Morgan Stanley Sees AI Data Center Buildout Driving 25% of US GDP Growth

摩根士丹利：AI数据中心建设将拉动美国25%的GDP增长 摩根士丹利预计，截至2028年在AI基础设施领域的2.9万亿美元支出将推动美国经济增长并引发巨大的盈利分化，从而使市场免受高企的地缘政治和政策风险的影响。 Michael Zezas解释了从传统的华盛顿共识转向多极世界如何推动全球再工业化、供应链回流和巨额资本支出。讨论强调了AI基础设施支出如何成为美国GDP增长的主要驱动力，尽管存在电网瓶颈，2025年至2028年间全球数据中心资本支出预计将达到2.9万亿美元。此外，对话还探讨了政策不确定性与市场波动性之间的脱节、市场集中与拓展趋势，以及美联储在政治过渡期的制度韧性。 摩根士丹利指出由自由贸易和有限政府干预构成的旧“华盛顿共识”已经破灭，其确立的“多极世界”投资框架表明，全球化红利期已成过去，地缘政治脆弱性与持续抬升的贸易壁垒已成为市场新常态。 (14:28) 两党政府在政府主导的产业政策、制造业回流以及国家安全供应链战略上保持着高度连贯性；英特尔（Intel）凭借《芯片法案》（CHIPS Act）的联邦资本支持，成功从此前备受估值惩罚的传统制造厂转变为核心战略受益企业。 (17:15) 孤立准确地预测某项政策落地并不等同于能精准预判市场反应，因为AI资本开支潮等强大的反向顺风完全可能压过宏观政策带来的拖累。 (22:11) 摩根士丹利预计2025年至2028年全球仅用于AI数据中心建设的累计资本开支就达2.9万亿美元，且当前AI物理基础设施建设已独立贡献了约四分之一（25%）的美国GDP增长。 (36:35)

Morgan Stanley Deputy Global Head of Research Michael Zezas argues that the post-war Washington consensus favoring free trade and deregulation has permanently collapsed into a fragmented multipolar regime. In this new era, market performance is primarily anchored by structural re-industrialization, bipartisan economic nationalism, and a massive capital expenditure cycle in artificial intelligence. Collapse of the Bipartisan Washington Consensus (14:28) Zezas contends that the political consensus prevailing from the 1980s to the 2008 financial crisis—characterized by bipartisan support for free trade, multilateral institutions, and minimal state intervention—has disintegrated. Driven by stagnant real wage growth, domestic opioid epidemics, and foreign conflicts, voters have rejected globalization. In response, Morgan Stanley treats rising protectionism, tariffs, and government intervention as permanent structural features rather than temporary political anomalies. This shift demonstrates remarkable continuity across presidential administrations. While many institutional investors surveyed in 2020 assumed a Biden administration would reverse Trump-era tariffs on China, the White House instead retained and broadened them while implementing aggressive industrial policy. The federal backing of Intel under the CHIPS Act highlights this dynamic, as government subsidies transformed a penalized legacy manufacturer building Arizona foundries into a strategic domestic asset, facilitating a share price recovery from $15 toward $80. Global Re-Industrialization in Europe and Japan (32:45) The transition to a multipolar world has forced U.S. allies to reconsider their reliance on American military protection and low-cost Chinese manufacturing. Catalyzed by security conflicts in Ukraine and the Middle East, foreign governments are ramping up capital investments to secure their own sovereign industrial bases, turning geopolitical necessity into an economic driver for equities. Zezas highlights that non-U.S. developed markets are emerging as prime beneficiaries of this realignment. In Europe, heavy capital expenditure directed toward defense and energy independence positions industrial corporations to gain operational efficiency from automation. In Japan, ongoing corporate governance reforms and returning inflation coincide with rising economic nationalism, prompting the reallocation of an estimated $7 trillion in household cash deposits into risk assets. AI Infrastructure Capex and Earnings Divergence (36:35) Morgan Stanley projects $2.9 trillion in cumulative global capital expenditures strictly dedicated to AI data centers between 2025 and 2028. The firm's economics team estimates that the physical construction of this infrastructure alone accounts for roughly 25% of current U.S. GDP growth. However, further growth faces operational bottlenecks outside of software, specifically electrical power grid constraints, labor shortages, and local zoning opposition. At the corporate level, adoption is driving a stark operational divergence. Morgan Stanley's data indicates that over 30% of covered global firms now report measurable earnings gains from AI, with adopting companies expanding EBITDA margins twice as fast as non-adopters. This divide is reflected in consensus 2027 earnings-per-share forecasts: AI infrastructure stocks are projected to grow earnings by 42%, compared to 10% for the overall S&P 500 and just 3% for the index excluding AI infrastructure. Geopolitical Disconnects and Market Resilience (28:42) A noticeable gap has opened between historic highs in the Economic Policy Uncertainty Index and a subdued CBOE Volatility Index (VIX). Zezas explains that equity markets often price existential geopolitical disruptions at very low probabilities, as investors recognize that worst-case systemic crises cannot be effectively hedged, whereas near-misses spark swift recoveries. Furthermore, isolated policy shocks, such as unexpected tariff threats, have been repeatedly offset by stronger secular tailwinds like the corporate AI spending wave. Regarding domestic political risks ahead of the 2026 midterms, Zezas advises institutional investors against reacting to short-term legislative noise. Although prediction markets assign roughly an 80% likelihood to Democrats taking the House, any resulting Senate majority is projected to be razor-thin (50 to 53 seats). Such margins will lack the filibuster-proof majority required to repeal existing corporate tax structures or reclaim executive tariff authority. Morgan Stanley maintains a constructive outlook across equities, supported by resilient corporate earnings, stable interest rates, and an FOMC governance structure insulated from political shocks. However, Zezas cautions that translating macro policy calls into portfolio returns remains difficult, noting that automated AI models increasingly risk producing overfitted 'research slop' unless paired with disciplined human judgment.

