Morgan Stanley’s Skelly Backs Tech Capex Over Small-Cap Equities

摩根士丹利Skelly看好科技资本支出而非小型股 企业人工智能带来的生产力提升与持续的高收益率，更有利于现金充裕的超大市值企业，而非面临脱媒风险的未盈利小型公司。 Dan Skelly深入剖析了摩根士丹利财富管理公司的宏观展望，重点阐述了尽管存在宏观经济不确定性，美国企业盈利增长与人工智能驱动的经营杠杆仍在持续支撑市场动能。讨论深入分析了当前AI资本支出与互联网泡沫时期的结构性差异、未盈利小型股的局限性，以及信贷市场分化的运行机制。Skelly还概述了其团队包含50只股票的投资组合策略，该策略在优质核心持仓与涵盖医疗保健、市场基础设施和长期科技龙头等高确信度的“重大构想”（Big Ideas）之间寻求平衡。 摩根士丹利财富管理业务经历大幅结构性扩张，占全公司总营收的比重从2005年的8%跃升至如今约60%，总管理规模达16至17万亿美元。 (0:00) 美国经济展现出强劲韧性，标普500指数第二季度盈利同比增长28%（中位数公司增长14%），多数企业在保住提价红利的同时通过应用AI工具实现了无额外雇佣下的生产力跃升。 (7:10) 美股小盘股存在显著的负向选择偏差，约80%营收过亿美元的本土企业选择留在私募市场，导致罗素2000指数中近40%的公司处于未盈利状态，面临AI颠覆与缺乏资本开支能力的双重压力。 (20:35) 摩根士丹利预计到2027年全球AI年化资本支出将攀升至1.5万亿美元，但数据中心电网扩容与发电设备等交付瓶颈可能诱发硬件供应链的“重复下单”风险。 (23:41) 标普500成分股之间的相关性降至近40年低点，市场正在严惩毛利率收缩的公司；与1999年互联网泡沫不同，当下的AI算力开支主要由现金充沛的成熟超大规模云厂商驱动。 (35:00)

Morgan Stanley Wealth Management portfolio manager Dan Skelly argues that resilient corporate earnings and AI-driven operating leverage support U.S. equities, justifying a disciplined, bottom-up equity strategy that prioritizes high-quality secular compounders over capital-constrained small caps. AI Productivity and Synthetic Operating Leverage (7:10) According to Dan Skelly, corporate America's earnings resilience is anchored in structural operational efficiencies rather than standard cyclical tailwinds. Second-quarter earnings showed 28% year-over-year index-level growth and a 14% median company increase, with revenue expanding 15% against a 6% nominal GDP backdrop without expanding corporate headcount. Companies have preserved post-COVID pricing power while tapping a 'synthetic operating leverage boost' through AI tools, operating in a temporary 'no hiring, no firing zone.' This productivity push also intersects with long-term labor dynamics. Demographic declines and an impending enrollment cliff in higher education are creating structural shortages across entry-level auditing, nursing, and industrial labor. Skelly and the panel view automation and robotics not as displaced-labor risks, but as essential structural offsets needed to sustain domestic manufacturing reshoring and corporate profitability. Capital Discipline Across the AI Infrastructure Buildout (35:00) Skelly challenges fears of a 1999-style tech bubble by contrasting end-customer profiles. Dot-com infrastructure providers like Cisco and Lucent sold equipment to unprofitable startups that disappeared overnight. Today, infrastructure is funded by cash-rich hyperscalers servicing established enterprise clients across healthcare, government, and manufacturing. Furthermore, unutilized computing assets can be absorbed directly into enterprise cloud migrations, which Skelly estimates are only in their fourth or fifth inning. Risks remain concentrated in physical deployment constraints rather than raw customer demand. With global data centers requiring an anticipated 40 to 50 gigawatts of power capacity over two to three years, electrical transformer shortages and grid delays risk causing supply chain double-ordering. Morgan Stanley analyst Brian Nowak projects annual AI capital expenditures will reach $1.5 trillion by 2027 (versus the $1.2 trillion consensus), and tech firms face roughly $320 billion in investment-grade debt and special purpose vehicle issuances by 2026. Small-Cap Fragility and Fundamental Market Discrimination (20:35) While macroeconomic expectations of interest rate cuts have driven tactical interest into small-cap stocks, Skelly highlights deep structural vulnerabilities in public benchmarks. Over the past 15 years, private equity monetization has kept approximately 80% of U.S. firms with over $100 million in revenue off public exchanges, leaving nearly 40% of the Russell 2000 unprofitable. This negative selection bias leaves small public firms exposed to a dual AI threat: lower-tier business models face direct disintermediation, while unprofitable balance sheets lack the capital to invest in enterprise AI productivity. Meanwhile, three-month intra-stock correlations in the S&P 500 near 40-year lows indicate that public markets are actively discriminating, aggressively penalizing companies that suffer gross margin contractions exceeding 1%. The 50-Stock Portfolio: Core versus 'Big Ideas' (64:57) Skelly's equity strategy traces its lineage to a framework established in 1995 by former Morgan Stanley strategist Byron Wien. The strategy runs a concentrated 40- to 50-stock portfolio divided into two distinct components: 40 to 45 'Core Holdings' sized at roughly 2% each to limit tracking error, and 5 to 6 high-conviction 'Big Ideas' targeted at multi-year secular transformations, such as early positions in digital payments or semiconductor leaders. Sector tilts are kept tightly bounded between 2% and 4% over- or underweight, ensuring that alpha is driven primarily by bottom-up stock picking. Rather than using mechanical stop-losses that create tax friction and portfolio churn, Skelly's team applies a quantimental exit discipline. Winning holdings are trimmed when valuation multiples stretch tracking error tolerances—such as Palo Alto Networks at 70x earnings—while losing positions are reassessed for fundamental thesis invalidation. Differentiating Cyclical Downfalls from Secular Volatility (76:30) Skelly emphasizes that risk management requires distinct playbooks for cyclical versus secular holdings. Cyclical downturns in sectors like energy or banking typically reflect severe structural impairments to earnings power, warranting rapid position exits rather than waiting for hypothetical V-shaped recoveries. By contrast, genuine secular compounders warrant patience during regulatory and headline panics, which historically offer tactical buying opportunities—as seen when Visa and Mastercard dropped 30% after the 2010 Durbin Amendment before multiplying tenfold. Skelly notes that the primary exit catalysts for secular winners are not government regulatory probes, but internal deterioration: market share loss, brand decay, capital misallocation, and misaligned executive compensation. Looking ahead toward 2027, Skelly projects that market leadership will reconcentrate in large-cap tech enablers and healthcare innovators as tougher operating comparisons challenge the broader market. An open uncertainty remains whether massive corporate tech debt issuances ($320 billion by 2026) and data center power grid bottlenecks will force an eventual cyclical digestion period in hardware spending.

摩根士丹利财富管理股票模型投资组合主管Dan Skelly指出，在强劲定价权与人工智能驱动的生产力支撑下，美国企业基本面依然稳健；投资者应通过聚焦核心复合增长与长期主题的精选策略，应对市场向高质量与超大盘龙头的轮动。 宏观韧性与企业AI生产力杠杆 (7:10) Skelly认为，美国经济正展现出抵御政策冲击与通胀压力的韧性，其底层逻辑已逐步摆脱传统周期品和标准收入分配模式。标普500指数第二季度总盈利同比增长28%，中位数公司盈利增幅也达到了14%；在名义GDP增速为6%的背景下，企业整体营收增长达到15%。这一表现得益于两大推手：一是企业在疫后与供应链瓶颈期大幅提价，并在成本回落后依然维持了高价与高利润率；二是企业在未显著增员或裁员的情况下，通过AI工具实现了“合成营业杠杆提升”，大幅拉高了人均产出。 AI资本开支周期与结构分化 (23:41) AI基础设施投入构成了第二季度标普500指数60%的盈利增幅，但硬件供应链面临部署端瓶颈。尽管大模型训练对算力的需求持续遵循扩展定律，摩根士丹利分析师Brian Nowak更预测年AI资本开支到2027年将达1.5万亿美元，但未来两到三年全球拟建的40至50吉瓦数据中心受制于电网连接和变压器短缺，可能引发类似光纤泡沫期的重复下单风险。与2000年网络泡沫由无盈利初创企业买单不同，当前AI基础设施主要由坐拥庞大净现金的超大规模云服务商采购，且企业云迁移仅进行到中途，未使用的GPU亦可灵活回流至核心云业务。 质量重估与小盘股的结构性劣势 (20:35) 当前美股呈现健康的板块轮动，资金从单一半导体硬件逐步流向医疗健康、软件与大型科技股，市场内个股三个月相关性触及近40年低点，毛利率下滑超1%的企业受到严厉惩罚。在此环境下，小盘股面临严峻的结构性挑战：全美营收超过1亿美元的企业约80%滞留在私募市场，导致罗素2000成分股中约40%缺乏净利润。Skelly警告，公募小盘股正遭遇双向风险——左尾面临缺乏护城河而被AI颠覆的威胁，右尾则因资金匮乏无法承担昂贵的AI部署成本以提升效率。 50只精选组合架构与买卖纪律 (64:57) Skelly团队管理着源自Byron Wien理念的40至50只股票长期集中组合。该组合自下而上构建，将行业偏离度控制在正负2%至4%之间：其中40至45只为约2%权重的“核心持仓”，用以锚定基准并控制跟踪误差；另设5至6只跨周期的“大构想”高信念主题标的，以博取超额收益。在买卖执行上，团队放弃机械止损，转而依据基本面假设是否受损以及机会成本进行动态平衡，例如在Palo Alto Networks估值升至70倍市盈率时获利减仓。此外，标普全球与纳斯达克等兼具刚性评级、交易所主业与技术生态的金融数据基础设施，被视为市场在软件恐慌中错杀的高确定性标的。 周期股止损与长期成长股的持有逻辑 (76:30) 结合20年风险管理经验，Skelly强调必须将周期股与长期世俗成长股严格区分。周期性企业一旦盈利滑坡，往往面临剧烈的基本面崩塌，必须果断快速止损；而真正的世俗成长股多能穿越短暂的宏观或监管冲击，正如Visa和Mastercard在2010年遭遇Durbin修正案下挫30%后仍走出十倍涨幅。对于投资组合而言，真正的致命威胁在于品牌护城河丧失、市场份额被蚕食或管理层资本配置失当，而非短期的监管调查。展望2027年，随着常规企业的营业杠杆面临更严峻的同比基数，市场领导权预计将再度收缩至超大盘科技巨头与医疗创新龙头。 该框架主要依赖于美股强劲的微观盈利与AI生产力红利，但未来仍面临电力基础设施交付滞后、企业巨额发债融资成本推升收益率，以及AI实际投资回报率不及预期的潜在考验。

## Background 该视频由 YouTube 频道 The Compound 于 2026 年 9 月 4 日发布。节目为《The Compound and Friends》第 258 期，主持人 Downtown Josh Brown 与 Michael Batnick 邀请到摩根士丹利财富管理（Morgan Stanley Wealth Management）投资组合经理丹·斯凯利（Dan Skelly），共同探讨极具韧性的美国经济、创纪录的企业盈利、AI 资本开支热潮、英伟达与博通的前景，以及大摩如何在高效市场中精选股票并布局至 2027 年。 ## TL;DR 节目围绕当前美股的基本面韧性与板块风格轮动展开深度剖析。斯凯利结合大摩内部研究指出，宏观经济在 AI 资本开支超级周期与生产力红利支撑下打破了传统周期律，标普成分股中位数企业利润激增 14%，而 6 月以来的调整实为健康的浮筹清洗与风格再平衡，不同于 2000 年电信泡沫时期的‘暗光纤’泥潭。针对选股策略，大摩推崇‘核心底仓搭档大构想’的高集中度打法，并强调动量因子与排他性数据壁垒（如 SPGI、NDAQ）的重要性，同时预言小盘股将面临两极挤压，而‘科技七巨头’（Mag 7）有望在 2027 年重领风骚。 ## Atmosphere & style 全片节奏明快且极富华尔街老兵的实战张力，在曼哈顿演播室的轻松调侃与极其硬核的机构数据剖析之间无缝切换。Josh 与 Michael 妙语连珠、频频抛出尖锐问题，而斯凯利则以大摩内部顶尖分析师视角和冷静从容的推演逐一化解，既有对财经媒体恐慌噪音的幽默解构，又有穿透历史周期的宏大叙事感。 ## Bulletin - [10:30](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=630s) 标普 500 企业盈利展现惊人韧性，成分股中位数利润增速达 14%，主要得益于企业将此前借通胀名义调高的价格沉淀为持久的合成经营杠杆，且尚未真正拉下裁员换 AI 的终极降本杠杆。 - [24:45](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=1485s) 华尔街对 2027 年 AI 资本开支的预测达 1.5 万亿美元（远高于一致预期的 1.2 万亿），巨头手握充足净现金，且闲置算力可无缝回流成熟云计算生态，与 2000 年电信泡沫的“暗光纤”存在本质差异。 - [52:54](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=3174s) 当前美股呈现出极端的去相关性与严厉的惩罚机制，过去 12 个月毛利率同比收缩超 1% 的企业均遭遇惨烈抛售，而缺乏 AI 降本对冲与定价权的传统餐饮消费板块承压尤为明显。 - [01:25:01](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=5101s) 摩根士丹利预测 2027 年美股风格或将重演 2023 年格局：大盘与高质量资产再次主导，科技七巨头（Mag 7）与率先拥抱 AI 的医疗巨头将重掌牛市领跑权，而小盘股表现或相对承压。 ## Key takeaways - [01:03](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=63s) 摩根士丹利通过次贷危机收购美邦、并购 E*TRADE 以及打造 Morgan Stanley at Work 获客平台，成功将财富管理营收占比从 2005 年的 8% 提升至当下的约 60%。 - [09:30](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=570s) 市场预期已从上半年的极度 FOMO 情绪回归至涨跌概率各半的 50/50 均衡状态，动量龙头股的阶段性挤泡沫让市场卸下了投机杠杆。 - [13:45](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=825s) 宏观就业处于“不裁员、不招聘”状态，企业在零员工增长下实现 15% 营收增长，未来通过 AI 重构人员结构有望推动利润率进一步释放。 - [15:30](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=930s) 少子化与制造业回流正在引发劳动力缺口，工业机器人与自动化并非抢占就业，而是维系社会与工厂正常运转的刚性生存工具。 - [20:24](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=1224s) 小盘股面临严峻结构性困境：利率上行推高再融资成本、私募股权吸纳了全美 80% 营收过亿的优质非上市公司导致罗素 2000 严重负向淘汰，同时还承受着 AI 颠覆的左尾风险。 - [27:40](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=1660s) 数据中心硬件折旧周期仅为 3 至 5 年，其建成直接锁定了源源不断的服务器与 GPU 替换更新需求，不同于传统工业厂房的一次性投入。 - [35:19](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=2119s) 2000 年电信泡沫的买方是无造血能力的投机初创企业，而当今 AI 资本支出的买方是手握巨额现金流与数千万跨行业企业客户的科技巨头。 - [49:23](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=2963s) 做空纳斯达克期货与中位数股票仓位激增，本质是领涨龙头见顶后的组合机械性对冲，以及对高杠杆和受负面事件冲击个股的定点做空，而非系统性宏观转向。 - [56:04](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=3364s) 美国大众消费与底层信贷依然稳健，零食巨头承压主因是 GLP-1 减肥药改变了饮食结构，各大银行信用卡及先买后付平台的不良违约率依然处于历史常态低位。 - [01:05:20](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=3920s) 摩根士丹利 50 支股票组合采取“40-45 支核心底仓（各占约 2%）控制跟踪误差，搭配 5-6 支十年级超级大构想博取超额 Alpha”的双层架构。 - [01:05:20](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=3920s) AI 赋能将一期药物研发周期从 6 至 7 年大幅压缩至 12 个月以内，生命科学耗材、研发工具与 CRO 机构已迎来重估拐点。 - [01:11:21](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=4281s) 标普全球与纳斯达克等数据密集型与强监管护城河资产被市场严重误判，其权威信用评级法定义务和专有交易数据无法被通用大语言模型替代。 - [01:17:18](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=4638s) 止损策略需区别对待：周期股盈利坍塌往往具有致命且难反转的负动量，必须果断斩仓；而长期结构性赢家遭遇监管诉讼或短期情绪杀跌时，往往构成买入良机。 - [01:22:31](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=4951s) 在信息完全透明、量化基金与充裕散户流动性交织的现代微观结构下，动量已成为决定个股短期生死与超额收益的最关键因子。 ## Detailed walkthrough ### [00:00](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=0s) 摩根士丹利的二十一年与办公大楼迁徙记 在播客录制前的闲聊中，Downtown Josh Brown 询问嘉宾丹·斯凯利（Dan Skelly）在摩根士丹利（Morgan Stanley）工作了多久。丹笑称自己大学一毕业就加入了大摩，至今已有 21 年，“感觉就像刚过了 20 年一样”。当被问及在哪个办公室办公时，丹透露他目前在第七大道 750 号（750 7th Ave），但很快就要搬回著名的百老汇 1585 号（1585 Broadway）大摩总部大楼。由于百老汇 1585 号历经了长达五年的翻新施工，许多团队不得不暂时“离线”外迁办公。丹调侃道，这已经是他 21 年职业生涯中第三次搬回 1585 号大楼了，而他的人生信条恰恰是“人人都爱三部曲（everyone loves a trilogy）”。 Josh 感叹大摩的体量极其庞大，每次他在纽约遇到大摩的理财顾问，顺嘴问起是否认识同在纽约的其他大摩顾问时，对方几乎从不认识。丹深有同感，坦言这正是大摩的惊人规模所在——机构分支遍布各地、人员众多，彼此之间很难全打过交道。 ```mermaid flowchart TD accTitle: 摩根士丹利核心战略转型历程 accDescr: 展示摩根士丹利从以投行为主转向财富管理为核心支柱的发展节点与收购历程 subgraph Phase1["传统业务主导期（2005年前后）"] A["高波动投行业务为主"] B["财富管理营收仅占全公司 8%"] end subgraph Phase2["战略转型与关键并购"] C["约翰·马克（John Mack）引入麦肯锡背景的詹姆斯·戈尔曼"] D["金融危机期间分阶段收购美邦（Smith Barney）"] E["收购线上券商 E*TRADE"] F["搭建股权激励平台 Morgan Stanley at Work 获取高净值客源"] end subgraph Phase3["今日万亿资产巨擘"] G["资产规模接近 20 万亿美元（含 MSIM 与财富管理）"] H["财富管理营收占比跃升至约 60%"] end Phase1 --> Phase2 Phase2 --> Phase3 ``` --- ### [01:03](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=63s) 詹姆斯·戈尔曼的财富管理大棋局 Josh 提到了前任首席执行官詹姆斯·戈尔曼（James Gorman）的远见卓识，称其彻底重塑了摩根士丹利。如今，摩根士丹利财富管理部门与投资管理部门（MSIM）合计管理的客户资产规模已接近 20 万亿美元（实际约为 16 至 17 万亿美元）。Josh 盘点了戈尔曼的一系列神来之笔：在金融危机最惨烈的时刻以极具眼光的方式逐步吃下花旗旗下的美邦（Smith Barney）——“先合资拿三分之一，再拿三分之一，最后全盘接管”；随后又果断收购互联网券商 E*TRADE，并将企业股权管理平台 Morgan Stanley at Work 打造为庞大的获客引擎。 丹回忆起自己 2005 年刚入行时的场景：当时财富管理业务仅占全公司总营收的 8%，而在经历了一系列结构性转型和并购整合后，如今财富管理板块的营收占比已高达约 60%。丹指出，前任 CEO 约翰·马克（John Mack）最早意识到了财富管理咨询业务的巨大价值，并从美林证券（Merrill Lynch）挖来了成功革新美林财富业务的戈尔曼。戈尔曼拥有麦肯锡咨询顾问的背景，将战略咨询思维、卓越的商业时机、精准的定价策略以及随之而来的估值倍数重估完美结合在一起。 丹还分享了一段有趣的个人履历：他在大二时曾在美邦的一家顾问办公室实习，大三进入大摩投资管理部（MSIM）的正规分析师培训项目，2005 年正式全职入职大摩财富管理部门。因此在美邦并购案落地后，无论他走进哪家遗留分支办公室，他既能以“美邦老兵”自居，也能以“大摩嫡系”亮相，完美游走在各类团队之间；而在公司内部，甚至还有经历过 1997 年摩根士丹利与添惠（Dean Witter）合并的更老一辈员工。 - **财富管理营收占比变迁**：2005 年仅占大摩总营收的 8%，如今备考合并各并购资产后已达约 60%。 - **总管理资产规模**：在财富管理与 MSIM 协同下，管理资产体量已逼近 20 万亿美元关口（约 16-17 万亿美元区间）。 - **关键资本运作**：次贷危机时期逐步吞并美邦（Smith Barney）、收购数字券商 E*TRADE、打造 Morgan Stanley at Work 获客平台。 --- ### [03:39](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=219s) 赞助商插播与节目正式开场 现场设备检查就绪，大家纷纷戴上耳机并将手机静音开启“勿扰模式”。随后播出领航先锋（Vanguard）针对固定收益与债券市场的广告赞助信息：面对复杂深奥的债市，先锋基金强调其主动型固收产品基于跨市场周期的稳定表现、低成本复利优势以及审慎的风险管理三大支柱，管理着旗下 40 只基金，以机构级标准为理财顾问和客户创造长期回报。 广告结束后，Josh Brown 热情开场，宣布本期《The Compound and Friends》在纽约市布莱恩特公园（Bryant Park）演播室现场录制。Josh 正式向听众介绍了首次做客节目的重磅嘉宾丹·斯凯利。丹目前担任摩根士丹利财富管理（Morgan Stanley Wealth Management）股票 MAPS（管理型咨询组合解决方案 / SMA 组合）投资组合经理，负责监管股票模型组合团队以及包括《AlphaCurrents》和《US Policy Pulse》在内的主题研究产品；同时，他还是大摩旗下包含 8 只多头 SMA（单独管理账户）策略的美国模型与股息股票策略首席组合经理。丹从 2005 年担任研究部门分析员起便扎根大摩，也是 CNBC 知名节目《Squawk Box》与《Closing Bell: Overtime》的常驻特邀嘉宾。 --- ### [06:10](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=370s) CNBC 连线背后的故事与宏观舆论脱节 Josh 夸赞丹在 CNBC 的出镜表现游刃有余，并询问他是否享受面对公众镜头表达观点。丹坦言自己非常喜欢，认为在当今媒体与技术高度交织的时代，媒体表达已经成为投资从业人员的基本功（"table stakes"）。他回忆起自己在 2015 年夏天的 CNBC 首秀，当时是早晨 6 点的《Squawk Box》档期，与贝基·奎克（Becky Quick）、乔·克南（Joe Kernen）以及安德鲁·罗斯·索金（Andrew Ross Sorkin）同台，刚开始确实有些紧张局促（"touch-and-go"），但随着时间推移逐渐找到了节奏。Josh 随即打趣，自己当年做过早晨 5 点的档期，曾问主持人这大清早电视机前究竟有谁在看，主持人告诉他“基本只有新加坡和亚洲交易员在看”。 切入正题后，Josh 指出当前投资界普遍面临着宏观负面舆论与股市持续走强的激烈拉锯。无论是机构专业人士还是普通散户，似乎都发现每天的新闻充斥着各种悲观叙事；而只要投资者被这些负面消息吓退离场，往往就会眼睁睁踏空标普 500 指数接踵而至的 20% 大涨行情。针对这一现象，Josh 请教丹如何帮助客户调和这种“坏消息满天飞，股价却照涨不误”的认知撕裂。 丹对此回应称，除了优质财经播客之外，他们经常建议普通散户“直接把电视新闻关掉”，不必时刻紧盯外界充斥的噪音。丹补充道，除了日常挑选股票的本职工作，他还是摩根士丹利全球投资委员会（Global Investment Committee, GIC）的成员，该委员会的核心职能是资产配置，专注于长期复利与分散投资。丹特别分析了高净值人群的投资心理误区：过去十年乃至数月前，整个市场的主导逻辑是高度集中（科技巨头），而近期市场广度正在明显扩大。许多通过创业或硬核科技创新积累财富的客户习惯了以“高度集中的方式承担风险”来致富，但这种高度集中的下注方式并不是长期守富的最佳路径；因此，引导客户重回分散配置、紧扣自身目标与风险承受能力，始终是大摩团队资产管理的核心着力点。 ### [09:30](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=570s) 市场情绪转向：从极度 FOMO 到五五开的平衡态势 主持人向来自摩根士丹利的丹·斯凯利（Dan Skelly）抛出一个核心疑问：眼下投资者的焦虑究竟落在何处——是担忧标普突破 8000 点、道指站上 60000 点从而错失牛市，还是恐惧随之而来的回调？斯凯利指出，这个提问的时机非常微妙。如果是在当年的五六月份，市场的普遍情绪毫无疑问是纯粹的踏空恐惧（FOMO）。然而，随着年中以来市场发生的风格轮动，以及上半年领涨动量股的剧烈调整，过度偏激的情绪得到了明显重塑。 目前的投资者预期已经回归到一个非常平衡的区间。现场嘉宾一致赞同，如果在当前节点发起一项民意调查，询问市场接下来的走势是上涨 10% 还是下跌 10%，受访者的预期比例大概率会呈现 50/50 的五五开格局；而在数月前的五月份，这种预期调查很可能会呈现一边倒的 70/30 贪婪态势。斯凯利认为，这种预期的重构让市场卸下了不少浮躁的杠杆与心理负担。 - **情绪钟摆的回归**：从五六月份极度的 FOMO（错失恐惧症）转变为当前多空双方力量相对持平的 50/50 均衡状态。 - **动量股挤泡沫**：上半年表现强劲的领涨龙头股经历了一轮风格轮动与阶段性调整，平复了市场单边做多的狂热氛围。 --- ### [10:30](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=630s) 韧性经济与财报奇迹：标普中位数企业 14% 增长的秘密 当讨论转向美国经济与股市的基本面时，乔希·布朗（Josh Brown）提出了一个核心问题：当前表现出强大韧性的，究竟是宏观经济本身，还是标普 500 成分公司的盈利能力，抑或二者皆有？斯凯利给出的判断是：宏观经济扮演了更坚实的基本盘角色。与四年前相比，当前的美国经济经历了一次根本性的范式转移——它逐渐摆脱了高度依赖传统顺周期行业的旧格局，叠加高收入群体的消费支撑以及 AI 资本开支超级周期（AI spending supercycle），形成了一个非常非典型的经济周期，能够轻松消化利率政策与通胀带来的冲击。 在宏观经济的托底之下，微观企业的业绩表现堪称天文学级别的亮眼。在第二季度中，标普 500 指数层面的每股收益实现了高达 28% 的同比增长；更值得关注且鲜被大众充分讨论的细节在于，成分股中位数企业的利润增速也达到了惊人的 14%。斯凯利与华尔街知名策略师亚当·帕克（Adam Parker）探讨后提出了一项假说： ```mermaid flowchart TD accTitle: 企业虚构关税溢价与经营杠杆释放机制 accDescr: 展示企业如何在关税冲击期提价并将其永久转化为营业利润率的流动路径 subgraph Phase1["成本冲击阶段（12-18个月前）"] A["关税与通胀预期升温"] --> B["企业普遍启动提价（传导60-70%成本）"] end subgraph Phase2["正常化与沉淀阶段（当前）"] B --> C["关税与成本阴霾逐步消散"] C --> D["终端商品与服务价格维持高位（拒绝降价）"] end subgraph Phase3["利润兑现阶段（未来12-24个月）"] D --> E["合成经营杠杆显现：中位数企业利润增14%"] E --> F["AI赋能现有劳动力，推动利润率进一步抬升"] end ``` 这种中位数企业（即排除大型科技巨头之外的普通 493 家企业）的利润爆发，其背后的根源在于一年前企业借由“关税与成本通胀”的由头大幅上调了终端售价，转嫁了 60% 到 70% 的成本压力。然而在成本危机平息后，企业并未调低价格，从而形成了持久的合成经营杠杆（synthetic operating leverage）。为了生动说明这一点，斯凯利拿自己刚搬到长岛郊区的经历打趣：在疫情后通胀期间，他的割草工人以成本上涨为由将维护报价调高了 15% 到 20%，但时至今日，工人的收费从来没有下调过一分钱——“加价容易降价难”正是绝大多数企业保住超额利润的核心写照。 --- ### [13:45](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=825s) “不招也不裁”的生产力红利与潜在的劳动力杠杆 除了未曾退让的定价权，支撑利润率扩张的另一条暗线是企业在生产力上的技术红利。斯凯利指出，目前宏观就业市场正处于一种“不裁员、不招聘”（no hiring, no firing）的温吞停滞状态。年初市场上广泛流传的“AI 鬼故事”——例如企业级 SaaS 软件彻底消亡、大规模白领失业潮等——已被市场现实逐一证伪。相反，财富 500 强企业的真实画像是在维持现有雇员规模不变的前提下，通过 AI 工具赋予现有员工更高的产出能力。 乔希·布朗与迈克尔·巴特尼克（Michael Batnick）对此进行了深入推演：在名义 GDP 增速约为 6% 的宏观背景下，第二季度企业的营业收入同比增长了惊人的 15%。在过去的周期性重资产经济中，想要实现 15% 的营收增长，工厂必须在流水线上增加 10% 的一线装配工来拧螺丝、装箱子；而在当前的数字化与 AI 时代，企业在员工零增长（Headcount 零增加）的情况下直接创造了巨额增量收入。更关键的是，斯凯利认为企业削减低效岗位或全面引入 AI 替代劳动力这一“终极降本杠杆”（labor lever）甚至还没被真正拉下。在未来的 12 到 24 个月里，随着企业在技术成熟后真正重构人员结构，利润率还将迎来更深一步的释放。 - **脱虚向实的效率飞轮**：名义 GDP 增速仅约 6%，但企业营收增长达 15%，完全由员工人均产出提升驱动，打破了过去营收增长必然伴随人力同步膨胀的周期铁律。 - **劳动力降本未完待续**：目前的利润率改善仅体现了现有员工搭载 AI 带来的效率红利；未来 1-2 年随着人员替代杠杆的实质拉动，企业经营利润率仍具备上冲空间。 --- ### [15:30](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=930s) 人口塌陷、再工业化与“弹窗广告”式的宏观噪音 在惊叹于企业生产力跃升的同时，三位对话者将视野拓宽至更为深远的全球人口结构与社会趋势。面对少子化与人口老龄化，乔希戏谑地调侃道：“只要大家少生孩子就没事了。”但事实上，全球多地的人口坍塌已成为不争的事实——韩国、中国、欧洲以及美国均在经历生育率下滑，甚至像雪城大学（Syracuse University）等高校因适龄学生生源枯竭而持续无法完成招生指标。斯凯利借用自身经历共情（他的妻子曾是雪城大学 Gary Gait 执教时期的女子长曲棍球校队主力），并用一句经典的黑色幽默总结了科技与人口交织出的未来幻象：“在 GLP-1 减肥药延年益寿、房地产市场因利率锁定而流动性冻结、再加上 AI 和机器人的全盘渗透下，我们的未来就是每个人都能活得极度长寿，但整天无所事事，而且买不起也住不进新房子。” 面对劳动力老龄化与人口缩减，自动化与机器人技术并非抢夺饭碗的恶魔，而是维系社会运转的刚需工具。无论是在即将爆发严重护士荒的医疗护理体系，还是在摩根士丹利工业研究团队重点跟踪的“制造业回流”（Reshoring）大潮中，新建立的大批本土先进制造工厂根本无法招募到足够的 18 至 35 岁年轻产业工人，其核心产能必须交由工业机器人和自动化产线承载。 | 宏观维度 | 传统认知 / 早期恐慌 | 现实情况与中长期演变逻辑 | |:--- |:--- |:--- | | **就业与招工** | 经济扩张要求企业同比招募大批工人 | “不招不裁”；AI 与自动化驱动人均产出大幅提升 | | **人口与社会结构** | 人工智能抢走基层与入门级就业机会 | 少子化导致护士、蓝领与工厂用工荒，自动化成为生存刚需 | | **制造业回流** | 重启传统工业基地，吸纳数十万年轻劳动力 | 新工厂主要依靠工业自动化与协作机器人填补结构性缺口 | | **宏观与地缘冲击** | 通胀与地缘冲突（如中东局势）将彻底终结牛市 | 政策与地缘冲击退化为“网页弹窗广告”，主线重回盈利与 AI | 最后，谈及眼前的市场走向与即将到来的大选及劳动节假期，斯凯利表现出了极具说服力的从容态度。他提出，如果站在更宏观的认知高度来看，过去动辄引发市场剧烈波动的政策冲击、通胀再起预期或是中东局势等宏观地缘危机，在今天的市场环境中更像是一个个转瞬即逝的“网页弹窗广告”（pop-up ads）——它们弹出来时确实令人分神，但稍纵即逝，资本市场的主流叙事在被短暂干扰后，永远会毫不犹豫地迅速切回核心页面：强劲的企业盈利韧性，以及 AI 驱动的产业超级周期。 ### [18:34](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=1114s) 充沛流动性支撑的牛市与健康的板块轮动 Dan Skelly 指出，即便从理论上无法断言即将到来的中期选举毫无影响，但正如 Josh Brown 所言，在极其坚挺的企业盈利基本面支撑下，市场可能出现的任何回撤或修正都会非常温和。Josh 随即提出一个常被忽视的宏观现实：当前金融体系中沉淀着海量的资金。在过去数十年里，类似的宏观冲击至少会让经济脱轨甚至直接引发衰退，而如今由于充裕的流动性，哪怕二级市场上出现任何风吹草动或负面情绪，资产几乎立刻就会被蜂拥而至的买盘吞噬——例如 Google 在重磅的 SpaceX 交易前进行的发债，转眼就被市场抢购一空。这种庞大的流动性缓冲机制正在以一种未被充分定价的方式重塑经济和市场的韧性。 Dan 对此深表赞同，认为这是长达 15 年大牛市积淀下来的红利——市场叙事从早期的 FAANG 演进到“美股七巨头”（Magnificent Seven），再推进到 AI 资本开支（AI Capex）。从延长本轮经济周期的角度来看，近期最健康的信号莫过于板块轮动。6 月份半导体板块及其他第一梯队 AI 资本开支受益股的回调，恰恰为医疗保健（Healthcare）板块的爆发、金融股的全年稳步走强以及小盘股的突然爆发腾出了资金空间。工业板块此前也经历了一波凌厉的上涨。Dan 还特别向摩根士丹利的同事、前主管 Mike Wilson 致敬，赞赏他早在去年底至今年初就精准做出了看多小盘股的判断。 - **流动性吸收能力**：市场中的海量资本形成了天然看跌期权，任何资产折价或优质发债（如 Google 面向 SpaceX 交易的债券发行）都会立即被巨量流动性抹平。 - **牛市演进路径**：FAANG $\rightarrow$ Magnificent 7 $\rightarrow$ AI 基础硬件投资（AI Capex） $\rightarrow$ 医疗、金融、工业等低估值板块的全面轮动。 --- ### [20:24](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=1224s) 小盘股的结构性困境：利率、私募分流与 AI 两极挤压 针对 Josh 提出的“标普 500 指数的强劲盈利增长是否能顺利外溢并持续推动中盘股和小盘股”这一问题，Dan 坦言小盘股领域的复杂度和危险系数要高得多。一方面，工业和金融等在中小型指数中占比较高的行业确实享有诸多独特的顺风，例如资本市场复苏周期、并购（M&A）活动激活带来的滚雪球效应（近期的生物科技板块便是一个典型代表）以及市场流动性催生的风险偏好（FOMO）。但另一方面，这些积极因素难以掩盖小盘股面临的严重逆风。 首先是利率上行的沉重压力，这对资产质量较低的小盘股群体打击尤为直接。其次是过去 15 年私募股权市场（Private Equity）爆发式增长带来的“负向逆向淘汰”：全美年营收超过 1 亿美元的公司中，约 80% 均未上市，这意味着散户在罗素 2000（Russell 2000）中买到的小盘股实际上是被私募市场筛选后挑剩下的标的，导致该指数中约 40% 的成分股目前处于亏损状态。更致命的威胁来自 AI：小盘股不仅面临着被 AI 彻底颠覆和中介化消解的“左尾风险”（Left-tail risk），而且由于缺乏盈利能力与充足资本去采购算力和部署 AI，它们很可能根本无法享受到 AI 采纳浪潮带来的生产力跃升红利（即“右尾收益”，Right-tail opportunity）。 ```mermaid flowchart TD accTitle: 小盘股的结构性困境分析 accDescr: 分析罗素2000指数面临的负向选择偏差、利率压力以及AI技术带来的两极不对称挤压。 subgraph MacroPressure["宏观与市场环境挤压"] Rates["利率居高不下<br/>劣质债务再融资成本激增"] PE_Drain["私募市场虹吸优质资产<br/>全美营收超1亿公司中80%为非上市"] Zombie["罗素2000指数负向淘汰<br/>约40%成分股尚未盈利"] end subgraph AIExtremes["AI 浪潮的不对称风险"] LeftTail["左尾风险：行业颠覆<br/>轻资产/中介型小企业被AI全面替代"] RightTail["右尾缺失：资本壁垒<br/>缺乏自由现金流投入昂贵的AI算力与架构"] end Rates --> Zombie PE_Drain --> Zombie Zombie --> LeftTail Zombie --> RightTail ``` - **形象比喻**：Josh 形容小盘股的走势就像“周五早晨飞往坦帕的航班”——绝大多数时候沉闷平稳，冷不丁来上几秒剧烈颠簸，往往在给投资者一段短暂的希望后紧接着长期的失望，直到大家心灰意冷时又突然拉升，其收益特征与大盘蓝筹截然不同。 --- ### [23:05](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=1385s) 政治噪音淡出与数据中心建设的“提前拉动” 在讨论政治对市场的影响时，主持人和嘉宾一致认同应当将华盛顿的噪音视为网页上的“弹窗广告”：余光能看到，但每个人都只想尽快点掉窗口。中期选举可能带来的国会分治往往不会改变经济运转的本质轨道，反而可能迅速给某些人造争议降温——例如眼下各地声势浩大、带有明显政治操弄色彩的数据中心环保抗议，很可能会随着选战结束而烟消云散，因为政客们失去了借由攻击微软等科技巨头来博取政治选票的动机。 然而，更深层次的产业隐忧在于选举周期引发的“需求提前透支”（Pull-forward）。Dan 提到他此前与摩根士丹利顶尖互联网行业分析师 Brian Nowak 交流时的洞察：在经历 6 月份的 AI 动量杀跌后，许多内存和算力硬件标的仍在 50 日均线附近反复挣扎；而在中期选举后，市场焦点会迅速转向 2028 年大选，届时科技巨头可能会为了规避未来的政策变数，选择将海量的数据中心基础设施建设进一步前置压缩至未来几年完成。这种前置抢跑虽在短期内制造繁荣，却也在悄然累积繁荣过后的反噬隐患。 --- ### [24:45](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=1485s) 业绩超预期却遭抛售？Y2K 宿醉担忧与 AI Capex 的真正周期 资本开支提前拉动所带来的最大阴影在于“宿醉期”（Hangover）。Josh 指出，历史上最典型的提前透支案例是千年虫危机（Y2K），当时企业狂热抢购 IT 设备，而宿醉在 2000 年第二季度全面显现，市场猛然惊醒：企业不可能每个季度都采购如此巨量的英特尔芯片。当前标普 500 指数中，来自超大规模云服务商（Hyperscalers）、前 50 大半导体/存储巨头以及戴尔（Dell）等服务器硬件厂商的利润占了大半江山；当前标普估值处于 21 倍远期市盈率的高位，一旦算力硬件采购的齿轮出现倒转，市场将遭受沉重打击。近期 Broadcom 公布了惊人的约 90% 增长，股价却在盘后重挫 5%（“因为没有达到 91%”）；Nvidia 即使交出大超预期的财报并指引供应端依然受限，股价也在冲高后动能不足。投资者一方面担忧盈利预期已经高得不切实际，另一方面哪怕面对前所未有的炸裂业绩也不再盲目追涨，这让不少人怀疑牛市是否已经见顶。 Dan 对此持坚定的长期乐观态度。他强调，资本主义机制下的长期结构性大牛市（Secular Bull）通常能持续 20 至 25 年，而结构性熊市的周期往往只有牛市的一半。以 2009-2011 年以黄金或通胀平减后标普筑底为起点，本轮牛市目前仅走过了约 15 年，整体长周期并未终结： - **业绩贡献的硬度**：2024 年第二季度 28% 的标普整体盈利增幅中，有整整 60% 直接归功于 AI 基础设施板块，Dan 认为这部分收益在 2027 至 2028 年依然非常安全（"Money good"）。 - **资本开支远超华尔街预期**：Brian Nowak 对 2027 年 AI 资本开支的预测高达 1.5 万亿美元，而目前华尔街一致预期仅为 1.2 万亿美元左右。摩根士丹利团队戏称“上调 AI 投资指引”已成为过去 18 个月每个季度的保留传统，各大巨头为了维持代际竞争壁垒以及中美大国科技竞逐，根本无法承受缩减投资的风险。 - **底层技术扩展定律（Scaling Laws）有效**：每当使用更庞大的算力集群对新一代前沿大模型进行预训练时，模型的输出质量与推理能力依然在沿着幂律曲线稳步提升，并未遭遇致命瓶颈。短期内的个股波动反而有助于清洗浮筹并拉长整轮科技繁荣周期的寿命。 ### [27:40](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=1660s) 数据中心的持续折旧与芯片更新周期 资本支出（CapEx）究竟为何不会在短期内骤然缩减？对这一问题的探讨，揭示了一个常被市场忽视的技术硬件规律：建立数据中心并非一劳永逸。不同于丰田（Toyota）建一座可以连续二十年生产 RAV4 的实体工厂，数据中心一旦建成，其内部的算力硬件就启动了不可逆转的折旧与迭代周期。无论业界将一枚生产于 2026 年的 GPU 视作 3 年还是 5 年折旧周期的资产，它绝对不是一个能用上 20 年的资产。 数据中心的建设实际上锁定了未来源源不断的服务器与芯片采购需求。即使数据中心本身的新建速度未来可能放缓，所有已建成的现有基础设施依然需要庞大的前沿技术硬件来喂养维护。Nvidia 以及近期公布爆表财报的戴尔（Dell）等硬件厂商，其业务底色都与这种持续性的设备更新紧密相连。 - **丰田工厂 vs. 数据中心**：传统工业厂房具有长达数十年、生产同一款车型的长资产寿命；数据中心内部的 GPU 和服务器更迭频繁，技术淘汰速度快。 - **内生性采购需求**：数据中心的大规模落成，本身就锁定了对最新芯片、服务器持续更新与替换的硬性需求。 --- ### [28:55](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=1735s) 部署瓶颈、暗光纤类比与对冲基金的谨慎轮动 虽然市场普遍热议 AI 芯片的“需求受限”（Demand Constraints），但更现实的隐患在于“部署受限”（Deployment Constraints）。摩根士丹利（Morgan Stanley）在 2024 年 6 月 1 日的一份研报中就曾对半导体势头提出审慎警示：产业链各方正提前为未来两到三年内规划的 40 至 50 吉瓦（GW）数据中心电力与建设规模订购芯片、服务器和电力工业组件。然而，除埃隆·马斯克（Elon Musk）这类具备极强工程执行力的案例外，并非所有数据中心都能按期交付。一旦建设延期，供应链内部便存在“重复下单”（Double Ordering）的风险——虽然当前并非现实危险，但在一年后可能成为实质隐患。 Josh Brown 现场展示了电信泡沫时期 Ciena 的历史走势图：当年光纤网络建设期间股价拔地而起、随后一落千丈的惨烈走势，与如今某些硬件图表惊人相似。 ```mermaid flowchart TD accTitle: 2000年电信泡沫与当前AI周期的承接对比 accDescr: 对比2000年暗光纤因缺乏替代需求而沉寂十年，与当前闲置算力可直接回流成熟云计算生态的差异。 subgraph Y2K ["2000年电信泡沫：光纤过度建设"] A1["超前铺设海量光纤"] --> B1["下游缺乏有效应用生态"] B1 --> C1["'暗光纤'沉寂逾10年<br/>(直至AWS与移动互联崛起)"] C1 --> D1["纳斯达克经历近90%惨烈下跌"] end subgraph Present ["当前AI周期：数据中心与算力"] A2["超前采购GPU与服务器"] --> B2{"是否存在建设延迟<br/>与重复下单风险？"} B2 -->|存在部署瓶颈| C2["算力重定向至成熟云计算"] B2 -->|企业上云仅处中局| D2["全球财富500强深度数字化"] C2 & D2 --> E2["极难重演全盘报废的'暗GPU'泥潭"] end ``` 针对这种担忧，Dan Skelly 依据摩根士丹利作为行业领先主经纪商（Prime Broker）的机构对冲基金持仓数据作出了关键反驳： 1. **机构仓位并未盲目狂热**：在经历回调后，对冲基金客户并未盲目逢低补仓上半年领跑的半导体与 AI CapEx 赢家，仓位远未回到前期高点；资金反而理性轮动至防御性板块（如医疗保健）或部分软件股以捕捉底部。 2. **云计算尚未见顶，提供坚实保底**：当前的 AI 浪潮并非建立在空中楼阁之上。各大超大规模云服务商（Hyperscalers）的企业上云渗透仍处于第四或第五局的中场阶段，远未步入尾声。 3. **“暗GPU”不同于“暗光纤”**：2000 年铺设的海量光纤在互联网破灭后没有替代用途，沦为“暗光纤”（Dark Fiber）足足沉睡了十年，直到 Amazon 等巨头成熟才得以消化，纳斯达克在长达 13 年后（从 2000 年一季度直至 2013 年）才重拾历史新高。而如今的数据中心芯片若遇上短期项目脱节，能随时回流并被日常云计算负载有效利用，演变成全盘荒废的概率大幅降低。 --- ### [32:49](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=1969s) 忧虑之墙与产业接力棒的交接 近期的财报季呈现出一个耐人寻味的现象：Nvidia 和博通（Broadcom）相继交出爆发式的超预期财报，但次日股价却反应平淡甚至回调。从“半杯水是满的”这一多头视角来看，市场没有因为优异业绩就将股价在两周内暴拉 70%，反而是极为健康的信号——它避免了一飞冲天后的急速崩盘，延长了牛市的寿命，让行情以一种“慢动作”的姿态稳健延伸。 市场始终需要构筑“忧虑之墙”（Wall of Worry）。正如 2022 至 2023 年美联储开启了 40 年来最激进的加息周期时，全市场绝大多数经济学家（摩根士丹利的 Ellen Zentner 等少数人除外）都笃信一场惨烈衰退即将降临。然而由于经济重心向服务业发生范式转移、K 型收入分化让高消费能力群体对日常通胀更不敏感，以及 2023 年后 AI 超级周期的启动，预期中的衰退从未发生。 未来两年的核心考验在于“产业接力棒的平稳交接”：**AI 采纳者（AI Adopters）能否通过生产率与利润率的切实飞跃，抵消 AI 基础设施资本支出可能出现的减速？** 这种生产效率的释放必须实打实地落地，成为支撑经济与估值中枢的关键动力。 --- ### [35:19](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=2119s) 客户质地的根本差异：1999 与当下的分水岭 许多经历过 2000 年互联网泡沫和 2008 年金融危机的散户投资者，至今仍受制于“世代偏见”与创伤记忆，时刻怀疑当前重演 1999 年的溃败。然而，当下的买方实力与 1999 年有着本质区别。 | 维度 | 1999 年科技网络泡沫 | 当前 AI 投资周期 | |:--- |:--- |:--- | | **主要资本支出方** | 依靠风投输血、毫无造血能力的新上市点康企业（Dot-coms） | 坐拥净现金资产负债表、年利润丰沛的科技巨头（原 FAANG / 现 Mag 7） | | **下游买方质地** | Pets.com、eToys、CDNow、DJ Direct 等上市数周后便关门大吉的投机公司 | 涵盖医院、保险机构、先进制造业与政府部门在内的 3000 万全球真实云客户 | | **硬件销售依赖度** | 思科（Cisco）、Sun Microsystems、Ciena 依赖寄希望于“下一轮 100 家 IPO”的虚幻需求 | 亚马逊（Amazon）等平台代表实体经济各类垂直行业统一代采底层基础设施 | | **资本运作特征** | 激进烧钱铺设基础设施，缺乏真实商业循环支撑 | 巨头过去十年沉迷万亿美元无风险回购；面对划时代技术，转向进行严肃的大规模 CapEx 布局 | 在 1999 年，像思科、朗讯（Lucent）、戴尔、Sun Microsystems 以及 Juniper 这样的设备商，其销售增长依赖于向刚上市一周甚至一年后就破产清算的初创公司卖设备；而今天，亚马逊采购硬件是为了服务其生态下 3000 万真实跨行业企业。 尽管生态系统里最大的两家买方大模型厂商尚未上市、暂未实现盈利、财务披露也不够透明（正如主持人所指出的现实注脚），但总体而言，本轮资本开支支撑者的资产负债表与商业纵深，与世纪之交的泡沫时代不可同日而语。 ### [37:25](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=2245s) 资本开支热潮与历史宿命：美国资本配置的卓越与周期陷阱 讨论聚焦于未经完整经济周期检验的私有领域成长性。Dan Skelly 指出，虽然出于合规限制不能对未上市标的透露过多，但公开数据表明，部分未上市 AI 企业的 ARR（年度经常性收入）呈现出令人咋舌的爆炸式增长——从去年底的 100 亿美元级别，短短数月内便蹿升至 600 至 700 亿美元。这种强劲的真实需求支撑着当下的供需格局。 Dan 强调了两大关键差异：一是本轮资本支出方（Spenders）的高质地；二是支出主体极高的信用度，这与 2000 年互联网泡沫时期的 Pets.com 有着天壤之别。从长达 30 年的技术图表来看，科技板块相对于标普 500 指数始终在走出更高的高点（higher highs）。无论历经 2000 年互联网破裂、2008 年金融危机、COVID-19 疫情，还是 2022 年因加息引发的久期杀估值，科技行业总能迅速走出泥潭并刷新历史纪录。这背后折射出“美国例外论（American exceptionalism）”依然强劲——美国依然是全球将资本配置到颠覆性创新与前沿科技中最出色的市场。 然而，历史的警钟从未停歇。无论是当年的互联网光纤网络铺设、19 世纪的铁路狂潮，还是 2015 年页岩油投资热，美国几乎每一次在准确押注大趋势的同时，都会在资本支出（Capex）上陷入“繁荣-萧条（boom-bust）”的过度投资怪圈。 ```mermaid flowchart LR accTitle: 资本开支周期的两难抉择 accDescr: 展示美国资本配置在创新引领与过度投资之间的循环特征 subgraph Alloc["资本配置优势"] A["颠覆性创新浪潮\n(AI/算力基础设施)"] --> B["超强资产负债表支持\n(巨头净现金充沛)"] end subgraph Risk["历史宿命风险"] C["资本开支过度扩张\n(光纤/铁路/页岩油先例)"] --> D["繁荣到过剩的回撤\n(Capex Boom-Bust)"] end B -.-> C ``` - **ARR 飙升**：追踪的部分非上市 AI 初创企业 ARR 规模短时间内从 100 亿美元飙升至 600~700 亿美元。 - **大数定律与市场理性**：Nvidia 和 Broadcom 在公布极佳财报后并未出现狂热暴涨，显示出市场定价走向理性，认识到这些巨头不可能无限期保持 100% 的净利增速。 - **轮动健康度**：资金并未离场，而是分流向 Snowflake 等具备增长潜力但市值远未达 5 万亿美元级别的标的。 --- ### [40:10](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=2410s) 机构上限、散户接盘与 3200 亿超大规模发债 市场轮动的另一个技术性因素来自共同基金（40 Act Funds）的持仓上限。当一家公司市值膨胀至极致，许多主动管理型基金在组合比例上存在硬性天花板（如单只股票不得超过 10%），无法继续超配。此时边际买家退居至高净值客户与散户群体。尽管散户过去几年在逢低买入方面表现敏锐，但个人财富在单一标的（如 Nvidia）上的暴露程度终有上限。 最终的归宿依然取决于超大规模云厂商（Hyperscalers）能否在天量资本支出中实现投资回报率（ROI）。Michael Cembalest 此前的研报指出，到 2026 年，超大规模云服务商与 Nvidia 通过新增投资级债券与特殊目的载体（SPV）融资规模将达到惊人的 3200 亿美元，这种庞大的吸金体量甚至在客观上推升了美国国债收益率。 对于如此巨额的发债是否会引发流动性枯竭或违约担忧，Dan 认为当前的认购需求和利差（Spreads）依然非常健康： - **买家群体扩容**：包括大型保险公司在内的充沛流动性正在极力搜寻优质投资级固定收益资产。 - **信用资质严重分化**：市场上只剩 Microsoft 与 Johnson & Johnson（JNJ）保持最高的三 A（AAA）评级，而像 Oracle 这样资产负债表偏薄弱的发行主体，其 CDS（信用违约互换）利差早已大幅走阔。市场正在精准进行风险定价，并非盲目承接。 - **主动权在手**：Alphabet 等巨头发债绝非出于现金匮乏的被动举债，而是在信用资质极优、资金成本低廉的情况下，充分利用非股权融资渠道进行资本开支运作。 --- ### [42:43](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=2563s) “循环融资”罗生门与高收益债利差背离 对于财经媒体每日热议的“循环融资（Circular Financing）”质疑——即设备供应商投资下游初创企业，下游企业再拿这笔钱购买其芯片和设备，酷似当年光通信泡沫末期的“卖方信贷（Vendor Financing）”——Dan 引用了 Jensen Huang（黄仁勋）在 2025 年秋季电话会上的公开表态： > “鉴于我们对供应链和合作伙伴未来增长拥有极其敏锐且独特的全景视野，再加上我们手中充沛的净现金资产负债表，作为我们公司的股东，难道你不希望我代表你们持有这些生态伙伴的股权与筹码吗？” 尽管供应商直接下场注资下游引发了市场对周期见顶的担忧，但 Dan 强调，巨头们不仅在经历互联网、移动社交、电商、数字广告、流媒体和云计算的完整周期中展现出了极强的韧性，而且在 AI、机器人、太空探索、量子计算及自动驾驶等下一代前沿技术中仍然占据绝对主导权。投资者应当警惕，但也应通过监测信贷市场利差来识别真实危险信号——正如 1998 至 1999 年联储加息周期以及 2025 年秋季 Oracle CDS 走阔对股价做出的精准预警一样。 随后讨论转向信贷市场正在显现的一抹诡异背离：整体高收益企业债（High Yield Corporate Bond Spread）利差依然波澜不惊，但属于更垃圾级别的 CCC 级债券利差却出现了明显的大幅抬升。针对这种分化，Dan 从底层行业基本面给出了解释： - **行业权重护城河**：高收益债指数中重仓的能源、原材料、化工和工业板块，当前在很大程度上受益于中东等地区的地缘政治动荡。 - **资本纪律重构**：以能源行业为例，2015 年页岩油崩溃前，高管薪酬机制严重锚定产量扩张，油价飙升必然导致盲目扩产与资本开支失控；2015 年崩盘后，高管激励机制彻底转向资本回报、自由现金流与股东分红，如今的能源龙头已蜕变为具备更稳健资产回报能力的防御性主体，从而拉开了与高杠杆私有股权支撑的边缘企业（CCC 级）的抗风险差距。 ### [46:42](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=2802s) 行业分化与做空热潮：美股空头仓位激增 讨论首先聚焦于能源、材料与工业板块的分化。Dan Skelly 与 Josh Brown 探讨了能源板块近期在基本面与成分构成上的变化：高分红和股份回购成为支撑，同时下游的原材料（如钢铁、铜）和工业零部件制造则深度受益于 AI 数据中心的扩建浪潮。这种分化反映出市场正在跨越“忧虑之墙”（wall of worry），即便短期内存在普遍的不信任感，从长远来看这种怀疑情绪反而有利于长期世俗牛市（secular bull market）的健康发展。 ```mermaid flowchart TD accTitle: 美股空头激增与市场分化逻辑 accDescr: 展示高盛空头数据与标普500多空分化背后的驱动因素及成因分析 subgraph MacroSignals ["做空情绪飙升"] A["高盛报告: 杠杆基金空头激增35%"] B["标普500中位数股票空头比例陡峭上升"] end subgraph GroupSplit ["标普两极分化"] C["空头极低组 (底十分位数 ~1.2%)<br/>Google、伯克希尔、大摩、苹果、礼来"] D["空头集中组 (顶十分位数 ~9.8%)<br/>Reddit、SMCI、Charter、Trade Desk"] end subgraph Catalysts ["驱动逻辑"] E["高品质资产：大盘多头头寸的机械性对冲"] F["低品质资产：降息预期回撤 + 个股黑天鹅"] end MacroSignals --> GroupSplit C --> E D --> F ``` 随后迈克尔·巴特尼克（Michael Batnick）展示了高盛（Goldman Sachs）的杠杆基金与纳斯达克 100 指数报告：对冲基金等杠杆资金正大举做空纳斯达克 100 期货，自 6 月中旬以来空头头寸飙升了 35%，逼近历史最高纪录。同时，“图表小子”马特（Matt）制作的数据显示，标普 500 指数成分股的做空比例中位数（做空股数占市值的比例）正以极为陡峭的趋势向右上角飙升。即使是全市场最不被看空的底部分位数（the bottom decile，空头比例极低的核心资产），其被做空的比例也在实质性抬升。这番数据一度让现场气氛变得严肃，甚至被调侃是不是突然转向看空，Dan 则直言这需要非常认真地审视，绝非在 CNBC 早间连线时能草率回答的浅显问题。 - **高盛数据**：杠杆基金对纳斯达克 100 指数期货的做空仓位自 6 月中旬以来激增 35%，逼近历史极值。 - **做空比例中位数**：标普 500 指数成分股的做空占流通市值比急剧攀升。 - **做空比例极低的巨头（第 99 百分位/底部分位数）**：包括 Alphabet（Google）、伯克希尔·哈撒韦（Berkshire Hathaway，Josh 吐槽“到底有谁会去做空巴菲特？”）、摩根士丹利（Morgan Stanley）、通用动力（General Dynamics）、沃尔玛（Walmart）、亚马逊（Amazon）、富国银行（Wells Fargo）、礼来（Eli Lilly）、苹果（Apple）、雪佛龙（Chevron）等。 --- ### [49:23](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=2963s) 结构性对冲与个股爆雷：空头飙升背后的真相 面对全市场做空比例上升的现象，Dan Skelly 提出了量化与结构层面的解释。回顾此前行情，截至 6 月底半导体触顶之前，市场呈现出极端的超集中化（hyper-concentrated）：标普 500 指数年内多达 90% 的涨幅仅仅由三个子行业贡献——半导体、IT 硬件和电力，全部直接挂钩 AI 数据中心建设。因此，左侧中位数股票做空上升，很大程度上是资金在这些领涨龙头见顶后进行利润轮动，同时使用做空中位数股票作为组合对冲手段（mechanical hedging）。 另一方面，查看空头仓位最重的顶十分位数（top decile，平均做空比例达 9.8%，处于历史第 70 百分位），其背后的驱动因素并非纯粹的宏观熊市逻辑，而是受到利率重定价以及一系列具有强烈个股特殊性（idiosyncratic）的事件所驱动： - **利率重定价的压制**：市场此前定价了多次降息（从年初的数次降息预期退潮至 50/50 的摇摆状态），资产负债表较弱、高杠杆（debt-to-equity ratio 高）的低品质股票因此遭受重估。 - **个股特有风波**： - **The Trade Desk (TTD)**：在强劲的科技板块中表现极度反常，业绩令人失望。 - **超微电脑 (SMCI)**：陷入财报与会计丑闻。 - **Charter Communications**：深陷有线电视传统业务衰退的困境。 - **Reddit**：受到 Google 搜索算法调整带来的流量分流威胁。 - **EchoStar**：直接面临与 SpaceX 星链（Starlink）的残酷正面竞争。 针对底部分位数中的巨头，讨论还提及了微软（Microsoft）的估值与情绪重构。Dan 指出微软在财报电话会议中对资本支出（capex）给出了更为克制温和的指引，获得了市场买账；而 Josh Brown 则补充了更深层的逻辑：市场热捧微软，是因为微软在事实上结束了与 Sam Altman 及 OpenAI 的独家绑定状态，逐步走向“公开分手”。投资者意识到原先的交易对微软并不划算，转向使用成本更低廉的开源/开放权重模型（open-weight models）来交付 AI 软件产品，反而能够有效保护利润率，使微软重新找回了商业理性。 --- ### [52:54](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=3174s) 市场机制运作良好：毛利率缩水成为最大死穴 巴特尼克反驳了“被动投资和指数基金正在毁掉股市定价效率”的常见偏见，直言当前美股市场运转得相当健康与理性——它正在极其精准地奖优罚劣。CNBC 图表显示，美股 3 个月内的股票间相关性（intra-stock correlations）已经跌至近 40 年来的最低点，“每只股票都在按照自己的节奏前进”。 ```mermaid flowchart LR accTitle: 财报季惩罚机制 accDescr: 展示企业毛利率下滑与股价暴跌之间的直接因果关系 A["毛利率同比收缩 > 1%"] --> B["定价权失效 / 缺乏AI提效"] B --> C["相对行业指数被极度抛售"] subgraph Examples ["典型受压行业"] D["餐饮快餐 (QSR): Shake Shack、Chipotle"] E["食品必需品: 承受油价与利率冲击"] end B -.-> Examples ``` 资深策略师亚当·帕克（Adam Parker）的一项研究更是指出了本轮财报季最严厉的惩戒法则：凡是过去 12 个月毛利率（Gross Margin）同比收缩超过 1% 的公司，其相对行业指数的表现均遭到了近乎毁灭性的抛售惩罚。当期的博通（Broadcom）便是一个鲜明例证——即使整体营收表现强劲，但由于毛利率下滑了约 200 个基点，财报发布前股价已经重挫，直接体现了市场对利润率失守的零容忍。 Dan 与 Josh 分析指出，毛利率正在成为筛选企业生存能力的试金石： 1. **定价权与通胀脱节**：许多传统消费与餐饮行业无法再通过持续提价来消化成本，失去了定价增益。 2. **缺乏 AI 降本增效的对冲**：以快餐及连锁餐饮连锁（QSR，如 Shake Shack、Chipotle 等）为例，这些企业正遭受牛肉价格上涨、包装纸制品成本上升等通胀反噬，且完全无法享受到 AI 基础设施带来的红利——“Shake Shack 又不能给数据中心供应炸薯条”。 3. **K 型分化（K-shaped economy）的反转隐忧**：虽然大盘指数权重被科技巨头主导，但全美数千只缺乏科技杠杆的普通中小盘股票与必需消费品（food staples）正在面临高油价、高利率黏性以及底层消费者购买力疲软的双重挤压。 ### [56:04](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=3364s) 零工经济、官方数据滞后与消费者的真实韧性 对宏观经济数据的争论往往集中在官方统计方法是否已经过时。现有的大量数据统计框架可以追溯到 20 世纪 50 年代，根本无法有效捕捉现代的零工经济（Gig Economy）以及人们在传统工资表之外获取收入的多样化方式。尽管普遍担忧中低收入群体正在承受压力，但该群体的消费韧性其实远超市场预期。不可否认，中产群体的边际恶化可能正处于初露端倪的早期阶段，但如果单纯因为某些特定企业的业绩波动就对整体宏观经济妄下断言，往往会陷入误区。 例如食品与必需消费品行业，百事公司（PepsiCo）及其旗下的菲多利（Frito-Lay）等零食业务已连续两至三年未能达成有机营收增长目标。五年前管理层曾信心满满地认为能够轻松应对行业冲击，如今却举步维艰；好时（Hershey）等公司的股价表现同样疲软惨淡。然而，这背后很大程度上是 GLP-1 减肥药带来的真实消费行为改变与品类冲击，而非整体经济崩盘的征兆。将麦当劳的单季度财报或零食巨头的困境直接外推为美国整体宏观经济不行，在分析逻辑上是完全错误的。 ```mermaid flowchart TD accTitle: Consumer Data Misinterpretation vs Reality accDescr: Diagram contrasting misleading micro signals with underlying consumer resilience factors. subgraph Signals["表面疲软信号与品类冲击"] A["必需消费品/零食承压<br/>(Pepsi / Frito-Lay / Hershey)"] B["GLP-1 减肥药普及<br/>改变饮食习惯与热量摄入"] A --- B end subgraph Reality["底层经济支撑与真实韧性"] C["零工经济 & 社交媒体创作者经济"] D["创业潮达到二十年新高<br/>(AI 降低技术与营销门槛)"] E["账外现金流与劳动力吸纳"] C --> E D --> E end Signals -.->|"错误推论：消费彻底崩溃"| F["错误宏观结论"] Reality -->|"实际表现：强劲韧性"| G["健康的整体消费与信贷指标"] ``` --- ### [57:21](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=3441s) 信贷违约真相与“虚假崩盘论”的心理陷阱 在日常生活中观察美国大众的消费状态，与网络舆论营造的末日叙事形成了刺眼的对比。现实中无论是高端商场还是平民购物中心都人流如织，航班场场客满，电影院座无虚席，普通人照常租赁新车、上班工作、按时还账。然而，互联网上始终充斥着一种诱人的悲观论调，宣称“这一切都是假象，底层已经彻底被掏空，华尔街只顾沉迷于美光科技（Micron）的财报而对眼前的断崖视而不见”。 华尔街大型银行的实际财报彻底粉碎了这种末日论调。过去三年里，每逢摩根大通（JPMorgan Chase）等发行信用卡的大行发布财报，外界和分析师总在急切追问底层收入人群是否已经支撑不住。如今大行高管甚至不再等待提问，直接在业绩电话会开场白中主动澄清：违约率处于历史常态区间甚至极低水平，车贷正常偿还，账单按期支付。不仅传统银行业如此，服务更低 FICO 信用分客群的“先买后付”（BNPL）代表平台 Affirm，其贷款违约率同样极低（仅约 2.3% 左右）。在企业盈利和就业依然维持增长的大背景下，期待出现所谓各大银行突然集体爆雷的“美国银行时刻（Bank of America moment）”，在经济学基本逻辑上概率极低。 - **创业潮与就业吸纳**：全美新企业注册成立数量触及二十年高点。生成式 AI 正在显著拉低营销、技术与流程壁垒，让拥有好点子的创业者能迅速扩大规模，进而持续雇佣员工，这部分新增就业并未被每月官方非农数据完全覆盖。 - **年轻一代的消费转移**：在年轻客群中，酒精消费和去酒吧的频次有所下降，资金正在流向体育博彩和各类游戏化下注（Gamification of betting），年轻人群依靠多元化手段维系着充裕的流动性预算。 --- ### [01:02:46](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=3766s) 摩根士丹利的资产配置与集中持仓选股之道 面向 2026 年下半年至 2027 年的市场展望，从大类资产配置（Asset Allocation）到具体股票筛选需要双轨并行。在大类资产配置层面，摩根士丹利团队已超配美股（Overweight US equities）一年有余，并坚定保持该配置。尽管美股经历了 AI 资本开支高度集中带来的波动，但健康的宏观环境和稳固的盈利增长提供了坚实支撑。为了防范 AI 交易过热带来的回撤风险，资产配置组合中通过配置实物资产（Real Assets）、多空对冲基金（Long/Short Hedge Funds）以及精选的新兴市场（EM）来实现风险对冲。 美股市场在遭遇波动时展现出了极强的自愈与轮动能力。夏季半导体板块（Semis）遭遇抛售调整时，标普 500 指数从高点到低点的最大回撤仅有 3%。这得益于健康的板块轮动：半导体虽占大盘约 18%，但当资金流出芯片股时，迅速被占比 35%~40% 的科技七巨头（Mag 7）其他成员，以及占比约 10% 的医疗保健（Healthcare）和占比约 15% 的金融板块（Financials）所承接，有效阻止了指数的系统性下挫。相比之下，韩国等高度依赖单一或两只核心股票的集中型市场，所受到的冲击则剧烈得多。 在具体个股精选层面，团队继承了传奇策略师拜伦·维恩（Byron Wien）自 1995 年在摩根士丹利股票研究部奠定的投资哲学： - **聚焦长期投资**：避免追逐短期噪音，建立跨越周期的投资组合连续谱体系。 - **高集中度组合**：严格管理包含 40 至 50 只优质个股的精选组合，在确保行业分散对冲的同时，依靠核心阿尔法驱动长期超额收益。 ### [01:05:20](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=3920s) 投资组合构建哲学：核心持仓与大构想的结合 Dan Skelly 详细阐述了摩根士丹利旗下 50 支股票投资组合的底层构建逻辑。在这个组合中，绝大部分持仓（约 40 到 45 支股票）属于“核心持仓”（Core Holdings），每支权重在 2% 左右。这些核心标的旨在紧跟大盘收益，控制组合的追踪误差（Tracking Error）。但仅有平均表现还不够，组合还会专门预留 5 到 6 个仓位留给高信念的“大构想”（Big Ideas）。Dan Skelly 指出，这种设计深度继承了 Byron Wien 的投资哲学：如果在完整的市场周期中能抓住几个真正的变革性机会——就像 10 年前的英伟达（Nvidia）、电子支付浪潮中的 Visa 和万事达卡（Mastercard），或是受益于生命科学爆发的赛默飞世尔（Thermo Fisher Scientific）——这些超级大牛股带来的超额收益就能在漫长的岁月中大幅跑赢平均指数。 面对当前的宏观环境，Skelly 对客户的核心建议是：尽管 6 月份需要对 AI 和半导体产业链的过热保持警惕，但眼下许多被错杀的板块已重新浮现投资吸引力。更重要的是，AI 的采用浪潮将持续整整十年，投资者应当重仓那些兼具“质量因子”（Quality）的 Fortune 100 行业巨头。尽管“质量风格”在过去几年乃至今年部分时间受到冷落，但最近三个月的轮动表明，正如首席策略师 Mike Wilson 所判断的那样，高质量资产正在重新成为市场宠儿。在利率长期偏高的背景下，企业是否能真正通过 AI 提效并拉开与对手的盈利差距，成了最硬核的考验。 ```mermaid flowchart TD accTitle: 50支股票投资组合架构 accDescr: 展现核心持仓与超额大构想结合的双层资产配置机制 subgraph CoreHoldings["核心底仓 (约 40-45 支)"] A["每支权重约 2%"] --> B["严控跟踪误差，匹配大盘基准表现"] end subgraph AlphaIdeas["大构想仓位 (约 5-6 支)"] C["押注十年级产业趋势\n(如早期的 Nvidia / Visa / Thermo Fisher)"] --> D["周期内贡献巨大超额Alpha"] end CoreHoldings --> E["50 支精选高质量股票投资组合"] AlphaIdeas --> E ``` - **组合仓位分布**：40–45 支核心股票（各占约 2%）搭配 5–6 个跨周期“大构想”重仓。 - **AI 赋能生命科学**：Dan Skelly 极度看好 AI 对医疗保健领域的颠覆，指出 AI 已经将一期药物发现（Phase 1 drug discovery）原本长达 6 至 7 年的研发周期直接压缩到 12 个月以内，大幅加快更安全药物的上市节奏。 - **医疗板块外溢赢家**：除了核心药企，生命科学工具、耗材供应商以及临床试验外包管理平台（CRO）同样具备巨大吸引力；该板块此前接连消化了疫情红利消退、2022 年加息周期以及成长资金被科技板块抽血的三重打击，当下已出现明显的重估拐点。 --- ### [01:07:35](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=4055s) 自下而上的选股与选错个股的现实挑战 Josh Brown 询问该组合是自上而下（Top-down）按行业看好医疗等板块，还是纯粹自下而上（Bottom-up）筛选盈利增速最高的个股。Dan Skelly 回应称，该团队 30 年来一直以自下而上为主，行业配置通常只在基准上下浮动 2% 至 4% 的小幅超配或低配，用以控制追踪误差。但借用 KKR 宏观策略主管、摩根士丹利老友 Henry McVey 的话，“你必须有自己的主见（You got to have a view）”，团队会把优质的底层个股逻辑与宏观行业观点进行融合。 随后讨论转向主动选股中最折磨人的场景：行业方向看对了，但挑中的公司由于管理层执行不力或者产品偏差而掉队。Josh 形象地打趣这就像“看准了德比大赛（Derby），结果买错了马”甚至是“本来要去肯塔基德比，结果坐错飞机飞到了贝尔蒙特（Belmont）”。Dan 直言软件行业（Software）就是过去七八个月中最典型的案例。今年 2 月中旬，随着新版 Claude 模型插件等技术问世，市场上涌现出极端悲观情绪，软件板块指数（如 iShares Expanded Tech-Software Sector ETF，代码 IGV）暴跌，整个行业甚至被按“即将被淘汰（Obsolescence）”的价格抛售。 - **逆向观点兑现与分化**：Skelly 在 2 月中旬逆势提示软件股严重超跌，他当时论证私营 AI 实验室绝不可能以“消灭所有美国大企业”作为向华尔街推销的商业模式；AI 与软件业更可能形成深度的“长期相互依存（Chronic interdependence）”。 - **个股选择的分化残酷性**：虽然大方向判断完全正确，但选股结果千差万别——选了 CrowdStrike 与选了 Adobe 的收益天差地别；摩根士丹利在该板块抓住了微软（Microsoft）和 ServiceNow 的反弹，并在一年多前清仓了 Adobe。 - **高估值止盈案例**：对于 Palo Alto Networks，团队在其估值触及 70 倍市盈率（P/E）的区间上限后进行了清仓，认为市场已充分消化其作为 AI 受益者的溢价，作为 50 支股票的集中组合不能承担过高的跟踪误差风险。 --- ### [01:11:21](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=4281s) 被市场严重误读的资产：数据密集型与强监管护城河 Josh 追问 Dan Skelly 在当前组合中，最坚信哪家公司被市场严重错误定价。Skelly 提出，今年 2 月市场掀起了一股将各类核心数据资产“定价为即将过时”的恐慌狂潮，波及了软件、保险经纪乃至财富管理行业。但真正拥有排他性数据壁垒和监管特许权的公司，不仅不会被大语言模型取代，反而会成为超级受益者： 1. **标普全球（S&P Global，代码：SPGI）**：虽然市场担忧其金融数据业务受到 AI 威胁，但市场完全低估了其发债评级（Issuance）业务的刚性护城河。在当前全球巨额再融资与债务滚动的浪潮中，债券投资者需要合法合规的权威信用评级，监管机构也强制要求评级，“债券投资人绝不可能只凭 Claude 的一句话就买单（They're not going to take Claude's word for it）”。 2. **纳斯达克（Nasdaq，代码：NDAQ）**：今年整体股价表现平平，但底层盈利持续保持 15% 左右的中高双位数增长，估值倍数依然受到压制。纳斯达克的核心由金融科技、交易所上市以及皇冠明珠般的“市场数据业务”三者协同构成。随着未来科技独角兽 IPO 窗口重新打开，加上对其独家专有数据进行 AI 优化与商业化，纳斯达克具有极高且尚未被充分计入的重估潜力。 ### [01:13:47](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=4427s) 市场信号、止损哲学与机会成本 Josh Brown 谈起自己近期在纳斯达克公司（Nasdaq Inc.）上的操作教训：尽管过去长期获利，但今年春天市场走势明确发出警告，触发了他的止损机制。当时他的困惑在于：AI 颠覆浪潮全面铺开，作为高度监管、坐拥独家数据且具备资本市场周期顺风的交易所龙头，纳斯达克怎么可能分不到一杯羹？摩根士丹利的分析师 Mike Cypress 也一直积极看好该股，财报连续两三个季度均有强劲的业务 AI 化证据，但股价反应却始终萎靡不振。这引出了核心议题：投资者究竟在何时必须承认自己看错了并果断斩仓？ Dan Skelly 直言，对于管理约 50 只股票组合、年换手率维持在 30% 到 40% 且面对税收敏感型零售客户的投资组合而言，僵化的“硬性止损”机制过于教条。团队采用的是量化与基本面结合（Quantimental）的二元决策树。在获利层面，像早前投资的 Palo Alto Networks 经过估值重塑（Rerating）并持有一年半到两年后，已经兑现了原有逻辑；此时若将其留在组合中，它就会从核心标的退化为第 40 号想法，而它的吸引力已经比不上备选池里的第 41 号新标的。当新资金流入时，如果投资者不愿意以现价增持它，继续持有就会产生极大的摩擦与机会成本。 - 摩根士丹利团队管理大约 50 只持股的投资组合，目标年换手率为 30% 至 40%。 - 每当建仓一只新股票，Dan Skelly 带领的 7 人投资经理团队会撰写长达 20 页的内部独立研究报告。 - 团队不仅借助摩根士丹利 25 年积累的卖方研究资源与华尔街公开研报，过去 12 年还系统性建立与平台上各类买方资产管理公司（Buy-side PMs）的深度交流网络，构建全方位信息马赛克。 ```mermaid flowchart TD accTitle: 投资标的审查与决策流 accDescr: 评估投资组合中股票继续持有或卖出的逻辑判断流程 Start["标的出现异动 / 达到持有期限"] --> CheckGain{"是获利重塑还是逻辑恶化？"} subgraph GainScenario ["获利与机会成本"] CheckGain -- "实现预期估值重塑" --> OppCost["成为组合中边际靠后标的"] OppCost --> NewCash{"愿意以现价追加新资金吗？"} NewCash -- "否" --> TrimSell["获利了结，换入更优质备选标的"] end subgraph LossScenario ["逻辑受损与风险归类"] CheckGain -- "股价反向走势 / 业绩受挫" --> CoreThesis{"原核心逻辑是否被破坏？"} CoreThesis -- "盈利/护城河暂时性受损" --> HoldManage["控制仓位，等待价值回归"] CoreThesis -- "盈利/护城河永久性破坏" --> FastCut["果断止损离场"] end ``` ### [01:17:18](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=4638s) 周期性标的与长期成长股的止损差异 在面对亏损与逻辑误判时，Dan Skelly 强调必须穿透表面，审视核心问题：竞争格局、行业监管、管理层治理是否发生了不可预见的突变？公司的盈利能力和竞争护城河究竟是遭遇“暂时性受挫”还是“永久性损伤”？基于 20 年的投资经验，他总结出一条极为明确的法则：对于周期性（Cyclical）亏损标的，砍仓速度必须远快于长期成长型（Secular）标的。 周期股的下行往往具有强烈的负动量，如 2014-2015 年的石油钻井板块或 2007-2008 年的银行股，这类股票一旦盈利坍塌，跌幅往往不是 10% 或 20% 的温和调整，而且极难出现 V 型反转。相反，真正的长期结构性赢家在遭遇挫折时往往能重新崛起。例如 2010 年《多德-弗兰克法案》出台多宾修正案（Durbin Amendment）限制借记卡刷卡费率时，Visa 与万事达卡（Mastercard）在半年内暴跌 30%，但随后却走出十倍涨幅；苹果公司在 2010 年代也多次因广告收入担忧或三星竞争而大幅回撤，最终均重回巅峰。因此，不同类型的股票必须采取截然不同的风险管理手段。 - 周期性股票的下行势头会自我强化，盈利坍塌往往伴随估值剧烈杀跌，反转概率极低，需第一时间止损。 - 长期结构性赢家（Secular Winners）若遭遇系统性或周期性负面情绪冲击，可通过控制仓位来管理风险，但不宜轻易彻底放弃。 - 历史佐证：2010 年受多宾修正案监管冲击的 Visa 与万事达卡曾单边下挫 30%，随后长线成为涨幅超 10 倍的大牛股。 ### [01:19:01](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=4741s) 监管风险与竞争/治理风险的本质分野 Dan Skelly 进一步将个股面临的非宏观风险拆解为两类：政府监管风险与核心竞争/治理风险。历史数据表明，政府与监管层面的威胁绝大多数时候是暂时性、被市场过度恐慌性定价的。Josh Brown 以自身持有 Live Nation 的经历佐证：每当有州检察长宣布对其提起反垄断诉讼、市场恐慌抛售时，往往反而是买入良机，随后股价便屡创新高，这种担忧之墙（Wall of Worry）恰恰是优质标的持续上涨的温床。 相反，真正的“死亡丧钟”来自于企业内部的竞争性风险与治理衰退。当一家公司开始丢失市场份额、品牌影响力逐渐褪色时，即便是曾经的优质白马股也难逃厄运（如耐克近期遭遇的挑战、露露乐蒙估值回落，而 Abercrombie & Fitch 却能逆势翻盘）。