面，平安基金研究总监兼基金经理张晓泉表示，预计今年AI的硬件端表现将优于应用 端。他认为，虽然AI取得了重大进展，但是目前全球最领先的模型仍未能达到AGI程度，因此，AI应用 落地仍存在不确定性。反观模型优化与迭代依然是相对确定趋势，而模型能力的持续提升本质上仍依赖 于底层AI硬件的持续投入与升级。 随着2025年国家出台一系列支持人工智能发展的政 ...

以下文章来源于起朱楼宴宾客 ，作者大卫翁 起朱楼宴宾客 . 金融狗记录这个大时代，播客同名 除了个别失意者，比如数字货币、各国国债以及石油之外，从股市到贵金属到资源品，从美股到A股 再到日欧甚至拉美股市，从量化到主动权益再到CTA，这是个涨声一片的年份。 猜中这个结果的人并不多，毕竟我们在年中经历了太多大起大落。而如果复盘我在年初对2025的几 个猜想，结论也是在这个大时代中，做猜想可真是越来越难了。我已经尽量讨巧地描绘出一些比较大 而化之的场景，没想到还是有好几个错得离谱。 比如第一个关于政府化债，事实上这应该不是解决中国问题的扣，它只是一个起始点。 本文来自微信公众号： 起朱楼宴宾客 ，作者：大卫翁，头图来自：视觉中国 去年的猜想文开头 ，我说"期待能再拥有一个各种资产的全年涨幅都和2024差不多的一年"，事实证 明我的格局还是不够打开。 2025年，我们拥有了一个对于投资而言称得上"梦幻"的年份。 2025年从春节的Deepseek Moment开始，关税战、反内卷，再到下半年对AI泡沫的争论不休和房地 产的彻底沉沦，市场关注的焦点不断切换，但化债已经不再是其中的一个。 它是不是停止了呢？当然也不是，根 ...

英伟达CEO黄仁勋5日宣布，最新Rubin平台AI芯片已进入量产，计算速度将是Blackwell平台倍数以上 增长，且成本更低，亚马逊AWS、微软、Google等云端服务供应商（CSP）最快下半年导入。 英伟达再推"地表最强AI芯片"，开启AI全新世代。法人看好，台积电、鸿海、广达及纬创等供应链将跟 著衝一波，2026年仍将是丰收的一年。 黄仁勋强调，Rubin打造的AI服务器与以往最大不同就是省去大量缆线，因此机壳也端出全新设计，让 组装一柜服务器的时间大幅缩短，若以10％的液冷AI服务器来计算，最短时间仅需五分钟。 面对AI泡沫言论，黄仁勋高喊："AI的竞赛已经开跑"，目前价值10万亿美元的传统科技市场正在大幅转 型，未来都将以AI为基础进行升级，"每个人都在努力迈向下一阶段"，预期AI计算模型及市场需求仍持 续爆炸性增长，使得AI基础建设市场呈现快速上升趋势，Rubin平台的问世将会替AI计算能力带来新一 代标准。 英伟达Rubin平台芯片由台积电以3nm独家操刀，并整合CoWoS先进封装制程。业界传出，台积电 CoWoS先进封装制程已经满到今年底，所有协力封测厂也开始全力支持，后续将有望同步受惠。 黄 ...