摩根士丹利（Morgan Stanley）研究部副主管Michael Zezas与节目主持人指出，基于自由贸易与小政府理念的旧“华盛顿共识”已彻底瓦解，全球正式进入由产业政策主导的多极化再工业化时代；尽管宏观政策不确定性高企，但在AI基础设施巨额资本开支与供应链重塑的驱动下，宏观经济与股票市场仍展现出较强的内在韧性。 华盛顿共识破裂与全球再工业化转型 (14:28) Zezas分析指出，1980年代至2008年间美国两党在自由贸易、有限政府干预及多边机制上的共识已不复存在。长期实际工资停滞、阿片类药物危机以及海外冲突导致选民深感被全球化边缘化，从而在政治上倾向选择能够打破现有体制的领导人。这一民意转向促成了两党政策的高度延续性：无论是关税还是非关税壁垒，政府直接介入产业政策和推动本土制造业回流已成为长期趋势。例如，2020年大选前市场曾错误预期Biden政府会撤销对华关税，但随后不仅关税得以延续，政府还通过《芯片法案》（CHIPS Act）直接资助Intel等企业在本土设立晶圆厂，推动其股价从15美元低点大幅回升。 这一多极化趋势正在向全球蔓延。欧洲领导层意识到无法完全依赖外部防御与供应链，正向国防、能源安全以及重工业自动化投入重金，摩根士丹利欧洲股票团队因此将欧洲重资产工业巨头视为“隐性AI受益者”。在日本，企业治理结构改革、温和通胀回归以及经济民族主义正在激活本土资本，该国家庭持有的约7万亿美元现金储蓄正逐步转向风险资产，进一步催化了非美发达市场的再工业化叙事。 地缘政治与政策不确定性下的市场定价逻辑 (29:52) 尽管经济政策不确定性指数升至历史高位，但芝加哥期权交易所波动率指数（VIX）仍处于低位，市场对传统黑天鹅事件（如霍尔木兹海峡地缘危机）的反应显著钝化。Zezas指出，当某个地缘事件的终局可能对全球经济带来毁灭性打击时，市场的理性定价往往是赋予其极低的发生概率。主持人Josh Brown引用纽约证券交易所传奇交易员Art Cashin的名言解释称：“当导弹升空时必须买入；若导弹击中，一切都无所谓了，若没有击中，所有利润都由此产生。” 此外，将单一政策预测转化为投资组合超额收益极其困难。Zezas以关税政策为例，若投资者在关税宣布前单纯依模型下调盈利预期并做空，将会遭受惨重损失，因为AI资本开支暴增与估值扩张等反作用力完全对冲了关税拖累。尽管预测市场（如Polymarket或Kalshi）能提供比传统民调更纯粹的概率数据，但在多变量交织的宏观环境下，孤立的政策预判往往难以直接映射为资产收益。 AI资本开支浪潮、基础设施瓶颈与盈利分化 (36:35) 摩根士丹利预测，2025年至2028年间全球仅用于AI数据中心建设的累计资本支出就将达到2.9万亿美元。该机构宏观团队测算，仅AI物理基础设施和数据中心的建设投资，就贡献了今年美国GDP总增长的约四分之一（25%），且该测算尚未计入未来的全要素生产率提升。然而，制约这一扩张速度的核心瓶颈已从软件转向物理现实，包括电网容量短缺、技术劳动力供给不足，以及各州日益抬头的邻避效应（NIMBY）。 在企业应用层面，摩根士丹利对全球3400家覆盖公司的追踪显示，报告实现切实AI盈利效益的公司比例已从2024年初的约20%上升至30%以上，且AI采用者的EBITDA利润率扩张速度是未采用者的两倍。市场一致预期2027年AI基础设施相关股票的每股收益（EPS）复合增长率为42%，而标普500指数整体仅为10%，若剔除AI基础设施后该增速则骤降至3%，印证了基本面盈利的显著分化。 市场广度扩散、选举噪音与美联储制度韧性 (45:51) 近期美股盘面呈现从超大盘科技股向外扩散的特征。尽管“科技七巨头”（Magnificent 7）年内涨幅回落至约70个基点，但纳斯达克100指数上涨19%，等权重标普500指数上涨11%，等权重科技ETF（RSPT）涨幅更是达到39%。Zezas认为，这一分化表明资本开支周期正带动更广泛的标的参与行情，大摩资产配置策略因而维持美股优于债券的观点，将固定收益主要视为提供票息收益和分散风险的工具。 针对2026年中期选举，预测市场显示民主党赢得众议院的概率约为80%，参议院控制权则呈五五开局面。Zezas分析，微弱的参议院多数难以跨越阻挠议事的门槛，此前推行的企业减税等重大法案很难被实质性废除，建议投资者过滤选举早期的舆论噪音。在货币政策治理方面，针对Kevin Warsh可能执掌美联储的讨论，市场表现相对平静，主要归因于美联储的利率决定依赖联邦公开市场委员会（FOMC）的集体决策机制，其制度框架能够有效抵御单一行政干预。 摩根士丹利目前对多数资产类别持有建设性展望，认为AI资本开支与供应链调整是长期的结构性动力；但该判断的前提假设在于企业盈利增速维持扩张、中东等局部地缘局势不发生极端恶化，以及货币政策利率路径能够保持相对平稳。