Dan Skelly 回忆自己 18 年前作为文科背景初级分析师进入团队时，主管要求他重复做过十几次的研究作业：统计组合内 50 家公司高管薪酬与股东回报率之间的相关性，找出脱离常轨的异常值。企业治理变更、管理层丧失市场信誉、资本配置缺乏连贯性，是投资团队始终密切跟踪的定性警讯。 - 监管冲击与诉讼恐慌往往具有暂时性且容易被市场过度折价，形成逆向投资契机。 - 市场份额流失、品牌价值退化以及管理层治理失控，往往对公司价值构成致命且不可逆的摧毁。 - 价值投资范式的转变：过去依赖单纯低市盈率（P/E）指标寻找廉价资产（如传统教学中买 Bethlehem Steel 这类高负债低市盈率标的）的机械做法已被淘汰；成熟投资者如今普遍意识到市场具备极高定价效率，高估值本身往往正是市场对卓越护城河与优秀治理给出的合理定价。 ### [01:22:31](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=4951s) 市场微观结构与动量因子的主导地位 Downtown Josh Brown 指出，很多受传统训练的投资人往往误将高估值股票归咎于市场“错误定价”（Mispricing），但事实并非如此——投资者之所以愿意为戴尔（Dell）支付比康柏（Compaq）更高的估值溢价，单纯是因为戴尔质地更优，“这绝非定价错误”。Dan Skelly 对此深表认同，并补充了促成当前市场高度定价效率的两大深层驱动因素： 1. **信息获取与传播速度**：信息的激增、获取的便捷度以及实时流动，从根本上提升了市场的自然定价效率。 2. **市场微观结构的剧烈演变**：在机构端，CTA（商品交易顾问）和量化基金大行其道；在零售端，市场流动性极为充沛。一方面，美国约有 40% 的房产完全没有房贷负债，老龄化人口手中握有大量过剩流动资金；另一方面，年轻一代投资者拥有前所未有的交易工具与投机渠道。如今市场充斥着动量交易者与趋势追踪者，经典沃顿商学院 MBA 所学的那套基本面估值模型已不再是主导力量。 这一转变使得“动量”（Momentum）成为过去五年及近来市场中最核心的单一超额收益因子。Dan Skelly 强调，摩根士丹利财富管理虽然始终定位于偏好高质量标的、分散投资、关注尾部风险的长期投资者，但“无视技术面和动量指标，风险自负”。在当今时代，任何人都能在 Google 上瞬间查出一只股票的市盈率（PE ratio），估值已是公开信息，真正的超额逻辑在于动量本身——即该股票到底处于机构吸筹还是派发阶段。 ```mermaid flowchart TD accTitle: 现代市场定价与动量主导逻辑 accDescr: 展示信息效率与微观结构如何共同推高动量因子的重要性 subgraph MacroDrivers["驱动因素"] A["信息流动加速<br/>与全网普及"] B["微观结构变迁<br/>CTA/量化基金/充裕零售流动性"] end subgraph Impact["市场现状"] C["传统静态估值普及化<br/>(PE数据人尽皆知)"] D["动量因子主导市场<br/>(趋势追踪与资金积聚)"] end subgraph Strategy["实盘应对策略"] E["紧盯技术位与拐点<br/>(如50日均线/板块轮动)"] F["捕捉非理性异常<br/>(如停火预期买入能源)"] end MacroDrivers --> Impact Impact --> Strategy ``` - **市场资金面特征**：美国约 40% 的住房无抵押贷款，构成老年高净值人群丰厚的资本缓冲，与高频交易的散户及量化资本共同构成了全新的流动性环境。 - **动量与估值共存**：基本面决定长期质地，而动量决定阶段性生死；在人尽皆知的公开估值倍数之外，辨别资金积聚（Accumulation）或流出是投资实操的关键。 --- ### [01:24:27](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=5067s) 板块拐点轮动与地缘政治交易异常 当前市场正处于关键的风格拐点。Dan Skelly 指出，此前大热的半导体、硬件设备和电力基础设施板块自 6 月以来普遍横盘震荡，不仅多次反抽受阻于 50 日均线，主经纪商（Prime Brokerage）渠道的对冲基金多头敞口也未见重新回补。与此同时，资金正加速流向防御性与改善型板块： - **医疗保健（Healthcare）**、**软件细分领域**、**金融板块（Financials）** 以及 **能源（Energy）** 正在接棒走强。 - **能源板块的套利策略**：Skelly 透露了团队今年屡试不爽的一个操作规则——每当地缘局势传出“停火”（Ceasefire）消息导致油价暴跌时，他们便在投资组合中逢低加仓能源股。事实证明，这些所谓的停火协议往往维持时间极短，地缘溢价的反复为理性投资者提供了绝佳的买点。 --- ### [01:25:01](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=5101s) 年末季节性走势与 2027 年“科技七巨头”再度领跑展望 探讨至年底及更长周期的市场前景时，主持人们提到大多数年份美股往往以接近历史高位的姿态收官，季节性偏向极为强劲。Dan Skelly 分析认为，在 10 月财报季和 9 月密集投资峰会的数据催化下，企业盈利加速势头有望延续；尽管受中期选举影响市场可能会短暂震荡，但整体大势依然向上。 展望 2027 年，Dan Skelly 提出了一个明确的宏观主题判断：“2027 年的走势可能非常像 2023 年”： - **巨头重掌江山（Bigger is Better）**：普通公司在面对关税成本传导与经营杠杆压力时将面临严峻挑战，难以实现强劲同比加速。 - **科技七巨头（Mag 7）与高质量资产回归**：市场可能会再次重现 2023 年那种由 AI 赋能核心科技巨头及率先拥抱 AI 的医疗保健领军企业主导的结构性牛市。 - **小盘股与平庸标的受压**：对于中小盘股（Small Caps）以及标普等权重指数中的普通成分股而言，2027 年可能会是相对艰难、缺乏超额表现的一年。 --- ### [01:26:29](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=5189s) 播客尾声与摩根士丹利研究渠道指引 在节目的收尾阶段，主持人 Josh Brown 和 Michael Batnick 盛赞 Dan Skelly 本期的深度剖析与专业输出。Skelly 向观众介绍了摩根士丹利财富管理的研究成果获取途径： - 个人及摩根士丹利官方在 LinkedIn 平台定期同步市场观察。 - 通过摩根士丹利全球投资委员会（Global Investment Committee, GIC）发布季度策略报告，涵盖其管理的单独管理账户（SMA）动态。 - 针对半导体、软件等快速变化的行业，定期面向客户发布战术性（Tactical）专题报告。 最后，主持人们致谢幕后团队（John、Duncan、Nicole 等），并提醒听众继续关注下周更新，本期节目在轻松愉悦的告别声中圆满录制完成。 ## Notable quotes > [00:00](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=0s) everyone loves a trilogy — Dan Skelly > [06:10](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=370s) table stakes — Dan Skelly > [15:30](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=930s) 在 GLP-1 减肥药延年益寿、房地产市场因利率锁定而流动性冻结、再加上 AI 和机器人的全盘渗透下，我们的未来就是每个人都能活得极度长寿，但整天无所事事，而且买不起也住不进新房子。 — Dan Skelly > [20:24](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=1224s) 周五早晨飞往坦帕的航班 — Josh Brown > [24:45](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=1485s) Money good — Dan Skelly > [42:43](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=2563s) 鉴于我们对供应链和合作伙伴未来增长拥有极其敏锐且独特的全景视野，再加上我们手中充沛的净现金资产负债表，作为我们公司的股东，难道你不希望我代表你们持有这些生态伙伴的股权与筹码吗？ — Dan Skelly (引用 Jensen Huang) > [46:42](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=2802s) 到底有谁会去做空巴菲特？ — Josh Brown > [52:54](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=3174s) Shake Shack 又不能给数据中心供应炸薯条 — Josh Brown > [01:07:35](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=4055s) You got to have a view — Dan Skelly (引用 Henry McVey) > [01:07:35](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=4055s) 看准了德比大赛（Derby），结果买错了马 — Josh Brown > [01:11:21](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=4281s) They're not going to take Claude's word for it — Dan Skelly > [01:22:31](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=4951s) 无视技术面和动量指标，风险自负 — Dan Skelly > [01:25:01](https://www.youtube.com/watch?v=hpTOnDgTjHQ&t=5101s) 2027 年的走势可能非常像 2023 年 — Dan Skelly ## Mentioned Dan Skelly, Downtown Josh Brown, Michael Batnick, James Gorman, John Mack, Adam Parker, Mike Wilson, Brian Nowak, Jensen Huang, Byron Wien, Henry McVey, Mike Cypress, Ellen Zentner, Michael Cembalest, Gary Gait, Sam Altman, Morgan Stanley, Morgan Stanley Wealth Management, MSIM, Smith Barney, E*TRADE, Morgan Stanley at Work, Vanguard, CNBC, Squawk Box, Closing Bell: Overtime, Nvidia, Broadcom, Dell, Ciena, Cisco, Sun Microsystems, Lucent, Juniper, Amazon, Google (Alphabet), Microsoft, Oracle, SpaceX, Starlink, The Trade Desk (TTD), Super Micro Computer (SMCI), Charter Communications, Reddit, EchoStar, Johnson & Johnson (JNJ), Berkshire Hathaway, General Dynamics, Walmart, Wells Fargo, Eli Lilly, Apple, Chevron, PepsiCo, Frito-Lay, Hershey, JPMorgan Chase, Affirm, Visa, Mastercard, Thermo Fisher Scientific, CrowdStrike, Adobe, ServiceNow, Palo Alto Networks, S&P Global (SPGI), Nasdaq Inc. (NDAQ), Live Nation, Nike, Lululemon, Abercrombie & Fitch, Snowflake, Shake Shack, Chipotle, Russell 2000, S&P 500, Nasdaq 100, Claude, OpenAI, GLP-1, Durbin Amendment, Dodd-Frank Act ## Video description On episode 258 of The Compound and Friends, ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Downtown Josh Brown⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ and Michael Batnick are joined by Dan Skelly, Portfolio Manager at Morgan Stanley Wealth Management, to discuss the resilient U.S. economy, record earnings growth, the AI spending boom, Nvidia and Broadcom, whether today’s data center buildout looks anything like the dot-com bubble, the rotation out of semiconductors, the return of healthcare and financials, risks facing small-cap stocks, why the Mag 7 could lead again in 2027, AI’s impact on corporate productivity and profit margins, the strength of the American consumer, stock-picking in an increasingly efficient market, and much more! This episode is sponsored by Vanguard. Learn more about Vanguard bonds at: https://vgi.vg/4i0TW97 Take The Compound's 2026 audience survey and help shape the future of the channel: ⁠https://www.surveymonkey.com/r/LHC8QHD ►00:00 - Cold Open ►04:38 - Intro ►06:10 - Talking Markets on CNBC ►07:10 - The Current Environment ►20:35 - Is Earnings Growth Sustainable and The 493 ►42:42 - The Circular Financing Question ►52:54 - The Stock Market is Functioning Well ►01:02:41 - How to Pick Stocks Like Morgan Stanley ►01:25:01 - Markets into Year End and 2027 👕 Check out The Compound shop: https://www.idontshop.com 🎙️Listen to our podcasts: TCAF: https://podcasts.thecompoundnews.com/show/TCAF/ Animal Spirits: https://podcasts.thecompoundnews.com/show/animalspirits/ Ask The Compound: https://podcasts.thecompoundnews.com/show/ATC/ The Compound Newsletter: https://www.thecompoundnews.com/subscribe Linkedin: https://www.linkedin.com/company/the-compound-media/ Instagram: https://instagram.com/thecompoundnews Twitter: https://twitter.com/thecompoundnews Tik Tok: https://www.tiktok.com/@thecompoundnews Investing involves the risk of loss. This podcast is for informational purposes only and should not be or regarded as personalized investment advice or relied upon for investment decisions. Josh Brown and Michael Batnick are employees of Ritholtz Wealth Management and may maintain positions in the securities discussed in this video. All opinions expressed by them are solely their own opinion and do not reflect the opinion of Ritholtz Wealth Management. See our disclosures here: https://ritholtzwealth.com/podcast-youtube-disclosures/ The Compound Media, Incorporated, an affiliate of Ritholtz Wealth Management, receives payment from various entities for advertisements in affiliated podcasts, blogs and emails. Inclusion of such advertisements does not constitute or imply endorsement, sponsorship or recommendation thereof, or any affiliation therewith, by the Content Creator or by Ritholtz Wealth Management or any of its employees. For additional advertisement disclaimers see here https://ritholtzwealth.com/advertising-disclaimers. #thecompound #investing #stockmarket #stocks [Source](https://www.youtube.com/watch?v=hpTOnDgTjHQ)

[0:00] Dan Skelly: everyone loves a trilogy [6:10] Dan Skelly: table stakes [15:30] Dan Skelly: 在 GLP-1 减肥药延年益寿、房地产市场因利率锁定而流动性冻结、再加上 AI 和机器人的全盘渗透下，我们的未来就是每个人都能活得极度长寿，但整天无所事事，而且买不起也住不进新房子。 [20:24] Josh Brown: 周五早晨飞往坦帕的航班 [24:45] Dan Skelly: Money good [42:43] Dan Skelly (引用 Jensen Huang): 鉴于我们对供应链和合作伙伴未来增长拥有极其敏锐且独特的全景视野，再加上我们手中充沛的净现金资产负债表，作为我们公司的股东，难道你不希望我代表你们持有这些生态伙伴的股权与筹码吗？ [46:42] Josh Brown: 到底有谁会去做空巴菲特？ [52:54] Josh Brown: Shake Shack 又不能给数据中心供应炸薯条 [67:35] Dan Skelly (引用 Henry McVey): You got to have a view [67:35] Josh Brown: 看准了德比大赛（Derby），结果买错了马 [71:21] Dan Skelly: They're not going to take Claude's word for it [82:31] Dan Skelly: 无视技术面和动量指标，风险自负 [85:01] Dan Skelly: 2027 年的走势可能非常像 2023 年

[0:00] Dan Skelly: everyone loves a trilogy [15:35] we're all going to live longer, but we're going to have nothing to do and nowhere to live. [20:35] catching a Friday morning flight to Tampa [62:15] In vitro GLP-1... I'm going to be the last person left in this world

[0:00] So, this is gonna be fun. How are you [0:02] How long have you been at, um at Morgan? [0:04] >> 21 years out of college. Only feels like [0:07] 20. [0:08] >> Wow. [0:09] >> So, uh I'm [0:10] >> What office do you work in? [0:11] >> Uh at the moment, 757th A, but on my way [0:14] back to 1585 Broadway. [0:16] >> They move you guys around. [0:17] >> They move us around. They keep the real [0:19] estate moving. [0:20] >> Okay. [0:20] >> Um but as you probably well know, 1585's [0:23] gone through like a 5year construction. [0:25] Yeah. So, some of us were kind of off uh [0:27] grid, as I call it. Yeah. or this will [0:29] be my third time going back to 1585 in [0:31] 21 years. [0:32] >> Wow. [0:32] >> And I always say everyone loves a [0:34] trilogy. So that's my mantra. [0:37] >> So Morgan Stanley is so big that [0:39] whenever I meet a financial adviser from [0:41] there, [0:42] >> I ask them about other financial [0:44] adviserss that I know. [0:45] >> Sure. [0:45] >> In New York and they never know each [0:48] other. [0:49] >> That's wild. [0:50] >> But that's the size of the firm. I'm [0:51] saying [0:52] >> 100%. [0:52] >> Right. because there's so many different [0:54] offices and it's just I mean I I've said [0:57] this you don't have to agree or disagree [0:59] but I have said um Gorman was incredible [1:03] as a CEO as a visionary [1:06] >> and that's why is it 20 trillion now the [1:09] do you know [1:10] >> just just under that yeah between wealth [1:12] and MSIM [1:13] >> yeah it's something like 16 17 trillion [1:16] >> so basically stealing Smith Barney [1:19] brilliant move yes during the during the [1:21] height of phases like really well [1:24] executed [1:24] >> joint venture then we'll take a third [1:27] we'll take another third fine we'll take [1:28] the whole thing [1:29] >> buying Erade [1:30] >> the Morgan Stanley at work platform as a [1:33] lead generator [1:34] >> my opinion I think that's the key yeah [1:37] that's like that was incredible [1:38] >> yeah so I mean it's it's and you were [1:41] there you were there you watched the [1:42] whole thing happen [1:42] >> watch the whole thing when I started [1:43] Josh in ' 05 wealth as a percent of [1:46] overall revenues was 8% of the firm and [1:48] pro- fora for all the different deals [1:51] you just alluded too. It's like 60% of [1:53] the firm's revenues now. I'm Dan. [1:55] >> Oh, absolutely. Good to see you. [1:56] >> So, uh, [1:58] >> right. So, I I think what was the guy [2:00] before? Uh, Mac. John. [2:01] >> John Mac. [2:02] >> John Mac. [2:02] >> So, I think John Mack understood the [2:04] value of Let's Let's go heavily more [2:07] heavily into advice, but Gorman actually [2:08] executed it. [2:09] >> Absolutely. And John was the one who [2:11] went out and found James, who was at Mel [2:13] at the time, and had really revamped [2:15] Merryill's wealth business. And James, [2:17] prior to that, as you probably also [2:19] know, was a McKenzie consultant. So he [2:21] brought he brought this strategic [2:23] consulting background as well [2:25] >> and you said it like t sometimes timing [2:27] is everything. So he had the strategy, [2:29] the timing, the pricing. [2:31] >> Yeah. [2:31] >> And you know the multiple and the [2:33] rerating has come together since that [2:35] point. [2:36] >> Yeah. I wonder if there are still Smith [2:38] Barney guys walking around saying I was [2:39] Smith Bar. I was legacy Smith Barney. [2:41] >> You're looking at one. So quick story [2:43] quick story for you. In college [2:44] sophomore year I interned at a financial [2:47] advisor's office at Smith Barney. Okay. [2:49] My junior year, I had an like one of [2:51] these official analyst programs at MSIM. [2:54] Actually, hired into wealth in '05 [2:56] full-time. So, after the merger, [2:58] depending on what office legacy branch I [3:00] would go into, I was either a Smith Vney [3:02] guy or a Morgan Stanley guy. Very [3:04] strategic, [3:05] >> right? And then there were also Morgan [3:06] Stanley Dean Witter guys [3:08] >> 100% [3:09] >> predating this from 97. Yeah. Yeah. All [3:12] right. Well, [3:12] >> so it's been quite a evolution. [3:14] >> Yeah. It's a it's a cool front row seat [3:16] that you've seen. Thank you, Josh. I [3:17] appreciate [3:18] >> to to see that all develop. Um, how are [3:20] we looking, guys? [3:20] >> Headones on everybody. Headphones on. [3:23] >> All right. [3:24] >> Mute your devices. [3:25] >> Okay. To have my devices on my person. [3:28] Okay. [3:28] >> Thank you. [3:29] >> All righty. [3:30] >> Yeah. Let me let me do Let me do not [3:32] disturb. [3:33] >> Do not disturb. Competent friends. [3:35] >> I think I'm doing that right. [3:39] >> Today's show is brought to you by [3:40] Vanguard. To all the financial adviserss [3:42] listening, let's talk bonds for a [3:44] minute. Capturing value and fixed income [3:46] is not easy. Bond markets are massive, [3:48] murky, and let's be real, lots of firms [3:51] throw some flashy funds your way and [3:52] call it a day. But not Vanguard. At [3:55] Vanguard, institutional quality is our [3:57] standard. The same discipline the [3:59] world's largest investors demand at [4:00] Vanguard cost. It defines our approach [4:02] to active fixed income across 40 funds, [4:04] built on three pillars. Consistent [4:06] performance across market cycles, a [4:08] lowcost advantage that compounds into [4:10] better outcomes for your clients, and [4:12] intentional risk management. That isn't [4:14] a great quarter, it's a great decade. [4:16] That's the Vanguard difference. Check it [4:18] out at vanguard.comaudio. [4:20] That's vanguard.comaudio. [4:22] All investing is subject to risk. [4:24] Vanguard Marketing Corporation, [4:25] distributor. [4:40] >> All right. Thanks, John. Oh, boy. What a [4:43] treat. This is going to be this is going [4:45] to be a very I'm feeling like this is [4:46] gonna be a very special episode. [4:48] Nicole's nodding her head yes. [4:51] John's John's saying thumbs up. All [4:52] right, guys. We uh we are coming to you [4:56] live from Bryant Park in New York City. [4:58] This is America's favorite investing [5:01] podcast. It's called The Compound and [5:02] Friends. First time listeners, we [5:05] appreciate you coming by. Longtime [5:07] listeners, thank you guys so much. We [5:10] have a a very special guest today. first [5:12] time, first appearance here on the [5:14] compound. His name is Dan Skellyy. Dan [5:16] is a portfolio manager of equity. Wait, [5:20] what is equity maps that portfolios? We [5:23] have a lot of acronyms at Morgan [5:24] Stanley. So, it's an SMA portfolio. [5:26] >> Portfolio manager of the equity maps at [5:29] Morgan Stanley Wealth Management where [5:32] he oversees the equity model portfolio [5:34] team and thematic research project [5:36] products including alpha currents. They [5:39] love that. US Policy Pulse. He is lead [5:43] portfolio manager for the US model and [5:45] dividend equity strategies, part of a [5:46] suite of eight long only SMAs. He has [5:50] spent his entire career at Morgan [5:51] Stanley starting in 2005 as his [5:54] associate in the research division. He's [5:56] a regular on CNBC's Squawkbox and [5:59] Closing Bell Overtime. Ladies and [6:00] gentlemen, please say hello to Mr. [6:03] Dankelly. [6:05] Thank you. Thank you. The crowd almost [6:07] cannot be contained. All right. Um, tell [6:10] me about tell me about the squawk [6:12] appearances. I I watched two of yours um [6:15] recently. You do a pretty good job [6:17] there. Do you like do you like doing [6:18] television? Do you like getting the [6:20] firm's views out to the public? [6:21] >> I I love it. And look, I think it's [6:23] become in this kind of media technology [6:27] um intersection. It's become table [6:29] stakes for kind of what we do, right? [6:31] And so remember my first appearance on [6:33] CNBC was a 6 a.m. slot in the summer of [6:37] 2015 and it was with Becky, Joe and [6:40] Andrew. [6:41] >> And you know it was I would say it was [6:43] touchandgo to begin and you know as you [6:45] know from having done this world for so [6:47] long you [6:48] >> over time you get your feel for it. You [6:50] get your your groove and it's been [6:51] really fun to do it over the years. I [6:53] spent a year doing the 5 a.m. Oh, [6:56] >> and uh I asked I asked I asked theo the [6:59] host like who who are we talking to and [7:01] she said basically Singapore. [7:04] >> Yeah, Asia Markets. [7:04] >> I said, "All right, let's do it." [7:06] >> All right. Thank you. Thank you so much [7:07] for coming here. [7:08] >> Thanks for having me, Josh. I'm excited. [7:10] >> So, you as we mentioned prior to uh [7:13] officially starting the show, you sit in [7:15] a really interesting seat at one of the [7:17] largest firms on Wall Street, one of the [7:18] largest asset managers in the world, [7:20] quite frankly. Um, I wanted to get your [7:22] take just overall on the current [7:25] environment. There's a little bit of a [7:26] push pull right now. Um, [7:30] I think a lot of the investing public, [7:32] professionals and retail investors have [7:35] arrived at this point where they almost [7:37] have learned that they can't afford to [7:38] pay attention to the news anymore [7:40] because it's almost all negative. And [7:44] every time they get carried away with [7:46] one of these negative narratives, they [7:48] miss the next 20% in the S&P. Um, how do [7:51] you how do you uh help people with that [7:54] kind of like, you know, all right, I [7:56] know the news is bad, but the stocks [7:59] keep going up any either way. Like, how [8:01] do you help people with that those two [8:02] opposing ideas? [8:03] >> Yeah, absolutely, Josh. And candidly, [8:05] aside from this podcast, of course, we [8:07] we often uh advise and counsel a lot of [8:10] um, you know, normal retail investors to [8:13] turn it off, like just stop paying [8:14] attention every moment of the day. And [8:17] what's interesting is [8:18] >> except when you're on then then keep it [8:20] on. No doubt. Got it. Or you [8:22] >> I say the same thing. [8:23] >> Exactly. So look, I mean I think a lot [8:25] of and the other additional biography [8:28] piece I would just add to is I'm a [8:29] member of the firm's asset allocation [8:31] committee which is called the global [8:32] investment committee. And in that effort [8:34] right away from my day job picking [8:35] stocks were really focused on long-term [8:38] compounding uh diversification. And as [8:41] you well know better than