1. Report Industry Investment Rating No relevant content provided. 2. Core Viewpoints of the Report - In 2025, the equity market showed an upward trend with significant fluctuations. The CSI All-Share Index recorded a 24.60% annual return, and the Shanghai Composite Index climbed to around 4000 points. The Evergreen Low-Volatility Strategy maintained low volatility and achieved nearly 15% return, while the Stock Fund Enhancement Strategy was limited by balanced fund selection but could still optimize the risk-return ratio. The Overseas Equity Allocation Strategy recorded a 13.98% return [3]. - The quantitative strategy allocation preference is Stock Fund Enhancement Strategy > Evergreen Low-Volatility Strategy > Overseas Equity Strategy. The report is optimistic about A-share investment opportunities, suggesting active short - term layout. The Evergreen Low-Volatility Strategy can be used as a base position, and the Overseas Equity Strategy still has some allocation value but with more risk points [4]. 3. Summary by Relevant Catalogs 3.1 Quantified Strategy Allocation Viewpoints - A-shares: The report is bullish on A-share investment opportunities. The Shanghai Composite Index has exceeded 4000 points, and the trading volume has increased in 2026. The market is expected to break through further in the spring offensive. The Stock Fund Enhancement Strategy is more attractive, and the Evergreen Low-Volatility Strategy can be used as a base position [4]. - Overseas market: After recovering from the tariff shock, the upward momentum of the overseas market has weakened and differences have increased. Although there is still some allocation value in the short term, there are many risk points, and the overall cost - effectiveness is slightly lower than that of A-share strategy combinations [4]. 3.2 Performance of Equity Fund Strategy Indexes - Evergreen Low-Volatility Fund Strategy Index: This week's return was -0.705% with an excess return of 0.283%. The one - month return was 0.717% with an excess return of -1.293% [5]. - Stock Fund Enhancement Fund Strategy Index: This week's return was -0.229% with an excess return of 0.759%. The one - month return was 2.294% with an excess return of 0.284% [5]. - Cash Enhancement Fund Strategy Index: This week's return was 0.019%, outperforming the CSI Money Fund Index with a cumulative excess return of 0.569% since the strategy's operation [5]. - Overseas Equity Allocation Fund Strategy Index: This week's return was -0.137% with an excess return of 0.186%. The one - month return was 1.834% with an excess return of -1.417% [5]. 3.3 Tool - based Fund Portfolio Performance Tracking - Evergreen Low-Volatility Fund Portfolio: It has maintained low volatility and small drawdowns, significantly outperforming the CSI Active Equity Fund Index. It has both defensive and offensive characteristics [13]. - Stock Fund Enhancement Fund Portfolio: Since the strategy has been running for a short time, its performance is similar to that of the CSI Active Equity Fund Index. It is expected to have stronger elasticity in an improved market environment [16]. - Cash Enhancement Fund Portfolio: After double screening, it has continuously outperformed the benchmark, with a cumulative excess return of over 0.41% since July 2023 [17]. - Overseas Equity Allocation Fund Portfolio: Since July 2023, in the context of the Fed's interest - rate cut cycle and the boost of AI technology, it has accumulated high - level excess returns, which can thicken the return of the equity investment portfolio [20]. 3.4 Tool - based Fund Portfolio Construction Ideas - Evergreen Low-Volatility Fund Portfolio: It aims to select funds with long - term stable returns in high - equity - position actively managed funds. By adding restrictions on fund valuation levels, it constructs a low - volatility active equity fund portfolio from the perspectives of net value performance and position characteristics [22][25]. - Stock Fund Enhancement Fund Portfolio: It aims to meet the needs of equity fund investors with different risk preferences. By analyzing the source of fund returns and constructing a portfolio based on the significant continuity of Alpha returns, it is expected to bring excess performance [23][26]. - Cash Enhancement Fund Portfolio: By constructing a money fund screening system, it helps investors select money funds with better returns, optimize the income of short - term idle funds, and reduce income volatility risks [23][27]. - Overseas Equity Allocation Fund Portfolio: Based on long - and short - term technical indicators and comprehensive index momentum and reversal effects, it selects QDII equity funds corresponding to overseas equity indexes to build a portfolio, meeting the needs of global allocation [24][28].

回首刚刚过去的2025年，尽管表面上美股和AI概念似乎是最大的赢家，但在达利欧看来，真正的大故 事其实是法币的集体贬值和黄金的强势崛起。财政刺激、地缘政治与贫富分化共同重塑了全球经济版 图，展望2026，三重风险正交织…… 下文为全球最大基金桥水基金创始人瑞·达利欧（Ray Dalio）发布的原文。 作为一名系统化全球宏观投资者，在告别2025年之际，我自然会反思过去一年发生的机制，尤其是市场 方面的表现。 尽管事实和回报率是无可争议的，但我看问题的角度与大多数人不同。虽然大多数人认为美股，特别是 美国AI概念股是最好的投资，并因此将其视为2025年最大的投资故事，但无可争辩的事实是，回报最 高（因此也是最大的故事）其实源于： 1）货币价值的变化（最重要的是美元、其他法币和黄金的表现）； 2）主要由于财政和货币刺激、生产率提高以及资产配置大幅撤离美国市场，导致美股表现 显著落后于非美股票市场和黄金（黄金是表现最好的主要市场）。 在这些思考中，我将退后一步，审视去年这种货币/债务/市场/经济的动态是如何运作的，并简要触及其 他四大力量——政治、地缘政治、自然灾害和技术——是如何在不断演变的"大周期"背景下影响全 ...