## Background 本期节目由 YouTube 财经频道 The Compound 于 2026 年 7 月 13 日发布。节目中主持人 Downtown Josh Brown 与 Michael Batnick 邀请到摩根士丹利（Morgan Stanley）全球研究副主管 Michael Zezas，深入探讨 AI 资本开支（CapEx）、数据中心建设、多极化世界秩序、2026 年中期选举以及为何政策预期极难直接折射为投资回报等核心宏观议题。 ## TL;DR 摩根士丹利全球研究副主管 Michael Zezas 指出，全球化红利退潮与多极化格局确立已成为不可逆的宏观体制变迁，驱动着欧美与日本开启以防务、供应链安全和再工业化为核心的庞大资本开支周期。针对炙手可热的 AI 浪潮，大摩认为当前仍处于极早期阶段，虽然电力与基础设施瓶颈拉长了军备竞赛的节奏，但微观财务数据显示采纳 AI 的企业 EBITDA 扩张速度显著领先，且今年美国约四分之一的 GDP 增长直接由物理基建投资拉动。面对 2026 年中期选举与地缘政治动荡，Zezas 强调华盛顿两党在产业政策与贸易壁垒上的高度延续性，并提醒投资者在多变量交织的复杂市场中切忌机械押注单一政策剧本，真正的超额收益依旧取决于跨资产研究与人类专家的审慎判断。 ## Atmosphere & style 整场对话在轻松惬意的老友闲聊（从影史经典到生活八卦）与极高密度的华尔街专业投研交锋之间无缝切换。主持人的风趣锐利、机智捧哏与嘉宾严谨宏大、深入浅出的机构首席风范碰撞出极佳的化学反应，既有“导弹在天上飞时买入”这种充满实战色彩的传奇交易智慧，也有“番茄切不切入水果沙拉”式对 AI 时代人类认知壁垒的生动解构。收听体验既像置身于曼哈顿高端投研例会般硬核清晰，又如酒吧老友叙旧般松弛自然，将枯燥复杂的宏观多变量博弈化作极具感染力与启发性的思想盛宴。 ## Bulletin - [36:14](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=2174s) 摩根士丹利预测 2025 至 2028 年间全球数据中心相关资本开支将达到 2.9 万亿美元，实体基建建设今年已直接贡献美国约四分之一的 GDP 增长。 - [40:28](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=2428s) 微观财务验证显示，MSCI 全球指数中能证明 AI 产生实质效益的企业占比已升至 30% 以上，AI 采纳者 2024-2025 年的平均 EBITDA 扩张速度是非采纳者的两倍。 - [45:43](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=2743s) 美股涨势呈现明显的广度扩散：科技巨头 Mag 7 年内基本走平（仅涨约 70 个基点），而科技等权重 ETF（RSPT）年内涨幅已高达 39%。 - [17:41](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=1061s) 地缘政治脆弱性促使全球告别单一效率最优导向，企业被迫斥巨资构建“反脆弱”供应链，欧洲防务重工与日本公司治理改革构成“再工业化”双主线。 ## Key takeaways - [14:01](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=841s) 旧华盛顿共识全面瓦解，选民对自由贸易与放任主义的疏离推动宏观体制转型，“多极化世界”与产业回流成为不可逆的长期趋势。 - [17:41](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=1061s) 地缘冲突暴露全球贸易咽喉要道，跨国企业从追求极端成本效率转向建立反脆弱供应链，推动天量资本开支流向本土再工业化。 - [19:42](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=1182s) 英特尔依托《芯片与科学法案》实现估值重估，证明安全优先与供应链本土化的产业政策跨越党派轮替，具有高度的长期连续性。 - [22:11](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=1331s) 宏观投资决策属于复杂多变量博弈，即使精准预测单项政策变动，市场定价路径仍可能因其他变量反向对冲而脱离直觉。 - [26:47](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=1607s) 大语言模型（LLM）目前在处理缺乏固定路径的跨资产多变量宏观推演上成功率有限，人类团队的深度协同研究仍是核心壁垒。 - [28:41](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=1721s) 面对极端地缘危机与尾部毁灭风险，理性的市场定价往往赋予极低概率，且白宫对资产价格的现实敏感度为股市提供了托底支撑。 - [32:45](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=1965s) 欧洲重工业成为 AI 赋能生产率的隐性受益者，日本则在企业治理改革与良性通胀推动下迎来居民现金资产入市重配的内生动能。 - [36:53](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=2213s) AI 基础设施建设受制于电力、能源短缺及地方社区阻力，但物理瓶颈反而拉长了整体投资周期的跨度，目前仍处于极早期阶段。 - [43:42](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=2622s) 市场预期 2027 年 AI 基础设施股票每股收益将大增 42%，而标普 500 指数剔除该板块后仅增 3%，非 AI 行业的实际盈利扩张亟需跟进。 - [48:12](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=2892s) 2026 年中期选举大概率陷入立法僵局，缺乏绝对多数意味着既有税改法案与总统关税裁量权均难以被推翻，选举情绪冲击适合逆向交易。 - [51:05](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=3065s) 生成式 AI 将蒙特卡洛模拟和海量契约分析门槛大幅拉低，投研竞争全面从人力信息差转向驾驭前沿工具与穿透底层大数据的能力。 - [55:01](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=3301s) AI 模型的迎合与过度拟合风险或导致“研报垃圾”泛滥，具备实战审慎考量与行业经验的人类专家网络仍是过滤噪音的关键。 - [56:52](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=3412s) 劳动力市场尚未观察到初级专业岗位的大规模结构性失业，AI 对青年就业的平滑承接暂时好于悲观预期。 - [58:01](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=3481s) 市场对“AI 杀死 SaaS”反应过度，企业级软件平台的底层业务流工业化封装、合规风控及私有数据集成壁垒无法被员工自建代码轻易替代。 - [01:00:23](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=3623s) 凯文·沃什执掌美联储稳定了华尔街预期，FOMC 的委员会集体决策机制构成了抵御单一行政干预与维持货币政策理性的制度防火墙。 ## Detailed walkthrough ### [00:00](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=0s) 录制前的寒暄、电影闲聊与姓氏发音 录制刚开始，演播室里充满了轻松惬意的闲聊氛围。主持人和嘉宾迅速拉近关系，彼此确认了共同的好友是 Colin Ross，原来大家还是大学时期的旧识。话题随即转向了日常的观影体验：有人提到在坐火车回家的路上刚看了重新数字修复版、长达整整 4 个小时的历史史诗片《阿拉伯的劳伦斯》（*Lawrence of Arabia*），并开玩笑说这种时长只能把它拆成几集电视剧来看；而 Michael Batnick 则打趣问这片子是不是杰克·吉伦哈尔（Jake Gyllenhaal）主演的，Josh 笑着反驳：“这是 1962 年的电影，里面的演员现在全都已经过世了，那可是影史上最伟大的一批演员。” 随后 Michael Batnick 吐槽 Josh Brown 极度喜欢剧透。Batnick 表示自己看电影喜欢“白纸一张”（raw dog）地去看，不想提前看烂番茄（Rotten Tomatoes）的评分，也不看影评，生怕被抬高或拉低预期；但 Josh 前一天刚看完斯皮尔伯格导演的新片《披露日》（*Disclosure Day*），一上来就直接给他来了一通全盘剧透，不仅评出了“7 分（满分 10 分）”，还直接揭底“外星人根本没出现”。此外，他们还讨论了该片上映前那种病毒式营销（Gorilla marketing），一度让人误以为真要公开外星人存在的秘密。在闲扯之余，大家还特意核对了嘉宾姓氏“Zezas”的准确发音——Josh 直觉认为它发音类似于“Jesus”，Michael Zezas 笑着证实确实如此，尽管这是一个希腊姓氏，在希腊语中本应读作类似于“Sezos”，但他们全家人在美国一直都习惯读作押韵于 Jesus 的发音。 --- ### [02:30](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=150s) 摩根士丹利的研究部掌门人与办公室日常 话题转回到摩根士丹利内部架构。Josh Brown 顺便向 Michael Zezas 询问了摩根士丹利现任的研究部全球主管是谁，Zezas 指出是 Katie Huberty，并强调“责任在她那里，她说了算”（"Buck stops with her"）。 当被问及办公地点时，Zezas 透露自己常驻在时代广场的百老汇 1585 号总部大楼（1585 Broadway），从 2010 年起就在那里办公。他在摩根士丹利已经工作了 19 个年头，早年公司刚起步时，他还在第五大道 522 号（522 Fifth Avenue）的办公室待过，最初正是在财富管理（Wealth Management）旗下的单独管理账户（SMA）平台上负责市政债券与企业债的信用分析。 --- ### [03:40](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=220s) 赞助商插播：Public 现代投资平台 在节目正式开场前插播了赞助商 Public 的广告。