anyone, the [8:43] last decade plus, maybe prior to even [8:46] the last few months, has been all about [8:47] concentration. And now the market of [8:50] course is broadening. And we're all [8:51] talking about broadening as per the last [8:52] several months. Um but at the end of the [8:54] day we've tried to counsel our clients [8:57] to stay focused on their goals, stay [9:00] focused on their risk tolerance, which [9:02] at times you know a lot of people who've [9:03] made wealth as you also know in a [9:05] concentrated fashion via [9:07] entrepreneurship entrepreneurship or [9:09] starting a uh coming up with a [9:11] innovation technology or something [9:13] really you know kind of uh innovative. [9:16] They're used to taking risk in a [9:18] concentrated fashion. It may not be the [9:19] most preferred way to stay wealthy over [9:21] time. And so having that diversified [9:23] bent is always top of mind for us. [9:25] >> Okay. Do you think more people are [9:26] worried about the next correction or [9:28] more people are worried about missing [9:30] out on S&P 8000 Dow 60,000? Like where [9:35] where do you think the bigger fear is at [9:37] the moment? [9:38] >> It's f it's interesting timing of that [9:40] question. If you had asked me that back [9:41] in May or June, it was definitely FOMO, [9:43] definitely missing out. Um, I think the [9:45] rotation and some of the implosion and [9:49] some of that first half momentum [9:50] leadership that we've that we've [9:52] experienced definitely has rebalance [9:54] that feeling, that sentiment. And so [9:56] today, I I think it's more balanced. I [9:58] don't think it's really one way or [9:59] another, but earlier in the year, it was [10:01] absolutely fear of missing out. [10:02] >> I think if you asked people took a poll, [10:04] next 10% move, I think it'd be close to [10:06] 50/50. [10:07] >> Yeah, I think you're right. I think [10:08] you're right. [10:08] >> And in May, it would have been like 7030 [10:11] >> 100%. Absolutely. And and what I would [10:14] say is, you know, look, we've kind of ph [10:16] we've coined this phrase that the [10:18] markets and certainly the econom [10:22] >> that was you or or um Ko Kramer I think [10:26] dovetailed off you but [10:27] >> he's welcome to do that. [10:29] >> All right. [10:30] >> So what I would say is we've come up [10:32] with this observation, this realization [10:34] as of the last year really that the [10:36] economy continues to be super resilient. [10:38] Everyone's been talking about it and it [10:40] really continues to look through these [10:42] policy shocks, these inflationary [10:43] pressures. [10:44] >> Is it the economy that's resilient or [10:46] S&P earnings that are resilient or both? [10:49] I think it's the economy mostly and I'll [10:51] I'll circle back to earnings no doubt [10:53] which is a great um comment but earnings [10:56] have been really astronomical and I'll [10:58] come back to my theory on that in a [11:00] second but the economy I think what [11:02] isn't as realized today and it certainly [11:03] wasn't four years ago Josh was this idea [11:05] of how much the economy has paradigm [11:08] shifted away from cyclicals away from a [11:11] normalized income distribution and add [11:14] on top of that the AI spending super [11:15] cycle and you've got I think a very non [11:18] atypical economic cycle. And so we've [11:21] seen this resilience. And on top of [11:23] that, coming back to your question on [11:24] earnings, look, earnings at the index [11:26] level did 28% year-over-year growth in [11:28] the second quarter. Um, if I look at the [11:31] median company, it was 14% growth, which [11:33] I don't think gets talked about enough. [11:35] So why is the median the average company [11:38] experiencing that much growth? It's not [11:40] just AI, and I know we're going to talk [11:42] about AI uh at Nauseium today, which I [11:44] look forward to. My my presumption I'm [11:46] trying to prove this with data and I [11:48] know our mutual friend Adam Parker is a [11:50] friend of the show and we have been [11:51] talking about this a lot recently [11:53] personally I presuppose that we had a [11:56] synthetic tariff trade related earnings [11:59] or economic cloud or hangover in the [12:02] first half of last year. Emerging out of [12:04] that right now 12 months later is a [12:06] equally or proportionate synthetic [12:08] operating leverage earnings boost for [12:10] the average company who had more pricing [12:12] than I think most people would have [12:13] perceived [12:14] >> and then didn't have to give it back. [12:15] >> Correct. We passed through on on the [12:17] tariff front which remember tariff was [12:19] the headline for like six or nine months [12:21] ago. [12:21] >> Trump really is a genius. Some people [12:23] would say that. [12:24] >> So every company had to take price [12:26] >> most companies the average company [12:28] >> just to muddle through [12:29] >> to muddle through but they passed on 60 [12:30] to 70% of it. [12:32] >> Right. and and the prices don't go back [12:34] down when the tariff emergency is over. [12:36] >> Dan, you think that's that's tariff [12:38] related? These this margin expansion. [12:40] So, this is the 493. [12:41] >> I think part of it is tariff related on [12:43] the pricing perspective, which I do want [12:45] to come back to as a new relatively new [12:47] suburban homeowner in Long Island. My [12:49] wife is Long Island. I'm New York City, [12:51] but I'm an adopted Long Island uh child [12:54] now. Son, uh I learned you're going to [12:56] love it. [12:56] >> Uh it's it's been a wild ride so far. [12:59] >> I never left. [12:59] >> How about that LIE? Um what I learned [13:02] from my landscaper after co is when he [13:05] had to take up price 15 to 20% because [13:08] of cost because of all these things he [13:09] he never took it back [13:11] >> and so it's just an it's a small [13:12] anecdote which speaks to this broader uh [13:14] thread. So I think part of this part of [13:17] your great question is no doubt pricing [13:20] surprising to the upside. Secondly, I [13:22] think coming through on this uh and it's [13:25] hard to obviously prove dollar for [13:27] dollar, but we hear it a lot in terms of [13:29] surveys and we're seeing it in the [13:31] transcripts is AI productivity on top of [13:34] existing workforce. What your margin [13:36] math at the moment on the chart doesn't [13:38] yet show is a labor lever being pulled. [13:41] >> So, this go higher. [13:42] >> Absolutely. [13:43] >> Wait, wait, say more. What's the labor [13:45] labor lever? [13:46] >> We're in this, if you zoom out, we're in [13:48] this no hiring, no firing zone. And [13:50] we've been there for a long period of [13:51] time, couple years it feels like. [13:53] >> And what I would argue is if you think [13:55] about what people were talking about in [13:57] terms of all the different AI boogeymen [13:58] in January, SAS SAS apocalypse software [14:01] going away, that's been thus far [14:03] disproven with some dispersion. We could [14:05] talk about it. Labor apocalypse also [14:07] disproven. What is happening at the [14:09] Fortune 500 level visav our data and our [14:12] surveys, it's productivity on top of [14:14] existing labor force, right? Right. And [14:16] so I think that is coming through in the [14:18] margin line. To to my earlier point, I [14:20] don't think what's coming through just [14:21] yet and I think it's a 12 to 24 month [14:24] time horizon is labor actually being [14:26] pulled in terms of additional margin and [14:28] earnings and that's going to be largely [14:30] AI driven. [14:30] >> Seriously, earnings are going up with [14:32] without additional headcount. [14:33] >> That's what it looks like at the moment. [14:35] So nor right so normally in order to [14:37] produce the revenue so I think the [14:39] revenue increase for Q2 year-over-year [14:42] was also an incredible number 15% [14:44] >> on 6% nominal GDP. [14:46] >> So historically in a more cyclical [14:48] analog economy that's more goodsh heavy. [14:52] >> You're not doing 15% revenue growth with [14:54] no headcount growth. No way. [14:56] >> Well said. [14:56] >> You need people on an assembly line [14:58] literally welding things together and [15:00] packing them in boxes. [15:02] You're saying now the next tailwind [15:04] might be companies continue to grow [15:07] revenue which translates into earnings [15:10] without the conccommittent addition of [15:12] another 10% you know labor force. I'm [15:15] not saying this is great society however [15:18] we're in the business of earnings and it [15:20] should be good for the audience. I think [15:21] you summarized it perfectly, Josh. And I [15:23] think this I think to your to your [15:25] embedded in your statement was this [15:28] longer term debate around socioeconomic [15:31] effect and [15:31] >> just have less babies, it'll be fine. [15:35] >> I mean, honestly, we can't solve that on [15:36] this show [15:37] >> that we talked about Asia markets coming [15:38] on like that is a phenomenon going on [15:41] across Asia, across Europe, and no doubt [15:43] across the US. Yeah. [15:45] >> And so, you know, we'll see. the the [15:47] joke I've been saying is in terms of [15:49] GLPs and longevity on top of a housing [15:52] stuck in the house locked in housing [15:54] market on top of AI and robotics is [15:57] we're all going to live longer but we're [15:58] going to have nothing to do and nowhere [15:59] to live. So that's like our future [16:01] >> and and uh a shortage of 18-year-olds [16:04] apparently. I was reading about Syracuse [16:05] University this this week. They're not [16:07] going to hit their admissions uh targets [16:10] yet again. And obviously there are some [16:12] Syracuse specific issues like the [16:14] weather, but the bigger picture is there [16:17] just aren't going to be as many young [16:19] people [16:21] prospectively. Um, [16:22] >> and the nature of work is going to [16:24] change too. Back to AI, what is the [16:25] entrylevel legal audit? What does that [16:28] all look like? Syracuse is very near and [16:30] dear to my heart. I was fortunate enough [16:31] to marry a former Syracuse laxer who [16:34] played for Gary Gate in her day. And um [16:37] so it's uh the article it's definitely [16:40] batted around [16:40] >> threw that one factor in amongst many [16:43] but um to your point Europe, China, the [16:45] Koreans are not reproducing anymore. So [16:48] it's a so I think I've always been glass [16:52] half full about robotics, automation, AI [16:54] likewise [16:55] >> in that we're sort of going to need it. [16:57] >> Like we're going to have a nursing [16:58] shortage here pretty soon. We're going [16:59] to have shortages of specific careers [17:02] and it'll only get exacerbated by a [17:04] slower population growth. Well said. And [17:06] oh, by the way, let's talk about there's [17:08] been so much myopic focus on AI and [17:10] there should be a lot of that is [17:11] justified. But let's focus for a minute [17:13] on some of the other massive initiatives [17:15] impacting the earnings picture, the [17:17] economic picture, etc. reshoring, which [17:19] I don't think gets enough press and [17:21] enough ink. But we're going to have [17:23] according to a lot of the work Morgan [17:24] Stelly's industrials team has done. [17:26] We're going to have a lot more [17:27] factories. We're already seeing evidence [17:28] of that. Will all of those factories be [17:31] filled by the next 18 to 35 year olds? [17:34] No. a lot of it's going to be automated [17:36] and so there's a an effect and an [17:38] initiative under reshoring and [17:40] production that also questions that [17:42] demographic risk. But I think the [17:44] robotics is no doubt the part of the [17:45] answer. So it sounds like you're fairly [17:48] sanguin on where we sit today. Not not [17:50] that you don't think a correction is [17:52] possible, but uh you you sound as though [17:56] the earnings growth um looks to be [18:00] sustainable based on these tailwinds [18:01] that you're talking about. My [18:03] presumption is the we have the midterms [18:05] coming up right around the corner. It's [18:06] going to be Labor Day this weekend that [18:07] that flew by. And the phrasing I'll go [18:10] back to the outset of this conversation [18:11] that I've come up with over the last [18:13] year is policy shocks, inflation [18:15] pressures, all of these factors and [18:17] dynamics that used to matter more to [18:19] markets are like pop-up ads today. They [18:21] kind of come and go and the main [18:23] narrative keeps coming back to earnings [18:25] and AI. And so, like, being [18:27] intellectually honest, knowing that [18:29] Liberation Day mattered for a minute for [18:31] the market, knowing that Iran has [18:33] mattered in March and April and other [18:34] points in time, can I can I [18:36] intellectually say that the midterms [18:38] aren't going to matter? No. But to your [18:40] point, Josh, because the earnings [18:41] backdrop is so strong, I think whatever [18:44] uh draw downs or corrective uh uh [18:47] experience we get is super moderate. [18:49] >> Well, how about this? I think one of the [18:50] reasons why all of these things that [18:52] we've dealt with over the years that [18:54] would have at a minimum derailed the [18:57] economy if not thrown it right into a [18:58] recession. I think part of the reason is [19:00] there's so much money in the system. [19:03] Absolutely. [19:03] >> And I think it's underappreciated how [19:05] much that is distorting not in a bad way [19:08] what otherwise could have happened in a [19:09] different generation. Now the assets [19:11] could shrink in a bare market and fear [19:13] can return obviously but think about [19:15] like all the the secondaries that we're [19:17] seeing. Anytime something goes bad, it's [19:19] bought up immediately. And that is [19:21] >> Google's issuance ahead of the big [19:23] SpaceX deal bought up immediately. [19:24] >> Really impacting the economy and the [19:26] market in an underappreciated way. [19:28] >> No, I think that's absolutely right. And [19:30] that's like the residual benefit of a [19:32] 15-year bull market. Started out as [19:34] fang, then went to Mag 7, then went to [19:36] AI Capex, and now I agree with you. The [19:39] most healthy thing I would argue in [19:41] terms of the duration of this cycle is [19:43] the rotation we just saw. It's as if you [19:45] really needed the semis in June and some [19:48] of the other first order AI capex [19:50] winners to roll over [19:51] >> here healthare take it [19:52] >> to get the health care sector to get [19:54] mathematically get the mag seven working [19:56] and we saw that in you know selectives [19:59] finance financial stocks working all [20:02] year and all year last year for the most [20:03] part um healthcare this year small caps [20:07] coming out of nowhere industrial had a [20:09] run I know they've pulled back but they [20:10] had a big [20:11] >> and kudos to Mike Wilson our other [20:12] friend and partner um you know many [20:14] years whom it used to be my direct boss [20:16] uh 10 years ago who had a small caps [20:19] call uh earlier this year late last [20:21] year. So yeah, I think that's been one [20:23] of the surprises as well. [20:24] >> Okay. Um are we going to see the [20:27] dramatic earnings growth gains that have [20:30] now spread from the S&P into the midcaps [20:32] and the small caps? Is that sustainable? [20:35] >> So I feel like that space is a lot [20:37] trickier. Yeah. Because on the one hand [20:40] you would argue like the sectors that [20:42] are disproportionately overweighted to [20:44] small and mid industrials financials [20:47] have a lot of as I've mentioned so far a [20:49] lot of idiosyncratic positives like [20:50] capital market cycle uh rates uh [20:54] building production no doubt Josh FOMO [20:57] for sure that risktaking liquidity you [20:59] mentioned great [21:00] >> all the M&A leads leads to more M&A [21:02] >> and it should and look at how the [21:03] biotech sector is acting of late as an [21:05] example so here all of that can be true [21:08] on the one hand and then on the other [21:09] hand I think it can be true that rates [21:12] backing up particularly for that lower [21:14] quality cohort of small caps should be [21:15] an issue right and we've talked about [21:17] this phenomena as of the last 15 years [21:20] we went through this massive [21:21] monetization cycle in privates and [21:24] something like 80% of the companies in [21:26] the US today that generate 100 million [21:28] plus revenue are private so basically [21:30] your small cap allocation as a private [21:32] in as a retail investor could have just [21:34] been in the private market and not in [21:35] Russell [21:36] >> it's an alt now Totally. And so that [21:39] you've had like almost this negative [21:40] selection bias in the public in the [21:42] Russell 2000 which is like whatever the [21:44] stat is 40% of that index isn't [21:47] profitable. And so like I think on the [21:49] one hand you have [21:50] >> positive drivers but on the other hand [21:52] beware rates number one. And beware AI. [21:55] Look face it. I think AI we we've talked [21:57] about it earlier in terms of the big [21:59] caps and when does that show up in [22:00] earnings but I think AI could be really [22:03] tricky for small caps in the sense that [22:05] number one in some of these industries [22:08] AI is going to dis disintermediate [22:10] certain industries completely and they [22:12] may be more small cap in nature number [22:14] two I would argue that the the AI [22:16] adoption wave we're going to experience [22:18] which is going to be a decade experience [22:21] may not also be felt in terms of the [22:22] right tail from the small caps because [22:24] in many of cases those less profitable [22:27] small caps don't have the capital to [22:28] invest in AI. So they might have a left [22:31] tail risk disintermediation they may [22:33] have a right tail risk in terms of not [22:34] participating in the AI adoption. [22:36] >> So small caps will do what small caps do [22:39] which is periodic uh moments of [22:42] inspiration followed by disappointment [22:45] and then when you're so disappointed all [22:47] of a sudden they start to rally again. [22:49] Very different return profile than large [22:51] caps. [22:52] >> I think that's well said. I think it's [22:53] like catching a Friday morning flight to [22:54] Tampa. You're going to have a lot of [22:57] periods of calm and you're going to have [22:58] second a couple seconds of turbulence. [23:00] >> I like I like that. I like the popup ad [23:02] metaphor even better though. That's a [23:04] good one. [23:04] >> Thank you. [23:05] >> Where Yeah, we see it. It's peripheral. [23:08] Can't wait to close that window and not [23:10] think about it. So, I'm that way with [23:12] the midterms. I have I have no opinion [23:14] of what's going to happen because I'm [23:15] paying as little attention as possible. [23:17] >> I think that's wise. So maybe the House [23:18] turns over, the Senate doesn't, and [23:20] nobody really does anything differently. [23:22] The one big thing that might change is [23:25] all these uh astroturfed data center [23:28] protests might very quickly go away [23:30] because there's no longer a political [23:32] opportunity to say how much you hate [23:34] Microsoft. Like all of a sudden, we were [23:36] talking about it, talking about it, then [23:38] the election comes and goes and nobody's [23:39] talking about it anymore. I could [23:40] picture that. [23:41] >> I think that's spot on. I was at a [23:42] client dinner earlier this week out on [23:44] the island with Brian Noak who's our [23:46] leader thought leadership uh thought [23:48] leading um industryleading internet [23:50] analysis as you well know and we were [23:52] talking about really the midterms but [23:54] also 28 and this idea that the midterms [23:57] like you just said are going to come and [23:58] go and then people are going to focus