文案指出当下的投资平台往往走向两个极端：要么是那种界面古老、宛如 1997 年设计的传统老牌券商，要么就是把投资当成体育博彩搞成线上赌场的散户交易软件。Public 则主打专为严肃投资者设计的现代综合投资平台，支持股票、期权、债券和加密货币。平台近期大力拥抱 AI，推出了可以帮用户监控市场、管理闲置现金流并执行自动化交易的专属投资 Agent（智能体），投资者只需输入 Prompt 并确认工作流即可启动。 --- ### [04:34](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=274s) 节目正式开场与 Michael Zezas 的职业履历 节目第 247 期正式拉开帷幕，主持人 Downtown Josh Brown 携手搭档 Michael Batnick 隆重介绍了本期嘉宾——摩根士丹利研究部副全球主管（Deputy Global Head of Research）Michael Zezas。Josh 详细回顾了 Zezas 极其亮眼的职业履历： - 他于 2007 年以信用分析师和投资组合经理的身份加入摩根士丹利； - 2010 年升任市政信用策略主管（Head of Municipal Credit Strategy）； - 2016 年进一步出任美国公共政策研究主管（Head of US Public Policy Research）； - 自 2013 年起，他连续多年入选《机构投资者》（*Institutional Investor*）全美固定收益研究团队； - 在 2014 年至 2021 年期间，他更是常年稳居 Smith's Research and Ratings 的全明星第一梯队（All-Star First Team）。 Josh 还不忘打趣调侃一句：“当然，到了 2022 年他的评选掉榜了，那一年确实艰难，不过我们今天就在节目里帮他重振雄风。” --- ### [05:52](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=352s) 摩根士丹利的财富管理霸权与 James Gorman 的战略布局 谈到摩根士丹利近年来的强劲表现，Josh 作为股东表示感到非常满意，并高度盛赞了前任 CEO 詹姆斯·戈尔曼（James Gorman）的战略远见。Josh 指出，华尔街许多大型机构的财富经理每天都在为去哪里找新客户而苦恼，但戈尔曼深刻领悟到了核心逻辑：要打造一个真正的财富管理帝国，必须源源不断地为顾问提供庞大的客户需求和流量漏斗。 为此，摩根士丹利精准出击，通过并购 E*TRADE，以及收购并整合现在的“Morgan Stanley at Work”（企业股权管理服务）业务，打造出强大的获客漏斗，源源不断地将普通散户和企业高管转化为高端财富管理客户。这使得大摩一跃成为华尔街增长最快的财富管理巨头。 ```mermaid flowchart LR accTitle: Morgan Stanley Wealth Management Funnel accDescr: Diagram showing how acquisitions feed into Morgan Stanley Wealth Management client pool. subgraph AcquisitionSources["获客渠道与漏斗平台"] direction TB E["E*TRADE 平台"] MSW["Morgan Stanley at Work"] end subgraph CoreWealth["核心财富业务"] WM["摩根士丹利高端财富管理顾问"] RC["研究智库与配置方案"] end E -->|"零售散户转化"| WM MSW -->|"企业高管与股权激励转化"| WM RC -->|"提供顶尖研究支持"| WM ``` Zezas 回顾道，自己职业生涯的前几年虽然只在这部庞大机器中贡献了一小部分力量，但他从事了长达十年的地方债研究，深知将前沿研究智库转化为可落地的客户解决方案，对于整个财富顾问网络起到了至关重要的赋能作用。 --- ### [07:23](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=443s) 宏观视角的转变：单线保险商崩塌与 Meredith Whitney 恐慌 Josh Brown 提出了一个关键问题：市政债券（Munis）通常被视为华尔街颗粒度最细、极度微观具体的“显微镜式”细分领域，Zezas 究竟是如何从这个领域跃迁为纵览全局的全球宏观政策专家的？ Zezas 回忆指出，这一转变源于 2008 至 2009 年金融危机期间市政债券市场发生的根本性范式转变： - **危机之前的单线保险模式**：在 2008 年金融危机之前，几乎所有地方市政债券都被单线债券保险公司（Monoline Insurers，如 MBIA、Ambac 等）打包包裹（wrapped）。在那种环境下，分析师几乎只需要评估这一两家保险公司的信用，所有地方债都被直接视同于 AAA 级安全资产，底层具体的市政信用基本不需要深入研究。 - **危机后的信用瓦解与主权叙事**：随着次贷危机爆发，单线保险商轰然崩塌，所有的安全背书瞬间消失。