on [23:59] 28 and what you might feel right is just [24:02] a very fast pull forward [24:04] >> in terms of yes we may have a [24:06] presidential election in 28 [24:08] >> yeah they every four years or so they [24:10] seem they seem to come back Okay, go on. [24:12] >> What I was just saying is Brian's view [24:14] and it makes it's intuitive to me is [24:16] past the midterms, right? And keep in [24:19] mind that AI capex and momentum [24:21] implosion in June, which has struggled [24:24] technically to come back to the 50-day [24:26] if you're like memory stocks or I think [24:28] it's very interesting when you get past [24:30] the midterms that you have maybe almost [24:32] a pull forward ahead of 28 in terms of [24:34] the data center trying to get as much [24:36] done ahead of that. [24:37] >> So that would be a negative catalyst. [24:40] Um [24:41] >> eventually it means to feel great in the [24:43] moment. [24:44] >> The hangover. [24:45] >> Well, so [24:47] the the last capex cycle that became a [24:50] pull forward was Y2K [24:52] >> and the hangover from that started to be [24:54] felt in the second quarter of 2000. So [24:57] wait, wait, wait. People are not going [24:58] to buy this many Intel chips every [25:00] quarter. [25:00] >> Oh, that's not good. So that's the thing [25:03] that I most worry about. the amount of [25:05] the earnings growth that's expected to [25:08] come from the hyperscalers the large the [25:11] 50 largest semis and memory companies [25:14] and then the Dell computers of the world [25:16] it's a large it's not all of the [25:18] earnings growth [25:19] >> it's a large amount of it [25:21] >> and if it goes into reverse [25:24] um I don't know that the market's going [25:25] to treat that well from 21 times [25:27] earnings [25:28] >> and also Broadcom today what do they 90% [25:30] growth whatever it was and the stock [25:32] fell fell 5% [25:33] >> it wasn't 91% Same with Nvidia recently [25:36] which you know blew out the guidance and [25:38] even said guidance amid capacity [25:40] constrained backdrop and the stock [25:43] worked well on the day and then really [25:45] hasn't followed through. [25:46] >> So what does this what does this tell [25:47] you about where the stock market is [25:48] today? Now the environment could change [25:50] but people are simultaneously worried [25:52] that the earnings are too high, the [25:54] earnings estimates are too high and even [25:56] when companies destroy earnings to the [25:59] tune that we've never seen before, the [26:01] stocks still aren't working. Is that [26:03] bullish? It doesn't really sound it. [26:04] >> I think it's in my opinion it's it's [26:07] good for the duration again. Like it [26:09] comes back to this idea of I think we're [26:11] in a longer cycle and like secular [26:14] here's the good news. The way of the [26:15] capitalism works the secular bulls last [26:17] a very long time. Like they last 20 25 [26:20] years on average and the good news is [26:22] also that the secular bears tend to be [26:24] half that duration. So we had a nasty [26:27] bare market as you well know coming out [26:29] of o out of the 2000 internet bust and [26:32] really since 09 010 it depends if you [26:34] adjust it for inflation and S&P price in [26:37] gold terms or not which I know which is [26:38] one popular way to look at it but we [26:40] think we bottomed in 10 or 11 in real [26:42] terms and so you know whatever we're 15 [26:45] years into that cycle I think the the [26:47] factors that are going to matter right [26:49] to Josh's point a minute ago 60% of 2Q [26:52] earnings 28% headline came from AI [26:55] infrastru structure. So I think we're [26:57] money good on that particular [26:59] contribution 27 20 I think we're money [27:02] good 2728 talked to Brian who I saw two [27:04] nights ago his numbers for AI capex in [27:07] 27 are 1.5 trillion the street is at [27:10] like 1.2 too. So, the street is still [27:12] low and that's been there's been a [27:13] catch- up. We all know for the last 18 [27:15] months or so. On our team, we call it [27:16] the quarterly tradition like every [27:18] quarter. You can bank on the numbers [27:20] going higher. Why? Because it's a [27:22] generational competitive uh risk among [27:25] the US players zooming out among China [27:27] as well. We can talk about it. Secondly, [27:29] on a more technical basis, the scaling [27:31] laws continue to work. Meaning every [27:33] time we train new levels or new models [27:36] of AI, new AI AI models on higher levels [27:38] of compute, the outputs and the results [27:40] continue to get better. So what is the [27:42] technical incentive for drawing down the [27:44] capex at this moment? So I agree that [27:47] >> one other thing to add to that that I've [27:48] been talking about that I think is maybe [27:50] underappreciated. It's not as sexy as [27:52] new data centers and it's not as sexy as [27:55] new GPU sales to new customers. But when [27:59] you build these data centers, you are [28:01] embedding guaranteed purchases of [28:04] servers and chips. As far as the I can [28:07] see the we could argue about the [28:09] depreciation schedule, and is a GPU [28:11] produced in 2026 a 5year asset, a [28:14] three-year asset. I don't know the [28:16] answer. I'll be I'll be the last person [28:18] that will know that answer. But the [28:20] point is, it's not a 20-year asset. [28:22] >> Correct. It's not the same as uh Toyota [28:25] building a plant that's going to make [28:26] Rav 4s for for the next 20 years. Like [28:29] you're going to need new chips all the [28:32] time. [28:33] >> And there's there's a story there for [28:34] Nvidia, [28:36] >> story there for Dell, which just had a [28:37] blowout earnings report this week. [28:39] >> Like that's a big part of the story that [28:41] I think people underappreciate. like now [28:44] that you built these data centers, even [28:46] if that slows down the pace of new [28:49] construction, we still have to feed all [28:52] of this existing infrastructure with [28:54] tons of technology. [28:55] >> So, I I think that's the most important [28:56] point that anyone's made so far on the [28:58] pot. And the reason I say that this [29:02] >> is that tennis clap. [29:03] >> Yes. [29:03] >> The reason I say that is because of the [29:06] following. We wrote a note back on June [29:08] 1st and I love writing. I wish I had [29:10] more time to write but I write you know [29:12] fairly infrequently but I wrote a note [29:14] on June 1st which I'll share with you [29:16] cautioning the semis momentum and our [29:18] takeaway was you hear all about demand [29:21] constraints to your good point I think [29:24] in under the hood in reality it's more [29:26] about deployment constraints you are [29:28] ordering servers and chips and all these [29:31] electrical components and industrial [29:32] components ahead of 40 gawatts 50 [29:36] gawatts of projected data center [29:38] construction production in the next 2 to [29:40] 3 years. Are all of those buildings [29:42] going to get built? Unless you're Elon, [29:43] are all those buildings going to get [29:45] built on time? And do you have a risk of [29:47] double ordering in the supply chain of [29:49] of semis of service? Absolutely. Is it a [29:52] risk today? No. But is a risk from year [29:54] for now? Could be. [29:55] >> This is one of the most bearish charts [29:56] that I've seen. I accidentally look at [29:57] this. I'll hold it up because it's not [29:58] in the dock. This is a chart of Sienna. [30:01] And as you know, Sienna was around [30:03] during the com bubble for the fiber [30:04] optic buildout. Straight up, straight [30:06] down. and it looks [30:09] eerily similar. Doesn't look great. [30:12] >> Yeah. And that's so that's a a perfectly [30:15] reasonable analog. I think a couple [30:16] things I would argue, right? So one is [30:19] when we look at our prime brokerage book [30:21] and we're the biggest wealth manager in [30:23] the industry and I like to I'll say the [30:25] best we're the biggest prime broker in [30:27] the industry on the institutional side. [30:28] All right. So if you look at the net and [30:30] gross exposures of our hedge fund book [30:32] and our hedge fund clients, a lot of our [30:34] hedge fund clients have not reged in [30:36] memory and a lot of the first half [30:38] winners in AI capex. They've been [30:40] waiting and I we we are paying very [30:42] close. [30:43] >> So the stocks came down and they didn't [30:45] they didn't get back into them or at [30:46] least not of the same size. [30:47] >> Correct. Not even close. And um where [30:49] have they rotated? They've rotated a tad [30:52] to Mag 7. Some somewhat of software [30:55] picking the bottom in some of those uh [30:57] areas, but it's been healthcare. It's [30:59] been some of these other parts of the [31:00] market. So, the reason I mention that [31:02] number one is number number one, I think [31:04] that's healthy. Like I think the idea [31:07] that everyone didn't jump back on the [31:09] train is a positive. They're kind of [31:11] waiting to see how things go in terms of [31:13] pricing, in terms of midterm political [31:15] football and data center. But again, I [31:18] think the healthiest thing that has [31:19] happened in terms of this cycle, Visa V [31:21] 2000 is the rotation that's happening. [31:24] And lastly, let's not also forget, as [31:26] Brian and I spoke about uh on Tuesday [31:29] night, that all the major hyperscalers [31:32] are in the early innings still of a [31:34] cloud transition. And yes, the cloud [31:37] business is being in effect supercharged [31:39] by the AI intersection today. But how [31:41] many big Fortune 500 and beyond [31:44] companies have fully transitioned to the [31:46] cloud? You're not in the ninth inning. [31:48] You're not in the sixth inning. You're [31:49] probably in the fourth or fifth inning. [31:51] So here's the key takeaway. The way this [31:53] kind of differs from 2000 is when you [31:56] laid all this fiber, you had no [31:58] alternative for the fiber. It just went [32:00] dark. And it was dark for 10 years until [32:03] Amazon became Amazon. So did it [32:05] eventually matter for creating US [32:06] exceptional exceptionalism? Yes. But not [32:10] We had to live through a 90% NASDAQ [32:12] decline on the on the way to somebody [32:15] inventing YouTube. Like we had to wait [32:17] we had to wait from first quarter of [32:19] 2000 to 2013 [32:23] >> uh I think for the NASDAQ [32:24] >> for the NASDAQ to fully make a new high. [32:27] Very great. Very good point. And um so [32:29] where I'll just put a fine point just to [32:31] end that comment is keep in mind that [32:34] taking both of our I think I think we're [32:36] on the same page wellfounded risk [32:38] factors around data center deployment [32:40] and double ordering a lot of those chips [32:43] can be reverted back to cloud and so do [32:45] we have dark GPUs the way we had dark [32:47] fiber maybe not as likely. [32:49] >> I like what you said about the duration [32:51] of the bull. So not having a bubble. So, [32:53] when Michael brings up like Nvidia [32:56] blowout numbers, a week later Broadcom [32:59] blowout numbers, why are these stocks [33:01] flat down? It makes no sense. Um, I [33:04] agree with what he's saying, but I also [33:06] like what you said. If we don't rocket [33:09] those two stocks up 70% right after [33:12] earnings, it gives it gives you more [33:15] potential upside over time. And so those [33:18] stocks can rally, [33:20] >> but we don't have to have the rise and [33:22] fall all take place inside of two weeks. [33:24] >> 100%. And I sort of like I like that [33:26] it's in slow motion. [33:27] >> Well, the glass half full version is [33:28] we're building the wall of worry. And [33:30] you need that. You need this. [33:31] >> You need the wall of worry. Thank you, [33:32] Michael. And by the way, the wall of [33:33] worry was like the Empire State Building [33:36] in 22 and 23. Remember when everyone [33:38] predicted, and I'll give kudos to Ellen [33:40] Zner, our colleague who at the time had [33:42] the um economics call. Remember that in [33:45] 2022, 2023, we've just gone through [33:47] through the most aggressive Fed cycle in [33:49] 40 years. And the consensus was [33:51] predicting this big bad recession that [33:53] never arrived. Why? Back to our earlier [33:55] conversation, we have paradigm shifted [33:58] away from goods to services, away from a [34:00] normalized income scale to a hyper [34:03] K-shaped income scale where the 10% is [34:05] driving 40% and is less elastic to [34:08] monthly changes in gas and food prices. [34:11] And and last but not least, of course, [34:13] post23, the AI super cycle. And so so my [34:16] point being is I agree with you, Josh. I [34:19] like the fact that you have and Michael, [34:21] I like the fact that you have these [34:22] clouds hanging over the kind of Uber [34:25] euphoria from happening. Number one. But [34:28] number two, I just want to come back to [34:29] I know I'm doing a little bit of a [34:30] weave, so I apologize. I just want to [34:32] come back to your comment about the [34:34] earnings growth risk because here's [34:36] where the onus on the baton being handed [34:38] off from AI infrastructure to AI [34:41] adopters is really crucial in the next [34:44] two years. Does the margin of [34:45] productivity boost from the AI adopters [34:48] more than offset what could be a [34:50] deceleration in the AI infrastructure [34:52] spending? That is the key handoff that [34:56] way. [34:56] >> It better. And by the way, let me just [34:57] say this because a lot of our retail [34:59] clients often everyone thinks and talks [35:02] and invests at unfortunately at times [35:04] with their generational bias and their [35:05] memory and a lot of our average retail [35:08] clients are still stung with the memory [35:09] of08 with 2000 etc. you know for not uh [35:13] good re for good reason and a lot of [35:15] people are talking about is this 99 over [35:17] and over again and one of the things I [35:19] would just argue is in terms of how it's [35:21] different is number one the quality of [35:24] the spenders today is so much different [35:26] from back then [35:27] >> that's such a great point [35:28] >> and yes we are talking about leverage [35:30] now but don't forget these were all mag [35:32] 7 formerly fang the industry group [35:34] formerly known as fang was was once upon [35:36] a time netcash balance sheets and so if [35:39] I'm an as an investor When I see a [35:42] generational technology investment, I [35:44] want to see more capex. I don't want to [35:47] see just buybacks for buyback's sake, [35:50] which is what they had done for 10 [35:51] years. Fang was capital light. They [35:53] generated excess cash. They bought back [35:55] stock. They did that to the tune of a [35:57] trillion. [35:58] >> People who hated the buybacks though now [35:59] hate the capex too. I don't know if [36:01] you'd be surprised by that for the [36:02] duration. So, I want to I want to um [36:04] double click on what you just said [36:05] because in 1999 [36:08] in order for Cisco to hit its growth [36:10] targets and Lucent and Dell and Sun [36:14] Micro Systemystems and Sienna and [36:16] Juniper and I can go on and on. In order [36:19] for those companies to hold up, [36:21] >> they were reliant upon selling to a [36:23] customer that had gone public a week [36:25] prior. Like literally like our [36:28] enterprise customers are Pets.com, [36:31] etoys, cd now, dj direct, and all these [36:36] things that didn't exist a year later. [36:38] >> The customers today are Amazon buying on [36:43] behalf of its 30 million cloud customers [36:46] who represent every sector of the [36:49] economy, hospitals, [36:52] >> uh insurance companies, manufacturers, [36:55] government. So it's it is not the same [36:58] as I hope we get the next 100 IPOs so we [37:02] can sell these people some some Sun [37:04] Micro products or some servers. Uh EMC [37:07] needs to sell uh you know some some [37:09] stuff. So let's hope we get another 500 [37:11] IPOs next year. That is not now the [37:15] asterisk is the two biggest players in [37:18] the ecosystem on the buy side are not [37:20] public yet. [37:21] >> Correct. [37:22] >> Not profitable. not particularly [37:24] transparent yet because they don't have [37:25] to be and not proven through any sort of [37:28] economic cycle. That's the wild card [37:31] that takes everything I just said and [37:33] invalidates it a little bit. [37:34] >> 100 I think it's really well said and I [37:36] think I'm somewhat limited as you know [37:38] on what I can say on the privates or [37:39] not. [37:39] >> Understood. [37:40] >> But what I would say in terms of what [37:41] we've publicly written about and talked [37:44] about is like the growth rates when you [37:47] track what they're doing. Some of these [37:49] companies on the privates were printing [37:50] $10 billion ARRS end of last year. Now [37:53] they're last month 60 70 billion. So the [37:56] growth rates are still astronomical and [37:58] that's what ultimately what you need is [38:00] the demand versus supply dynamic to [38:02] still be in our at our favor and we [38:04] think it is. Let me just say two things [38:06] quickly because I want to get them in. [38:07] So one is the quality of the spenders is [38:10] different. Secondly, I don't think what [38:12] gets discussed enough is the credibility [38:14] of the spenders. And this is also in [38:16] vast contrast to the Pets.com analog. [38:20] Look at I'm not I'm a fundamental [38:21] investor, but look at the technicals of [38:23] tech relative to the S&P over 30 years. [38:25] It always makes higher highs. I think [38:27] that is incredibly profound technical [38:29] signal. Why do I say that? We go through [38:32] our warts, whether it was internet bust, [38:34] whether it was '08, whether it was [38:35] COVID, whether it was 22 duration [38:37] selloff, etc., etc., etc., But we always [38:39] recover to new highs. What does that [38:42] remind me of? The US, despite some of [38:45] our political pitfalls and despite some [38:47] of our issues, of which there are [38:48] several, is still the single best [38:51] allocator of capital to new innovative [38:53] technologies and enterprises anywhere in [38:55] the world. American exceptionalism [38:57] lives. The historical analog, however, [39:00] is that while we always figure out the [39:02] next big thing and allocate to it [39:03] appropriately, we almost always boom [39:05] bust on the capex. So that was true of [39:07] internet fiber. It was true of shale [39:10] 2000 2015 rails going even back to my [39:14] blowing off my history books going back [39:15] to rails. No doubt here's where this [39:18] could be different. Once again it goes [39:20] back to our comment about Nvidia and [39:22] Broadcom are not screaming after [39:24] phenomenal earnings. I also like that. I [39:27] like the fact that the market is [39:29] signaling one that they're treating [39:32] those companies with rationality and [39:34] frankly a lot of its law of large [39:36] numbers and lastly there's other games [39:38] in town. They're new there are private [39:39] companies out there that are going to be [39:41] coming. [39:41] >> I like that argument that it's not that [39:43] they're not impressed by Nvidia. It's [39:45] that they know they're not going to do [39:47] 100% earnings growth next year but in [39:50] the meanwhile there are snowflakes out [39:52] there. like there are other companies [39:56] that do have that sort of potential that [39:59] the world is waking up to [40:00] >> that don't trade at a 5 trillion market [40:02] cap [40:02] >> and the money is shifting from one to [40:04] the next. All right. I [40:06] >> I would buy that as a as a great reason [40:08] for the rotation and for why it's so [40:10] healthy. [40:10] >> By the way, last point is 40x mutual [40:13] funds. Think about the mechanical [40:16] ceiling or the issue with a lot of those [40:18] funds have on being relatively in [40:21] absolute overweight Nvidia, right? [40:22] right? Like they get kept out at a [40:24] certain level. [40:25] >> They can't go to 10%. They can't go [40:26] >> overweight. So, who's the incremental [40:28] buyer? A lot of our wealth clients and [40:29] and frankly to their great um uh [40:32] intuition have bought the dip. When you [40:34] look at the trading uh patterns and the [40:37] statistics in the last year, 5 years, 15 [40:39] years, they've been better buyers on the [40:41] dip and a lot of them have gone back to [40:43] Nvidia. But a certain point, if you're a [40:44] retail in investor, how much of your [40:46] overall wealth can be in Nvidia? So all [40:48] of this hinges on the hyperscaler's [40:50] ability to make an ROI on all these [40:53] investments. Symbolist a former guest uh [40:56] guest of the show uh [40:57] >> I watched the program was excellent. [40:59] Yeah, Michael Michael's great. [41:00] >> Uh hypers scalers and Nvidia new [41:02] investment grade debt and SPVS. It's 200 [41:04] 200 uh $320 billion in 2026. [41:07] >> This is from his note. [41:08] >> So he he argues that perhaps there's so [41:11] much money. [41:11] >> I mean this is dramatic. [41:13] >> There's so much money here that this is [41:14] help pushing up [41:15] >> government bond yields. [41:17] Um, is this going to work? [41:22] >> So, at the moment, the demand we're [41:24] seeing and even the terms and the [41:26] spreads we're seeing for a lot of this [41:28] issuance is still relatively benign. And [41:30] again, it goes back to my point that [41:32] these the names you all cite on this [41:33] chart were all net cash balance sheets. [41:36] And is there dispersion between the [41:38] oracles of the world and the Microsofts? [41:40] Absolutely. Microsoft and J&J are the [41:42] only two triple rated uh AAA rated names [41:44] in the market left. Oracle CDS spreads [41:47] trade at you know very wide levels. So [41:50] is the market in my opinion visav that [41:53] example pricing in some of the risk in a [41:55] dispersed manner as it should? Yes. So [41:57] the market is fully aware of this but [41:59] Michael you made the point earlier in [42:01] terms of just how much liquidity is out [42:03] there and what types I'll also add what [42:06] types of new buyers are out there [42:08] including the insurance community which [42:10] I think has been another source of major [42:12] demand for this paper. [42:13] >> That's such a good point getting back to [42:14] the point I made earlier. God forbid the [42:16] the price on these come down, the yields [42:17] the yields go up. Uh title wave of money [42:20] waiting to buy these. [42:21] >> Agreed. Agreed. [42:22] >> The world wants this paper, otherwise it [42:24] wouldn't be issued. [42:25] >> Absolutely. [42:26] >> And it's not being issued out of [42:27] desperation. They're calling up Alphabet [42:29] and saying, "Hey guys, like there's an [42:31] opportunity here, you know, given your [42:33] credit rating and your cash flows, [42:36] there's an opportunity to do something [42:37] that's potentially better than equity [42:39] financing if you don't want to just keep [42:41] doing they're going to spend anyway." [42:43] Yeah. So, okay. Um, how do you handle [42:45] people asking about the circular [42:48] financing question? Because that's I [42:50] would say five days a week on CNBC that [42:54] that's being debated. It's not going [42:55] away. [42:56] >> Um, there are some great answers for it. [42:58] I'd love to hear your answer for it. So [43:01] if you listen to Jensen's comment on the [43:03] October call of 25 on that same um risk [43:07] or that same feedback, it was, hey, we [43:10] have a really unique line of sight in [43:11] terms of some of our supply chains and [43:13] some of our partners' future growth. And [43:16] he's talking about the private labs. And [43:17] he said, "Given that line of sight, [43:20] given our net cash balance sheet, as an [43:22] investor, wouldn't you want me to take [43:24] some type of uh leverage or take some [43:27] type of skin in the game in terms of [43:28] some of our partners?" And frankly, I [43:31] thought it was a really convincing [43:32] statement. [43:33] >> Wouldn't who want him? His own [43:34] shareholders, [43:35] >> his own balance sheet [43:36] >> because I know the shareholders of these [43:37] other things love him doing that. [43:39] >> No doubt. No doubt. And and look, to [43:41] your earlier point, does it create a [43:43] risk factor? Absolutely. Um, and so you [43:46] know what I would say is I think it goes [43:49] back to you it's a it's a more of a [43:51] qualitative answer but it goes back to [43:52] my comment about 2000 to 2026. Look at [43:56] how all these companies have managed [43:57] through every technology wave whether it [44:00] was internet yes social um e-commerce [44:04] online ads streaming entertainment cloud [44:07] now AI and oh by the way the next round [44:09] which doesn't get talked about enough [44:10] Adam Jonas talks about it quite a bit [44:12] robotics space quantum and autonomous [44:15] and by the way the MAG7's going to [44:17] dominate in a lot of those sectors as [44:19] well right and so my opinion is [44:22] >> he's saying we have a really unique line [44:24] of sight in terms of how all these [44:25] businesses are um being allocated, how [44:29] they're being in allocated in the [44:30] enterprise, I meant to say, and we have [44:32] the net cash balance sheet. So, wouldn't [44:34] as an uh shareholder of our stock, [44:36] wouldn't you want us to take that [44:37] opportunity? [44:37] >> So, vendor financing was a big issue [44:40] during the boom bust and [44:42] >> people remember like this is one uhoh [44:45] when the equipment suppliers are giving [44:48] money to the buyer who's then going to [44:50] buy their equipment. that's usually [44:53] closer to an end than a beginning or [44:55] that's what that's the thing that people [44:56] are worried about. [44:57] >> So watch the spreads, right? So the debt [44:59] market just like in 98.99 when you had [45:01] the Fed raising and you started to see [45:04] the bond spreads on the tech name [45:06] starting to widen out. We're watching [45:07] that too, right? And so look, I mean [45:10] that was go back to the fall of 25. [45:12] Oracle CDS really was the uh predictor [45:16] of Oracle stock in the coming year. [45:18] >> Dan, this is weird. So we're looking at [45:21] in red is the high yield corporate bond [45:24] spread. Okay. [45:24] >> Mhm. [45:25] >> And in black or gray is the triple C. So [45:29] really junky junk. And they go in the [45:31] same direction almost all the time. And [45:33] there's been a very very notable [45:36] divergence where high yield corporate [45:38] spreads haven't budged. Good. But the [45:40] shittier stuff is ticking up in a pretty [45:44] meaningful way. [45:45] >> What's in trip? What's in [45:46] >> I was just gonna say is that like [45:48] private private equity private equity [45:50] backed companies [45:51] >> even even worse rated stuff. [45:53] >> So take well take it with a grain of [45:54] salt from the equity guy but my [45:56] extrapolation of this data series is [45:59] when I look at the underlyings of high [46:01] yield I look at energy I look at [46:03] materials I look at chemicals I look at [46:06] um industrials. One a lot of those [46:09] industries are benefiting fundamentally [46:11] from the geopolitical issue in the [46:12] Middle East right now. Second, if you [46:15] just stick with energy for a minute, [46:16] right, pre the 2015 shale bust, capex [46:20] bust, uh, a lot of executive [46:23] compensation frameworks were [46:24] incentivized towards production growth. [46:26] So, commodity prices would scream [46:28] higher. What would all the EMPS do? They [46:30] put out extra production growth. What [46:31] changed dramatically post the 15 [46:33] implosion? A lot of the executive [46:35] compensation structures today are more [46:37] balanced in terms of capital return. And [46:39] so, what have a lot of energy names [46:40] become today? They've become higher [46:42] dividend payers. they become higher [46:44] buybacks. Secondly, let's go downstream [46:46] to materials and some of the industrials [46:49] components. Okay. Well, we just talked [46:50] about the AI data center buildout. [46:52] They're fundamentally build benefiting [46:54] from that buildout. So maybe there's [46:56] more of an idiosyncratic [46:58] >> steel copper like you name. [47:01] >> I feel like this series is diverging [47:03] because in part because of AI and [47:05] because of some of the compositional and [47:07] fundamental changes in the energy [47:09] sector. [47:10] >> I think that's right. Especially on the [47:11] energy side. We were talking earlier [47:12] about the wall of worry and how there [47:15] does seem to be a persistent level of [47:17] disbelief which in the short term is [47:20] maybe something to pay attention to but [47:21] longer term that's healthy for the [47:23] continued secular bull market. John [47:25] chart six please. So this is from [47:27] Goldman leverage funds versus the NASDAQ [47:29] 100. [47:29] >> Never heard of them. [47:31] >> Bears are piling into NASDAQ 100 [47:34] futures. Short positions have surged 35% [47:37] since midJune and are now near record [47:39] levels. So, I want to share another [47:41] chart that we made, John, chart 7. So, [47:44] chart kid Matt made this and we're [47:47] looking at the median S&P 500 stock [47:50] short interest as a percentage of market [47:51] cap going up and to the right in a [47:54] pretty steep way. [47:55] >> Uh, and on the right side, you have the [47:57] bottom desile. So, these are the ones [47:59] that are least shorted and even the ones [48:00] that are the least shorted are going up [48:02] in a material way. Is there anything [48:05] going on in here? Like my first reaction [48:07] was well maybe dismiss this because [48:09] there's a lot of other different hedging [48:11] >> charts are 30-year charts. There's [48:13] something going on [48:14] >> but they're moving. [48:16] >> Wait, did you just get bearish now? [48:19] >> I just got very pensive. [48:20] >> So let me let me let me So I've got [48:21] >> I have to take this all in. [48:22] >> This is a lot. So while you're thinking [48:24] I've got the stocks that are [48:25] >> This is not a squawk box exercise. Hold [48:27] on. [48:27] >> You could you could think. So the stocks [48:29] that are in the B. So the least shorted [48:31] stocks make sense. [48:33] >> That this 99th percentile. Yes. So, [48:36] what's in that basket is [48:39] Google [48:39] >> probably Apple. [48:40] >> Google and Bergkshire because who the [48:41] hell is short Birkshshire? Morgan [48:43] Stanley number three. Let's hear it. [48:46] Morgan Stanley number three, General [48:47] Dynamics Dan Skelly, [48:49] >> Walmart, Amazon, Wells Fargo, Eli Lilly, [48:51] Apple, AET, and Chevron. [48:53] >> So, these are stocks that nobody wants [48:54] to short. And even them, it's at the [48:57] highest it's been since [48:59] >> this be mechanical people hedging even [49:01] bigger long positions. So the only the [49:04] only way this is fascinating and I I'll [49:06] come back to you with a fuller answer [49:07] because I want to study this. Well well [49:09] well Chris Metley is our leader on the [49:12] quantitative derivative strategies desk [49:14] in ID Boston's been around forever. [49:15] Phenomenal [49:16] >> do you say this is my quant [49:17] >> this is my quant. [49:19] >> All right [49:20] >> um from Boston um [49:23] >> here's the only thing I could say [49:24] because I'm just still focused on this [49:26] left hand side chart on the median [49:28] stock. Remember that as of late second [49:32] half 25 [49:34] until June 22nd when the semis locally [49:37] topped is we had this really [49:40] hyperconentrated market in AI capex. So [49:42] let's year to date through June we had [49:45] 90% of S&P's attribution come from three [49:48] industry subgroups semis IT hardware and [49:51] power all tethered to data centers and [49:53] so if I think about like what's going on [49:54] the lefth hand side of the chart simply [49:57] put is I think you have people rotating [50:01] out of those names and also at and also [50:05] shorting the median stock as a counter [50:07] to that [50:08] >> there's interesting happening it's [50:09] hard to so all right the top desk so the [50:11] most shorted By the way, look at the y- [50:12] axis. Obviously, the bottom decile, [50:14] nobody's short. Okay, it's it's at an [50:16] all-time high, but it's 1.2%. If you [50:19] look at the top decile, so the most [50:20] shorted, it's 9 9.8%, which is in the [50:23] 70th percentile. And these are the [50:25] names, the most shorted stocks. [50:26] >> Reddit, 13%. I don't even know a lot of [50:29] these names. KMBB, I [50:31] >> helpful to see this. [50:32] >> Okay. Charter, SMCI, I don't know. I [50:35] don't know all these names. Uh, Trade [50:37] Desk, I know. So I know this feels like [50:40] this feels like a rates move. So like [50:42] the curve we were priced for three rate [50:44] cuts from Jan till Iran March April and [50:48] now that's completely flipped and I [50:49] don't I I can't I didn't see what [50:51] happened this morning but I think we're [50:52] now 50/50 price for this month in terms [50:54] of rate. So if I look at this cohort on [50:57] the left hand side it's super low [50:58] quality less balance sheet strength. On [51:01] the right hand side it's much higher [51:03] quality. So to me it feels like the [51:05] short interest in the lower quality [51:07] things rerated as rates have rerated the [51:09] debt to equity ratio feels like profit [51:12] margins cash flow [51:13] >> there's idiosyncratic stories in here [51:15] that like the trade desk has been an [51:18] abomination in a very good tech sector [51:21] not representative of anything else um [51:24] SMCI is accounting scandals uh charter [51:28] is cable like [51:30] >> Reddit has the Google algorithm change [51:32] Right. Reddit has to contend with the [51:34] source of all its traffic changing how [51:36] it sends people. Like you can uh Echko [51:39] is competing with SpaceX. Have fun with [51:41] that. [51:42] >> Like you can go through this and come up [51:44] with uh a reason why this is not a [51:47] market story. It's an idiosyncratic [51:49] story. [51:49] >> So Dan, you'll like this. The only the [51:51] only not to interrupt, Michael, the only [51:53] uh outlier on the right hand side, [51:55] Charles, I'm not sure why Microsoft [51:57] wouldn't be there [51:58] >> because I think what was really crucial [52:00] to their narrative in the last month was [52:03] how they tempered the capex comments on [52:05] the earnings call a month ago and the [52:07] stock reacted so positively towards that [52:10] and it feels like people are now [52:11] siphoning out within the group maybe [52:14] Microsoft, Amazon, a Apple's in a [52:17] different category, but they are in a [52:18] different group at the moment than Meta. [52:20] Oracle. [52:20] >> It's funny that you say that. I don't [52:22] think that's at all behind the rally in [52:24] Microsoft. I think what people really [52:25] liked was the public divorce with Sam [52:28] Alman [52:29] >> and the end of exclusivity with Open AI. [52:31] >> I think people felt that Microsoft was [52:33] not getting the best end of that deal [52:35] >> and that they might be better off using [52:37] cheaper openweight models to fulfill [52:40] those AI software product commitments to [52:42] their users. [52:43] >> And I think that that's probably where [52:46] Microsoft gets its mojo back. Some [52:48] combination. [52:48] >> Absolutely. No, that's a great point. I [52:50] I think they went 180 from being tip of [52:53] the spear investing in front to now [52:55] being more rational. I think that's a [52:57] great point. [52:57] >> I want to introduce two charts to uh [53:00] present evidence that the stock market [53:02] is functioning pretty well. I think [53:04] fundamentally I think it's responding [53:06] appropriately to stocks that are doing [53:08] well and stocks that are doing less [53:09] well. John, let's start with chart four. [53:12] So, this is uh this is a CNBC chart. [53:14] threemonth intra stock correlations is [53:16] that is at basically as low as it's been [53:19] in almost 40 years. Like each stock is [53:22] marching to the beat of its own drum. [53:23] >> But Michael, I thought index funds were [53:25] ruining the market. Is that not true? [53:28] So, Adam Parker, your friend, our [53:30] friend, uh, John, let's jump to chart [53:32] nine, has an awesome, awesome, awesome [53:34] chart that shows the mean 12-month [53:37] industry group relative return of [53:40] companies with year-over-year growth [53:42] margin contractions of more than 1%. [53:44] Okay? And these stocks are getting the [53:46] kicked out of them relative to the [53:48] index. And what he's saying is basically [53:49] if your gross margins are missing, you [53:52] are in big big big trouble. M [53:54] >> and we're seeing that over and over and [53:56] over again with these earnings reports. [53:58] >> Oh, Broadcom today, by the way, their [54:00] gross margins fell 200 200 basis points. [54:02] Blowout numbers, yeah, their margins [54:03] weren't great and the stock was down 25% [54:05] going to the print. [54:06] >> Crucial crucial point, Michael. So, let [54:08] me jump in on two things. I think the [54:11] reason why this particular vector is [54:13] compounded at the moment. One goes back [54:16] to what I said earlier in terms of this [54:18] synthetic tariff related cloud in first [54:21] half of of last year being now [54:23] synthetically boosted by pricing. So if [54:26] you're not getting the gross margin [54:28] benefit from that pricing right now, [54:30] you're in this camp. And secondly, maybe [54:32] a little bit of this on the gross margin [54:34] line even is on the AI productivity as [54:36] well. And so like if your peer group is [54:38] experiencing some modicum of AI [54:41] adoption, productivity and you're not, I [54:43] think that's why you're getting triple [54:44] dinged on this particular data series. [54:46] >> Okay, you know what's in this in this [54:49] the stocks that are missing on gross [54:51] margins? It's a lot of food service. [54:54] >> Okay, [54:54] >> it's a lot of QSR restaurants. It's your [54:57] Shake Shacks, your Chipotles, [54:59] >> the K working in reverse, which is a [55:01] thing now. [55:01] >> It it's just right. It's just not a It's [55:04] not a lot of fun to be in that business [55:05] right now. Like I Shake Shack is not [55:08] supplying French fries to data centers. [55:10] You know what I mean? Like they have [55:11] less to look forward to. They had an [55:13] earnings blow up a couple of months [55:14] back. They blamed [55:16] >> paper goods, the price of beef, and it [55:18] was all Iran war oil related, whatever. [55:22] >> But they don't have that offsetting. [55:24] Yeah, but look how much money we're [55:26] making from AI. like they don't [55:28] >> there's probably a thousand stocks like [55:29] that in in you know in the US market of [55:32] 3,800 stocks [55:34] >> they're small [55:35] >> they don't actually matter to the return [55:37] to the index [55:38] >> it's 100% and there it's helpful Josh to [55:40] your point to see it through the lens of [55:42] which business models because like let's [55:43] just drill down for a minute on um food [55:45] staples which are going through two [55:48] massive uh headwinds at the moment [55:50] cyclically you've got the backup in [55:52] rates and the volatility mostly to the [55:54] higher um side of the chart on oil. So [55:58] the K shape which has been incredibly [56:00] resilient on the lower rung. I mean we [56:01] talk about everyone focus on the upper [56:03] run. Let's talk about the lower rung for [56:04] a minute which I mean I could argue and [56:06] I think there's been this debate around [56:08] does the official data capture all of [56:10] the gig economy and all of the ways [56:13] people are making money off balance [56:15] sheets [56:15] >> because the data come from comes from [56:17] the 1950s. We didn't have that [56:18] >> right. We could use an update there. [56:20] Yeah. So I think on the one hand the [56:21] data is skewed but on the other hand [56:23] let's face it like the middle lower [56:25] income cohort has been I think more [56:27] resilient than most has most have [56:29] perceived or anticipated until now. I [56:32] think now you are seeing some [56:34] degradation on the middle income worker [56:37] maybe visavi in the beginning stages but [56:40] just the macro factors and then just [56:42] drilling down into food staples because [56:44] I as a generalist I can um talk to all [56:47] of these sector experts. I think the GLP [56:50] thing is real. I think when you look at [56:51] Pepsi going back five ten years uh five [56:54] years ago, I think they were incredibly [56:56] convicted that they were going to sail [56:58] through the issue. And when you look at [57:00] Fredo Lelay and some of the food [57:02] businesses missing their organic revenue [57:04] growth targets for like 2 to 3 years [57:06] running, I think it's a very So the [57:08] mistake very is to look the mistake is [57:11] to look at McDonald's earnings and try [57:14] to extrapolate something about the [57:15] economy cuz it's just not the way. [57:19] >> Look at the stock of Hershey. Horrible. [57:21] That's years. Now Dan, I am a salt of [57:23] the earth person and I am equally likely [57:25] to be shopping at Americana Manhattan [57:28] >> as I am to be at Roosevelt Field and [57:31] Michael and I have been harping on this [57:32] for like 3 years. Every time the banks [57:36] report earnings, um, less so Morgan [57:39] Stanley, the banks that issue credit [57:40] cards, it's almost like the audience is [57:43] clamoring for them to say the bottom [57:47] desile, the bottom quintile is breaking. [57:50] And now they don't even wait for the [57:52] question anymore. Like JP Morgan, they [57:55] lead with it. They're like before anyone [57:57] asks us, [57:58] >> everyone's paying their bills. [58:00] Delinquencies are well within norms and [58:03] at historic lows and car loans are being [58:06] paid off. And just anecdotally, like who [58:09] is filling up all these flights? The [58:12] movie theaters are packed again. [58:14] >> Like you walk through the mall, it [58:15] doesn't matter which mall, there are [58:17] people in it. So, even if you don't know [58:19] the data, if you're a normal person and [58:22] you talk to the people in your life and [58:24] you look around and see what they're [58:26] doing, they're leasing new cars, they're [58:28] going to work, they're paying their [58:30] bills, [58:32] >> and then for some reason there's like [58:33] this internet component that is [58:35] insisting that it's all fake and nobody [58:38] is paying their bills and it's about to [58:40] collapse. And that's the hardest part [58:42] for me cuz the arguments are so [58:46] seductive [58:47] >> that we're at some sort of peak [58:49] something and it's being hollowed out [58:51] from the bottom and you up there on Wall [58:53] Street, you don't see it yet. Have fun [58:56] with your Micron earnings, but you'll [58:58] see it soon. All when [59:00] >> I keep waiting. Do Do you feel that that [59:03] maybe we get to a point where we're [59:05] ignoring that risk too much or am I [59:08] ignoring it just about the right amount? [59:10] It's really an important comment and I [59:13] think super well said and I'll say a [59:15] couple things. First is number one I [59:18] think it goes back to my statement which [59:20] is unproven at the moment a minute ago [59:22] about the gig economy. It goes back to [59:24] co which you know a crisis doesn't [59:27] usually create new trends but it [59:29] certainly can accelerate things that [59:30] were already in place and I think the [59:32] whatever you want to call it social [59:34] media influencer economy visav covid got [59:37] supercharged and it's still going [59:38] forward. starting businesses, [59:40] >> new business, I was going there next. [59:42] >> The entrepreneurial sphere. [59:43] >> Absolutely. I was going there right next [59:45] so we're on the same page. So new [59:47] business formations that are two decade [59:48] high is part of that visav AI. I [59:51] absolutely think so. AI in many ways is [59:54] lower lowering the barriers to entry to [59:56] being an entrepreneur if you have a good [59:57] idea. Now does everyone have a good [59:59] idea? No. But do I have more scale in [60:01] advertising and marketing and process [60:03] and technology if I have AI? Yes. And so [60:05] what are those new businesses doing, [60:06] Josh? They're hiring people. And maybe [60:08] it's not all counted on the official [60:10] monthly payrolls, but they're hiring [60:12] people. And and so my point is I I know [60:15] your comment about the big bank's [60:17] earnings transcript so well because we [60:18] we see the same thing. And it's true [60:21] like the data is too diverse and too [60:24] rich and too widespread to try to poke [60:26] holes in it, right? It's it's travel, [60:28] it's eating out, it's all these [60:30] different parts of the economy. Yeah. [60:31] And so in my opinion, it's not a [60:34] >> kind of one-off phenomena. There's [60:35] something underpinning it. And I think [60:37] these these comments are underpinning [60:38] it. [60:38] >> Don't you think one day there will be a [60:41] Bank of America or a Capital One [60:43] financial conference call where they [60:46] actually do say, "Okay, we're seeing an [60:49] uptick in delinquencies or bills passed [60:52] due like 30 days past due, whatever." [60:55] But by the time they do that, we will [60:57] all be like, "Yeah, no duh." [60:58] >> I was just going to say, "Look at the [60:59] earning." Won't it be obvious to all of [61:01] us, [61:01] >> right? How unlikely would it be that [61:04] earnings prior to that moment hold up [61:07] >> how unlikely would it be that earnings [61:08] continue to trend quarter over quarter [61:10] not year-over-year in the short term how [61:12] unlikely would it be that they continue [61:14] to to accelerate very unlikely so so [61:16] we've talked a lot about idiosyncratic [61:19] nuances which I think today this cycle [61:21] is unlike any other cycle it's been a uh [61:23] underlying thread of all of our comments [61:25] but to be fair economics 101 just goes [61:28] back to earnings and and job growth and [61:30] it's very unlikely in this earnings [61:32] picture to have that future BFA moment. [61:34] >> The lower shape of the K. I was looking [61:36] at a firm's delinquency rate. Which [61:39] customer do they serve? [61:41] >> A firm. [61:41] >> It's not AMX, right? Buy now pay later. [61:44] >> Yeah. So lower FICO 20. Yeah. [61:46] >> The delinquency rate is nowhere. It's [61:49] nowhere. It's 2.3%. Whatever it is, [61:51] >> because that's the pay later part. You [61:52] don't understand. You just don't [61:54] understand economics. [61:55] >> I'll add to that. I don't know if it's [61:56] >> They say pay later. You don't know how [61:58] much later. I don't know if it's GLPS at [62:00] the um 25year-old cohort level, but what [62:03] I do understand from a lot of our [62:04] clients and our partners anecdotal [62:07] comments about their children and the [62:08] next generation is that people aren't [62:10] going out as much maybe to bars, maybe [62:13] drinking, but they are gambling, they [62:15] are making wages, uh wages, and maybe on [62:18] average they're doing better than we [62:20] would have anticipated. So maybe they [62:21] have this excess capital from this [62:24] proliferation in the gamification of [62:28] betting and gambling, all these other [62:29] things going on. [62:30] >> So instead of overeing, they're betting. [62:32] >> They're over betting. [62:33] >> So it's so funny like the next [62:34] generation, they're just going to do the [62:35] GLP1 right into the right into the womb. [62:39] Like in vitro g I'm going to be the last [62:41] person left in this world. [62:42] >> Yeah. Less of a copay. [62:43] >> Okay. Uh all right. I want to I want to [62:46] make sure we get to a couple more [62:47] things. stock selection. [62:50] You're managing money. You're also [62:52] advising other people who manage money. [62:55] You're on the committee. You're in all [62:57] these conversations. [62:59] What are you telling people about the [63:01] end of year run? Hopefully, it's a run [63:04] >> and into 27. [63:06] >> Like I don't know if it's your opinion, [63:07] the house opinion. I don't know where [63:09] one ends where, but talk about our our [63:11] audience are investors, [63:13] >> right? They love they love the stock [63:15] market. people listen to compounded [63:17] friends love the stock market. So talk [63:18] about stock selection, what you guys are [63:20] excited about, what you're looking at, [63:22] and maybe what you're uh leaving behind [63:25] from the first half of 26. [63:27] >> Yeah, all the above. So let me bifurcate [63:30] it between my asset allocation hat and [63:32] my day-to-day PM equity hat. So on the [63:34] asset allocation front, uh Josh, we've [63:36] been overweight US equities for well [63:38] over a year. And I would say while that [63:41] has certainly served us well, we've also [63:43] no doubt experienced this AI spending AI [63:47] highly correlated concentrated [63:48] phenomenas as well. And so at the asset [63:51] allocation level, we've also told people [63:53] allocate towards real assets, allocate [63:55] towards long short hedge funds, [63:56] >> counterbalance, [63:57] >> no doubt. And um and you know in some [63:59] instances allocate towards EM and we saw [64:02] what happened when AI caught a cold this [64:04] summer. We saw what happened in the US. [64:06] Rotation was the outcome. Peak to trough [64:09] draw down. The S&P was 3%. Because of [64:11] that concept I alluded to earlier. Semis [64:14] went to MAG 7. And mathematically MAG 7 [64:17] might be 3540% of the market. Semis are [64:20] much bigger, but they're 18%. If I add [64:22] in healthc care at 10% and financials at [64:25] 15%. That's what really offset that um [64:29] cold metastasizing for the US. Not so [64:32] for the Cosby. So the Cosby had truly [64:35] concentrated to stock. So we've told [64:38] people to be selective in EM. So what [64:40] I'm trying to outline Josh at the broad [64:42] level is asset allocation. We've been [64:44] overweight US and are sticking to that [64:46] for a lot of the reasons we've talked [64:47] about resilient economy um accelerating [64:50] earnings growth, but we've also tried to [64:52] plug and play through diversified asset [64:54] classes. On the stock level, again, the [64:57] healthiest thing I would say because I'm [64:59] most focused on long-term investing, you [65:01] know, my team originated in 1995 under [65:04] Byron Ween uh in the equity research [65:07] department. and and Byron who is truly a [65:10] visionary thought of investing in a [65:12] concentrated portfolio along a [65:14] continuum. And we run 40 to 50 stock [65:17] portfolios. And the idea being in a 50 [65:20] stock portfolio, you want to have the [65:22] majority of your names, 40 names, 45 [65:24] names be core holding, plus or minus the [65:27] market in any given year. [65:28] >> They're going to be two-ish percent. [65:30] >> Two-ish percent. Correct. So, not super [65:32] uh big absolute sized, but I also want [65:35] to build in maybe five, six positions [65:37] that are big ideas. And this was truly [65:39] Byron's philosophy. If I can get big [65:42] ideas right over a cycle, be it Nvidia [65:45] 10 years ago, Visa, Mastercard and [65:48] Processing, be it thermoffisher and the [65:50] proliferation of life science. If I can [65:52] get a lot of the big ideas right over a [65:54] cycle, they can now offset the average [65:56] over a longer period of time. So what [65:58] we're telling clients today through our [66:00] portfolio [66:01] >> I like that approach. [66:01] >> What we're telling clients today through [66:03] our portfolio process to to conclude is [66:06] yes you wanted to be cautious AI and [66:08] semis in June. I think there's a lot [66:11] more uh interesting ideas there today, [66:14] but coming back to our overall thesis, [66:16] this adoption wave is going to be a [66:18] decade long. And so we want to own a lot [66:20] of those Fortune 100 companies that are [66:22] quality oriented. And I know it's a [66:25] off-sighted cliche to be quality [66:27] oriented. Quality has not worked this [66:29] year. It has not worked in several [66:31] years. But when I think about the [66:32] rotation as of the last three months, [66:35] what's working again? Again, kudos to [66:36] Mike Wilson. quality has been coming [66:38] back in vogue. Rates are higher. The [66:41] onus on whether or not you can make [66:43] money and out compete your peer set on [66:44] AI adoption is higher. And I think this [66:47] really augers for a quality diversified [66:49] equity portfolio across healthcare. No [66:52] doubt AI in many instances is taking a [66:55] six to sevenyear timeline in terms of [66:57] phase one drug discovery to like less [66:59] than 12 months. One of the things I'm [67:01] most excited about um is seeing that [67:03] play out and seeing drugs come safer [67:07] drugs come to market faster I think is a [67:10] phenomenal out outgrowth of all this [67:12] spend [67:13] >> and and and the downstream winners [67:15] beyond that the life science tools the [67:17] consumables the lab uh and clinical [67:19] trial managers keep in mind healthcare [67:21] couldn't comp the COVID comp then it had [67:23] to deal with the 22 rate cycle then it [67:25] had to deal with everyone who had growth [67:27] equity capital going to tech so [67:29] healthare used to have a growth equity [67:31] capital. So that's one sector we like on [67:33] the AI adoption theme, capital markets. [67:35] But are you top down on this 50 stock [67:37] portfolio? You saying we're really [67:39] bullish on healthcare, so let's find [67:41] healthcare stocks that fit our thesis or [67:44] are you arriving at that top down [67:46] because from the bottom up, these are [67:47] the best earnings growers. How are you [67:50] doing that? [67:50] >> It's more of the latter. For 30 years in [67:52] a 50 stock context, we've been bottom up [67:54] focused. So do we make sector calls? [67:57] Yes. But our sector calls are are modest [67:59] in nature, two to four points over [68:01] underweight any given sector as a way of [68:04] controlling tracking error and risk. So [68:06] to your good question Josh, if we like a [68:08] stock from the bottomup thesis, we might [68:11] overwrite the sector call. But at the [68:13] same time, as our old strategist and our [68:15] friend Henry McVey has always said, who [68:17] started out as a financials analyst now [68:19] at KKR, great friend of Morgan Stanley, [68:21] uh you got to have a view. And so we [68:24] think you you no doubt have to marry the [68:25] bottomup selection effect with the [68:27] sector calls. Are there scenarios where [68:30] you did get you got the sector right? [68:33] You got the you thought you got the [68:35] stock right, but then there's an [68:37] execution misfire on the part of [68:39] management and you say we're still [68:41] bullish on this theme. We're riding the [68:43] wrong horse. You must have to do that [68:45] all the time. [68:45] >> Let's talk about software. [68:46] >> How hard is that? [68:47] >> Oh, that's the hardest part. That's kind [68:48] of the hardest part. Hardest part. [68:50] hardest part because in a you know going [68:52] back to the process [68:53] >> right horse wrong jockey or right race [68:55] wrong horse or whatever [68:56] >> no doubt I was supposed to get on a [68:58] flight to the derby and I ended up at [69:00] Belmont yeah [69:01] >> um so what I would argue is in a 50 [69:04] stock portfolio with a 2 to 3% tracking [69:07] error which is our process um you live [69:10] and die by concentration and so we try [69:12] to temper the sector bets as a function [69:14] of controlling that risk factor but the [69:16] stock selection effect drives really [69:18] everything at the end of the day and So [69:19] you asked the right question which was [69:21] got the sector call right or wrong. What [69:23] happened to the stock selection? No [69:25] doubt software has been I think the the [69:27] most acute realization of that question [69:29] in the last seven or eight months. So I [69:31] told you earlier I don't write often but [69:34] when I do I've actually been fairly um [69:36] you know directionally correct this [69:37] year. I cautioned on semi semi June 1st. [69:40] I said software is oversold in [69:42] midFebruary. And my conclusion in [69:44] software literally [69:45] >> my conclusion in software it was at the [69:48] time if you look at IGV [69:51] >> the CL the new claw plugin had just come [69:53] out in midFeb and the whole sector was [69:55] being priced for obsolescence freaked [69:57] everybody out. My statement was [69:59] >> I think it's very unlikely that as [70:01] private entities the business plan for [70:03] these labs pitching Wall Street future [70:05] investors is our our plan over the long [70:07] term is to put a lot of corporate [70:08] America out of business. I think that [70:10] would be a very uh futile business plan [70:13] to the street. More likely in my opinion [70:15] back in February was forget about [70:17] obsolescence. You're going to see [70:19] chronic interdependence between AI and [70:21] software. What's going to be most likely [70:23] is dispersion of outcomes across the [70:25] spectrum. And so I'd like to say across [70:28] the holdings we had, we had 100% hit [70:30] rate across that outcome. We were right [70:32] on the concept, but there's been so much [70:35] variation in terms of the [70:36] >> like if you take the bet on Adobe based [70:39] on what you just said versus take the [70:41] bet on Crowd Strike, it's like, okay, we [70:44] were right. Software was oversold, but [70:46] we didn't get the full benefit. And [70:48] that's got to be when you're doing 50 [70:50] stocks, that's got to be the thing [70:51] that's like, oh man, I hope this is the [70:53] right one. [70:53] >> 100%. And look, we had [70:55] >> it's it's hard for everybody. [70:56] >> We had puts and takes for us. [70:58] Microsoft's rebound, Service Now's [71:00] rebound. Uh we were out of Adobe over a [71:03] year ago. We were in PaloAlto as of the [71:05] last two years, but it got to 70 times. [71:07] And in the last two or three months, we [71:09] said it's the top end of the PE range. [71:11] The market has correctly discounted. [71:13] It's an AI winner, but we run a 50 stock [71:16] portfolio at 70 times. It was a high [71:18] tracking error winner, so let's take it [71:20] out. [71:21] >> What's the stock in that portfolio that [71:23] you are most convinced the market [71:24] misunderstands [71:25] >> in the software space or just overall? [71:27] >> Just period. Like what? I have mine in [71:29] my own portfolio. I think the market [71:30] Uber I think the market is Uber wrong [71:33] personally. Um you look like you're [71:35] short. [71:36] >> You look like you're bearish. [71:36] >> I'm long only. So So I've been in and [71:38] out of Uber over the years. So let me [71:40] let me put this. [71:40] >> What's What's your version of that where [71:42] I'm saying to people I give you two [71:43] names? [71:43] >> Yeah. No, give me as many as you want. [71:44] >> I'll give you two names. So [71:47] >> back in go back to February. So it [71:50] seemed like every other week a big [71:52] liquid part of the market was being [71:54] priced for absolescent. We talked about [71:55] software at one point. It was insurance [71:57] brokerage. at one point wealth [71:59] management CB [72:00] >> Michael we're also on the same page [72:02] >> spgi so [72:03] >> morning star [72:04] >> that's where I was going to go the idea [72:06] that data forget about software at the [72:08] moment the idea that datacentric [72:10] businesses were also being priced for [72:13] obsolescence and in many instances have [72:15] not rebounded [72:16] >> so S&P as an example in one of our full [72:19] disclosure one of our holdings does [72:20] multiple things it has the issuance [72:22] business but it has the data business [72:24] what they just started talking about [72:25] potential alternatives [72:28] But I don't think the issuance business [72:30] is frankly getting enough bid in a world [72:32] of going back to your debt charts in a [72:35] world of massive financing. [72:36] >> Yes, bond investors need ratings. They [72:38] don't they're not going to take Claude's [72:40] word for it. [72:40] >> Thank you. [72:41] >> Regulated. It's regulated. They have to [72:44] be rated. [72:44] >> Regulated. Go back to healthcare. I'm [72:46] with you. So So it's highly regulated [72:48] dataentric businesses. And last one is [72:51] NASDAQ because when I think about that [72:53] business, I have these future franchise [72:56] IPOs on the horizon. [72:57] >> It's at highs. [72:58] >> NASDAQ for the year, it might be at [73:00] highs, but for the year pointtooint is [73:02] flattish. The last time I looked at it, [73:04] you should be at more. The earnings [73:07] continue to grow mid- teens to high [73:09] teens. The multiple has stayed rarely [73:12] flattish. And to me, again, it was [73:15] caught up in that February March selloff [73:17] of datas going out of business. So [73:18] NASDAQ is three businesses. They have a [73:21] small fintech business. [73:22] >> They have a data business, which is the [73:24] crown jewel. And then they have the [73:25] exchange business, which it is what it [73:27] is. [73:27] >> All three of those things feed each [73:29] other. So they're not really three [73:31] distinct businesses, but that's how they [73:32] report, [73:33] >> right? [73:33] >> I remember thinking NASDAQ is the bet [73:36] because all of this IPO activity, [73:40] listings, great for the exchange. Then [73:43] you think about the data business, which [73:45] is [73:45] >> and AI optimization on the data. And the [73:47] market said, Josh, you I got stopped out [73:49] of NASDAQ. I think I made money on it, [73:51] but I had been in it for a while. [73:52] >> And it's been a great name for a long [73:54] time. [73:54] >> Yes. The market told me this spring, to [73:56] your point, that I was wrong, and I [73:58] couldn't proc Wait, we're saying there's [74:01] this whole wave of AI disruptors coming [74:03] along and NASDAQ's not going to capture [74:06] its share of that upside. It makes no [74:08] sense. [74:08] >> Doesn't make sense. Highly regulated uh [74:10] industry data, the cyclical tailwind [74:13] from the capital market cycle. And look, [74:16] our analyst Mike Cypress, who's a a [74:18] great partner and friend of our [74:19] business, has been positive on the [74:21] stock. And as the last two to three [74:23] quarters of evidence shows they continue [74:26] to generate AI uh strong evidence of AI [74:30] utilization in their business and the [74:32] market is just not roaring. So those [74:33] that was the example [74:33] >> around here, we don't believe in [74:34] trouble. That's going to go, right, [74:35] Josh? [74:36] >> That's going to go. Um, [74:37] >> everybody loves a trilogy. [74:38] >> When do you know when do you know when [74:40] you know you're wrong on a stock and [74:42] it's time to take action? [74:43] >> When it's too late. I know it's always [74:44] different, but what are like the big [74:46] picture things that you think about? [74:48] >> So before I answer that, just to also [74:51] reference having a hard and fast [74:53] stop-loss in a context of 50 stocks [74:56] where we try to generate 30 to 40% [74:58] turnover a year with a tax-sensitive [75:01] retail type client, I think is too [75:03] rigid. And so to your question, we've [75:06] adopted quantitative and quantal to use [75:09] Adam's phrasing, quantimental inputs, [75:12] but it's also just at the end of the day [75:14] a decision tree usually binary. [75:16] >> Did you see Adam's piece on stop losses [75:18] this week? [75:19] >> Absolutely. We just we had a call the [75:20] other day on it. [75:21] >> Yeah. Yeah. Okay. Go on. [75:22] >> And so look, I would argue that um [75:26] there's the good news and the bad news. [75:27] I told you Palo Alto earlier when I have [75:30] something in the in the high tracking [75:31] error bucket that we've taken a [75:33] idiosyncratic risk on and it rerates and [75:36] we've owned it for our time horizon one [75:38] and a half two years and it's everything [75:40] that's worked out. It becomes my 40th [75:42] idea isn't as good as my 41st potential [75:44] idea. It's friction and opportunity cost [75:46] in the portfolio. [75:46] >> You have because if that's still in the [75:48] portfolio, it doesn't matter that it [75:50] just doubled. [75:51] >> You're allocating fresh cash to the [75:53] portfolio. Do you really want to buy [75:55] that stock now? [75:56] >> Opportunity cost So that's the good [75:57] scenario. Let's talk about the much more [75:59] difficult conversation which is when [76:00] something goes against you and it goes [76:02] back to in our world the original [76:04] thesis. [76:05] >> So we build out a every time we write up [76:08] a new idea. It's a 20page proprietary [76:09] note that my team uh builds. I have a [76:13] sevenperson PM team. Yes, we use Morgan [76:16] Stanley's Research and have used them [76:17] for 25 years and we really benefit from [76:20] that relationship. We use street [76:21] research and since I took over the team [76:23] 12 years ago, I realized, hey, we have [76:25] all these asset managers on our platform [76:27] trying to talk to our wealth management [76:28] audience. Why don't we talk to their [76:30] PMs? So, we talked to a lot of the buy [76:32] side as of the last 12 years and have [76:34] established great relationships there. [76:37] >> My my [76:37] >> That's a great filter because those [76:39] people are putting risk on in those [76:41] stocks or taking [76:42] >> Absolutely. And in some cases, they look [76:43] like me in terms of process. In some [76:45] cases, they have a shorter horizon or a [76:47] riskier horizon. But I love the mosaic [76:49] of all the input there and we have great [76:51] relationships there and I can recommend [76:54] some friends for the for the pod. Um but [76:56] what I would just say is when I look at [76:58] that second scenario, it comes back to [77:00] the original thesis. Did something [77:02] unforeseen happen in terms of [77:05] competition, regulation, management, [77:07] governance and is the earnings power [77:10] temporarily impaired or permanently [77:12] impaired? is the competitive moat [77:14] temporarily in question or permanently [77:17] impaired and you have to make that [77:18] trade-off. So, two things I will tell [77:20] you, one on the stop-loss front and one [77:22] on just my 20 years um deducing kind of [77:27] qualitative signals on the stop-loss. [77:30] Cut your cyclical losers faster. Give [77:33] your secular [77:33] >> faster than the growth [77:34] >> faster because because correct because [77:37] the cyclical momentum almost always [77:40] begets more cyclical downside momentum. [77:42] almost always whether it was oil rigs in [77:44] 2014 2015 whether it was banks in 0708 [77:48] and like when the earnings drop out on [77:50] the cyclicals you're not talking about [77:52] an earnings haircut of 10% or 20% [77:55] >> less likely for a V-shaped recovery in a [77:59] material stock for example [78:00] >> 1,00% and so and let me give you the [78:02] counter to that [78:03] >> that's a good Are you writing these down [78:05] >> okay [78:05] >> the counter to that is on the secular [78:07] names and I'm going to bring in my [78:09] second comment in in a segue here on the [78:12] secular names, maybe manage them down, [78:14] maybe risk manage them as they beget [78:17] negative momentum. But if they're really [78:19] those Byron, big idea secular winners, [78:22] you know, Mike Durban, Durban amendment [78:24] 2010, Visa and Mastercard, they were [78:25] down 30% in a six-month period. They're [78:27] at 10 baggers since then. Um, you know, [78:30] Apple were worried about um advertising, [78:33] worried about Samsung, you know, 10 and [78:35] 20. So, if they're really the secular [78:38] winners, [78:39] >> they it can almost always come back. [78:40] Whereas the cyclicals different story [78:42] >> and you're treating different types of [78:44] stocks differently in the risk [78:46] management. [78:46] >> There's no there's I wish there was [78:48] there's no formula but it's one of those [78:50] experience judgment things we've learned [78:52] over 20 years. And the second comment [78:53] I'll come back really quickly to end two [78:56] types of risk factors [78:57] >> to end you're not getting out of here [78:59] >> on this on this topic on the risk [79:01] factors competitive risk versus [79:03] government and regulatory risk. Here's [79:05] where I've also extrapolated a lot of uh [79:08] signal over time on the government [79:11] regulatory front. It's almost always [79:13] temporary and over overkill in nature. [79:16] And there's exceptions to that, but [79:18] Durban amendment 2010 as a case in point [79:20] I just mentioned. It's almost always [79:23] more fear and more quickly discounted [79:26] incorrectly by the market than not. [79:27] >> I learned that lesson every six months. [79:29] I'm a shareholder in Live Nation, so I I [79:32] I know this. Oh, great. new high soon as [79:35] soon as people start freaking out about [79:37] the next attorney general is suing them. [79:39] >> And we've talked about this a little bit [79:40] in terms of the macro, but often times [79:42] you need a wall of worry for a stock fat [79:44] call to work too. Last comment on this [79:46] topic. [79:47] >> By contrast, the competitive risk is so [79:49] vastly different. And this is where it's [79:51] more nebulous by nature and just harder [79:53] to discern. When companies are losing [79:56] market share, when a brand is starting [79:59] to fade in terms of its relevance, that [80:02] is almost always the death nail of many [80:04] businesses. And and here's the hardest [80:06] part for me because we only fish in the [80:08] ascendant and highquality cohorts. It [80:11] happens to everyone. It even happens to [80:13] high quality [80:13] >> Nike [80:15] which I [80:15] >> Lululemon was a quality stock at one [80:17] point, [80:18] >> right? And then and then they go things [80:20] go in and out of favor. [80:21] >> But Abbercrombie is back. Interestingly, [80:23] that stock is on fire. [80:25] >> The com the competitive dynamic in terms [80:28] of e uh brand and market share, brand [80:31] equity and really just governance, [80:34] management change over time. One of my [80:36] first managers when I started on the [80:37] team 18 years ago, first exercise he had [80:40] me do as a first year associate, fresh [80:43] out of, you know, the analyst program [80:44] and I was a liberal arts major. Dan, do [80:47] a study of executive compensation across [80:49] all of our 50 holdings and tell me about [80:51] the correlation between compensation and [80:54] pay and shareholder return and try to [80:56] find outliers. And he had me do that [80:58] exercise probably 11 or 12 different [81:01] times. Um, and so I mentioned governance [81:04] earlier before. Changes in governance, [81:07] lack of credibility, lack of capital [81:09] allocation, consistency, those are some [81:11] of the qualitative things that we pay [81:13] attention to. You know, it's so funny [81:14] because the way that you do things lines [81:17] up so much better with my actual lived [81:19] experience in the stock market, [81:21] >> but when I started, and maybe you [81:24] experienced some version of this, this [81:26] was not the way things were done. It was [81:28] much more about metrics [81:30] >> and it was much more about math and it [81:32] was much more about well, this PE ratio [81:35] is lower than that PE ratio. [81:37] >> Remember when valuation mattered, [81:39] >> right? But so so but that's that was [81:42] like everybody coming out of Wharton. [81:44] That's what they were taught. And so [81:45] they would look at like the steel sector [81:48] and they would say Bethlehem steel is [81:50] the lowest PE ratio. Yes, it's got the [81:53] highest debt but you know we're buying a [81:56] dollar for50 rather than IshPat steel [81:59] which is you know a dollar for 80ent. [82:02] And that was just the way people thought [82:04] about the I don't think anyone thinks [82:06] about the market. I think more people [82:07] now think about it the way you and I do, [82:10] which is that um the price is behaving [82:14] in whatever way it's behaving relative [82:16] to valuation because the market is [82:19] smart. [82:20] >> The market has figured out [82:22] >> the market has figured out that this [82:23] stock should be an elevated valuation. [82:25] That's not a negative. I'm not saying [82:27] it's a positive and we only want to buy [82:29] expensive stocks, but oh my god, are we [82:31] saying it's expensive by accident? [82:34] >> We can't be saying that. No, I [82:35] >> it's amazing how many people were [82:37] trained thinking that expensive stocks [82:40] relative to their peers were [82:42] accidentally they call it a mispricing. [82:45] >> What do you mean a mispricing? No, it's [82:47] not mispriced. They want to own Dell cuz [82:49] Dell's better than Compact. What part of [82:52] that don't you understand? [82:53] >> Right. [82:53] >> So, I think people have come along to [82:55] that. [82:55] >> Well said. And I think the two things I [82:57] would just add to that Josh are number [82:58] one is the access to information, the [83:01] speed of information and the [83:02] proliferation of information has helped [83:05] that efficiency in nature. And secondly, [83:08] it's the it's the um micro structure of [83:10] who's invested in the markets. Think [83:12] about the proliferation of CTAs and [83:14] quant funds on the institutional side. [83:16] Think about all the liquidity on the [83:18] retail side. You know, I didn't mention [83:19] it earlier, but I meant to. I made a [83:21] mental note of it, but something like [83:23] 40% of homes in the US are owned [83:25] outright, no mortgage. So, think about [83:27] an aging population with excess capital [83:30] on hand. And maybe from the other part [83:32] of the spectrum, the younger uh um [83:34] cohort who has more access to investing, [83:37] gambling tools, and just think about the [83:39] momentum and trend followers in the [83:41] market versus classically trained [83:43] Wharton MBAs doing valuation work. The [83:46] micro structure has changed. The [83:47] technology and the speed of information [83:50] has changed making momentum one of the [83:52] single most important factors in markets [83:54] this year through June 22 and the last 5 [83:57] years. And so what we say is look, it [83:59] can be true on the one hand that we're [84:01] quality oriented. We're diversified. [84:03] We're long-term investors. We have one [84:05] client in mind who's wealth management [84:07] and we manage the tales. That can be [84:09] true on the one hand, but on the other [84:10] hand, uh don't pay attention to the [84:12] technicals and the momentum at your own [84:14] risk. That's right. It's nice to be able [84:16] to punch into Google what is the PE [84:19] ratio in this stock, but everybody knows [84:20] that already. So, the momentum is the [84:23] actual story. The stock is under [84:25] accumulation or it's not. And watch for [84:27] when momentum changes. We are at a [84:29] crucial inflection right now. Whereas [84:31] the affformentioned semis hardware and [84:33] power names have either traded sideways [84:36] since June, failed to recover the [84:37] 50-day, not had the PB hedge fund guys [84:40] regrow back up. That's going on on the [84:43] one hand, and we've talked about it. [84:44] Healthcare is working. Soft parts of [84:46] software is working. Financials are [84:47] working. Energy is working. You know, [84:49] one of our very simple tools this year [84:52] has been every time we've had a [84:53] ceasefire is to buy more energy in the [84:55] book because every every ceasefire deal [84:57] has been pretty shortlasting. And so, [85:00] yes, you have to pay attention to the [85:02] momentum. You have to also try to [85:03] exploit some of the anomalies out there, [85:06] too. Last thing, um, you think the year- [85:10] end setup is favorable given the [85:12] earnings growth and all these tailwinds [85:13] we talked about? You're feeling pretty [85:15] good about like I mean most years we go [85:18] out close to the highs, so it wouldn't [85:19] be surprising. [85:20] >> Yeah, it's been tough to bet against [85:21] seasonality the last several years. So I [85:24] would say per our pop-up ad exchange [85:26] earlier, I think at best we trade [85:28] sideways into the midterm. As you stated [85:30] Josh earlier, I don't think there's any [85:32] real profound outcome from the midterm. [85:34] But then I think it's all about, you [85:36] know, in terms of October earning [85:37] season, I think all the trends and data [85:40] we're seeing suggest and the conference [85:42] season this September should also u [85:45] predict pretty solid acceleration there. [85:47] And then it becomes what happens in 27. [85:49] So, let me leave you with my 27 thought, [85:51] which is in my mind 27 could look a lot [85:54] like 23 where it's a return of more [85:58] monolithic mag 7 outperformance. I think [86:01] the average company is going to have [86:03] trouble comping that synthetic tariff [86:06] operating leverage. So, if I were to [86:08] make a a more of a thematic call for [86:10] next year, it could be one of these [86:12] ideas where it's [86:13] >> bigger is better again. [86:14] >> Bigger is better, quality is back and it [86:16] could be a broad it could be a broader [86:18] market, but you definitely want to be in [86:19] kind of the 23 kind of AI enabler mag 7 [86:23] names and maybe some of the healthcare [86:25] adopters and maybe it's not as good a [86:27] year for small caps or the average [86:28] stock. [86:29] >> I love it. Good stuff. Did [86:30] >> you have fun on the show today? It [86:31] >> was really fun. Thank you. [86:32] >> Okay. Uh we're h you're happy on your [86:35] end. [86:36] >> Okay. Hey, we're going to take we're [86:37] going to take a brief recess and then [86:38] we're going to do some Israel Palestine [86:40] stuff. Is that okay? Is that good? All [86:42] right. You You were awesome on the show [86:43] today. Put the headphones back on. We're [86:45] going to give you your flowers. [86:46] >> I want to tell people where they can [86:48] learn more and get more of your [86:49] insights. So, you're going to do squawk [86:51] from time to time. Um do you guys [86:53] publish anything anywhere? Are you on [86:54] LinkedIn? Tell us where [86:55] >> we do and thanks for that um that heads [86:58] up. So, Josh, we I'm on LinkedIn and the [87:00] firm Morning Stanley's on LinkedIn. We [87:02] are available through those channels [87:04] channels and then yes we publish through [87:06] the global investment committee I [87:08] publish a quarterly letter which I'll [87:10] share with you both in terms of my fund [87:12] the SMA and then at times as per the [87:14] software and semi comments we'll write [87:16] more tactical ad hoc client facing [87:18] material so we'll share all those with [87:20] you [87:20] >> that's great and I know the audience [87:21] definitely uh is going to want to hear [87:23] from more from you thank you so much for [87:24] joining us for a lot of fun really [87:26] appreciate it guys great job this week [87:28] >> John Duncan Nicole the whole team [87:30] appreciate it. Thank you so much for [87:33] watching. Thank you for listening. We'll [87:34] see you next week. Thanks again.