市场突然意识到，投资者必须直面底层每一个发行实体的真实资产负债表与财政状况，分析师必须逐一剖析底层信用。 - **与宏观危机的共振**：这一转变恰逢欧洲主权债务危机的爆发，全球宏观叙事开始大规模渗透进固定收益市场。当时市场上充斥着各种极端的恐慌论调，最著名的便是分析师梅雷迪斯·惠特尼（Meredith Whitney）公开预言美国将掀起数百亿美元规模的地方政府违约潮。当时 Zezas 奔走于各大对冲基金之间，极力向他们厘清市政债券底层信用与真正系统性主权违约之间的本质区别。 ### [08:20](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=500s) 从市政债策略师到华盛顿政策观察家 Michael Zezas 回顾了自己的职业生涯起点。在欧债危机与美国地方财政承压时期，市场曾热议加州是否会变成下一个希腊，但两者基本面有着天壤之别——希腊每年需要展期的债务规模相当于其财政收入的 150%，而加利福尼亚州的本息偿付支出（P&I）仅占财政收入的约 5%。这种微观层面的讨论让他深刻意识到，公共政策对市政信用评级与客户投资组合的重要性。作为固定收益产品，市政债不仅受货币政策影响，更深植于财政与税收政策的交叉点。 ```mermaid flowchart TD accTitle: 从市政债策略到多资产公共政策研究的转型路径 accDescr: 展示 Michael Zezas 如何将微观政策分析拓展为跨资产的宏观政策框架 subgraph Phase1["微观与基本面阶段"] A["市政债策略分析<br>（超越利差与展期滚动的微观计算）"] B["公共政策交叉分析<br>（医改法案与税法变动下透视财务报表）"] end subgraph Phase2["宏观框架拓展"] C["华盛顿政策观察<br>（2016年大选跨资产情景推演）"] D["多极化世界范式<br>（长周期结构性政策框架建立）"] end A --> B B --> C C --> D ``` 当时同行的传统做法往往是测算不同期限（tenor）的利差收益与展期收益（carry and roll down），提出“配医疗债而非公路债”这类琐碎战术。拥有公共政策硕士学位的 Zezas 则选择拉开维度，从宏观趋势和制度背景出发，搭建贯穿税收政策和奥巴马医改（ACA）的分析框架，直接推导法案变动对具体地方政府资产负债表与利润表的影响。2016 年大选前夕，研究部请他牵头梳理“如果特朗普胜选会发生什么”的跨资产预案，尽管当时主流观点认为“他不可能赢”。大选结果揭晓后，客户的问题铺天盖地而来，这使他阴差阳错地成了摩根士丹利内部专职公共政策与政治研究的核心人物。 - 加利福尼亚州债务本息支出（P&I）仅占财政收入约 5%，与希腊高达 150% 收入的展期债务压力存在本质差异。 - 摩根士丹利市政债分析从单纯的利差/滚动测算，转向结合奥巴马医改与税改法案的宏观基本面建模。 - 2016 年针对特朗普胜选的跨资产情景分析，促成公共政策研究在部门内由临时专题转为常设研究方向。 --- ### [11:16](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=676s) 白宫登上棋盘：“回到 2005 年”的幻觉破灭 特朗普首任初期，金融市场进入了前所未有的震荡期。Josh Brown 形容当时的白宫不仅是棋盘上的一颗棋子，甚至俨然成了横冲直撞的“皇后（the queen）”——总统可能随时发推特指责辉瑞（Pfizer）的药品定价，导致其市值瞬间蒸发 500 亿美元。面对海外央行和大型对冲基金“这种动荡还要咬牙坚持多久才能回归常态”的追问，Zezas 和他的团队明确意识到：市场不可能再回到 2005 年的常态。 摩根士丹利凭借跨资产与跨区域的协作文化，建立了税改情景与各行业估值映射的应对路径。但政策研究很快面临“追着自己尾巴跑”的疲惫感。Zezas 团队选择抽身退步，尝试寻找能够支撑数年以上的长周期分析框架。2017 年至 2018 年关税战全面铺开，特朗普频繁召集大企业 CEO 进白宫训话。虽然事后看来许多恐慌并未彻底引爆不可逆的灾难，但当时整个市场感受到的制度性震动极其剧烈。 --- ### [14:01](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=841s) 新华盛顿共识与“多极化世界”范式确立 对于白宫打破常规的行为，Zezas 认为这并非历史偶然，而是选民底层偏好根本性转变的投射。冷战后期至 2008 年金融危机期间的“旧华盛顿共识”——即两党普遍认同的自由贸易、政府不干预经济、维护多边国际组织以及输出软实力——在选民心中已失去吸引力。普通劳动者经历了 15 年实质工资停滞，面对阿片类药物危机与接二连三的对外战争，自然不再支持自由放任主义的世界秩序，转而倾向“为波动性投票”，渴望有人打破现有体制。 正是在 2010 年代末期，摩根士丹利团队正式提出了“多极化世界（Multipolar World）”这一核心研究范式。早在 2020 年，团队就明确判断：“全球化红利最好的时期已经过去，地缘政治脆弱性正在上升，贸易壁垒只增不减。”尽管当时很难精确预测每项政策何时会转化为具体可投资的主题，但这种从底层选民偏好推导宏观制度变迁的研究，最终经受住了时间的检验。 - **旧华盛顿共识的瓦解**：1980 年代初至 2008 年金融危机期间确立的自由贸易与多边主义共识，被追求产业回流与强硬干预的新政治诉求所取代。 - **选民诉求转向**：中底层群体因 15 年工资未实质增长、社区面临阿片类药物危机与对外战争代价，对 NATO、NAFTA 及中国加入 WTO 等宏观叙事产生强烈疏离感。 - **摩根士丹利 2020 年核心研判**：选民偏好重塑驱动宏观体制变革（regime change），全球化红利退潮与地缘脆弱性上升已成为长期不可逆的投资前提。 ### [17:41](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=1061s) 跨党派政策的延续性与供应链“反脆弱”转型 在讨论美国政治与经济周期的交织时，摩根士丹利（Morgan Stanley）全球研究副主管迈克尔·泽扎斯（Michael Zezas）回顾了以往的大选民调。在 2020 年大选前夕，摩根士丹利对投资者进行问卷调查，大多数受访者认为唐纳德·特朗普（Donald Trump）连任对市场更有利；而那些认为乔·拜登（Joe Biden）获胜对市场更有利的投资者，普遍在另一道问题中理所当然地认定“拜登会削减并撤销关税”。 然而事实证明这种预判错得彻底。拜登政府不仅没有退回到自由贸易的旧常态，反而从根本上强化了限制措施，只是策略有所不同——面对同中国的经贸谈判，任何一位总统都不会主动单方面下调关税作为筹码。泽扎斯指出，无论关税还是非关税壁垒，亦或是政府强力干预的产业政策，其前进的大方向早已极其明确。随着时间推移，这种政策意图最终直接演变为了巨额资金砸向的“再工业化”（Reindustrialization）、回流（Reshoring）以及本土化建厂。 与此同时，美国在全球舞台上行使着截然不同的姿态，加之俄乌冲突以及美伊对峙等关键地缘政治危机接连爆发，彻底暴露了全球贸易的“咽喉要道”（Choke points）。全球企业高管层终于彻底清醒：企业再也无法仅凭“全球化最优效率”来配置资源，而是必须转向投资银行家所说的“反脆弱策略”（Anti-fragile strategy）。为了绕过这些咽喉节点和美国设立的对华经贸红线，全球企业不得不花费漫长的时间与天量资金重新调整产能工具与供应链设备。 ```mermaid flowchart TD accTitle: 供应链反脆弱转型逻辑 accDescr: 地缘冲突与政策倒逼跨国企业从全球化效率转向反脆弱策略 subgraph ExternalDrivers ["外部驱动力量"] A["地缘政治危机<br>(俄乌、美伊冲突/航运咽喉)"] B["政府产业政策与贸易壁垒<br>(跨党派保护主义/回流补贴)"] end subgraph CorporateShift ["企业战略转向"] C["告别单一全球化效率最优"] D["构建'反脆弱'供应链 (Anti-fragile)"] end subgraph EconomicImpact ["经济与资本开支"] E["跨国供应链重构与去中心化"] F["天量资本开支投入再工业化 (Reindustrialization)"] end ExternalDrivers --> CorporateShift CorporateShift --> EconomicImpact ``` ### [19:42](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=1182s) 英特尔案例与产业政策的长期黏性 乔什·布朗（Josh Brown）提出了一个最为典型的样本——英特尔（Intel）。此前，英特尔宣布放弃完全依赖中国台湾代工、立志成为立足北美本土的晶圆制造厂，并在亚利桑那州大举建造微芯片工厂，这一举措曾遭到全市场的冷嘲热讽。投资者对其估值疯狂压缩，“估值乘数暴跌了近 20 倍，市盈率几乎降至个位数，股价走势惨淡得像一家传统汽车制造巨头”，公司随后更是陷入亏损泥潭。 然而，在经历疫情断供与高通胀洗礼后，从白宫决策层到整个行业都幡然醒悟：高端芯片决不能单纯依赖单一海外代工基地，必须把制造留在美国本土。英特尔顺势迎来彻底重估，并在搭上 AI 浪潮后股价从约 15 美元一路狂飙至 80 美元，市值出现巨大反弹。 泽扎斯顺着这一话题点明，英特尔正是产业政策延续性的极佳范例：美国政府通过拜登任内两党合作签署的《芯片与科学法案》（CHIPS Act）直接入股并大力补贴英特尔，这在历史上是一次重大的政策转向。针对市场过度关注的 2026 中期选举，泽扎斯认为选举结果往往是“噪音大于信号”，因为无论谁掌控国会，底层核心的产业政策与保护主义取向都具有极强的延续性。当下的宏观格局正由两大引擎驱动：一是像 AI 这样带来巨额资本开支（Capex）的“外生技术冲击”，二是由产业政策驱动的、耗费数万亿美元重构全球商业体系的长期进程。虽然从传统经济学家的视角看，以意识形态和地缘安全重构贸易“在经济学上并非最优解”，但这种策略在政治选票上屡试不爽。 - **英特尔估值修复路径**：从自建本土晶圆厂被讥讽、估值暴跌至个位数惨遭亏损，到借助《芯片与科学法案》支持与 AI 顺风，股价实现从 15 美元到 80 美元的惊人逆转。 - **产业政策两党共识**：无论白宫易主还是国会中期选举轮替，以供应链本土化和安全优先为导向的产业政策展现出高度一致的长期黏性。 ### [22:11](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=1331s) 从预测市场到多变量博弈的投资盲区 迈克尔·巴特尼克（Michael Batnick）指出，投资者往往很难直接从政策推导到投资标的，这种高难度促成了 Polymarket 等预测市场的爆发。面对纷繁复杂的地缘政治，投资者即便精准猜对了政策，也猜不透市场参与者会如何定价；与其费尽心思去琢磨“民主党拿下众议院买什么股票”或“CPI 超预期是买黄金还是卖黄金”，不如直接在预测市场上对“事件本身是与否”进行点对点押注。 泽扎斯回应称，摩根士丹利内部确实将预测市场作为评估胜率分布的有效工具，用来弥补民调可能存在的系统性误差和过度虚假的精确度，但在利用预测市场做投资对冲时却极其复杂，因为投资决策永远是一个极其繁复的多变量方程（Multi-variable problem）。例如关税政策从预期走向现实的过程中，必须经过经济学家测算有效税率、输入通胀预测模型、推演美联储货币政策反馈，最终层层渗透到消费行业分析师和各类微观资产估值模型中，根本不可能套用“某种政策必然导致某种资产涨跌”的刻板剧本。 乔什·布朗随即补充了一个生动的思想实验：如果在上一年 1 月向华尔街预告，两个月后总统会站在白宫玫瑰园摆出展板，对各国列出骇人听闻的极高关税，随后哪怕因股市大跌而妥协回调，平均有效关税最终仍将推高到 15% 甚至更高。如果市场在 1 月完全相信了这个剧本，所有分析师的第一反应必然是集体大幅下调工业股盈利预期，并大幅压缩标普 500 指数的估值倍数。然而实际发生的情况却是相反——AI 叙事带来了爆发式的盈利扩张，市场估值整整抬升了一整年。这有力证明了宏观投资的终极困境：即使你百分之百笃定政策本身的走向，市场在复杂多变量交织下的定价路径依然可能与直觉背道而驰。 ### [26:47](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=1607s) 跨资产协作文化与大语言模型的局限 讨论从宏观经济预测的复杂性切入。去年许多华尔街同行在关税或政策风声一出时便立刻做出衰退预测，而摩根士丹利（Morgan Stanley）在首席全球经济学家塞思·卡彭特（Seth Carpenter）及其经济团队的带领下并未跟风。他们认为这绝非单变量问题，除了谈判本身的巨大不确定性，市场上还存在大量相互对冲的宏观平衡力量。 ```mermaid flowchart TD accTitle: 复杂宏观研究分析路径对比 accDescr: 对比传统大模型与人类跨资产跨领域研究团队在处理复杂多变量问题上的能力差异。 subgraph LLM["大语言模型（LLM）"] L1["宏观多变量问题"] --> L2["单一或通用提示词"] L2 --> L3["处理跨资产/跨区域关联困难"] L3 --> L4["成功率低 / 缺乏深度"] end subgraph MS["摩根士丹利协作研究范式"] M1["宏观多变量问题"] --> M2["跨资产与跨学科团队协同"] M2 --> M3["融合多变量与宏观对冲因素"] M3 --> M4["阿尔法收益与客户高粘性互动"] end ``` 迈克尔·泽扎斯（Michael Zezas）指出，摩根士丹利推崇跨资产、跨区域的深度协作，因为影响任何单一资产类别的本质上都是多变量问题。阅读数据与客户互动指标证明，这类跨学科整合报告的客户参与度要比普通报告高出一个标准差以上。更重要的是，这恰恰是大语言模型（LLM）目前极难处理的任务。即便未来可能出现更高级的多智能体协同（agent orchestrations），但目前如果直接把一个包含大量变量的大型宏观主题丢给通用模型去推导，成功率极低。大摩内部及外部优化专家的测试均印证了这一点。因此，深耕多变量跨资产研究不仅在 AI 时代构筑了核心阿尔法（alpha），也成为了不可替代的客户服务壁垒。 - 跨学科、跨资产的研究报告在客户互动度指标上普遍超出常规报告一个标准差以上。 - 目前的大模型在面对缺乏明确路径、多变量交织的宏观大主题时，推理成功率相当有限。 --- ### [28:41](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=1721s) 地缘政治恐慌与市场脱节：导弹在空中时买入 话题转向当前市场中极其诡异的割裂现象：政策不确定性指数（Policy Uncertainty Index）与波动率指数（VIX）之间出现了罕见的巨大剪刀差。通常情况下两者同向联动，但眼下即便地缘危机接二连三，股市却波澜不惊。回顾过去三十年，市场常年流传着“一旦轰炸伊朗、霍尔木兹海峡被封锁，全球经济就会彻底熄火”的终极黑天鹅设想；然而真实冲突爆发时，VIX 可能也就冲到 50 维持了五分钟，随后股市照样反弹。 乔希·布朗（Downtown Josh Brown）引述了纽约证券交易所传奇交易员阿特·卡辛（Art Cashin）的经典轶事。20 世纪 60 年代初卡辛刚入行第一周，古巴导弹危机爆发，场内疯传“古巴导弹已经升空”。年轻的卡辛惊慌失措地四处抛售，老板一把揪住他的衣领质问：“你在干什么？”卡辛喊道：“导弹都在天上了，我在卖出啊！”老板厉声喝止：“不，导弹在天上的时候你必须买入！因为如果导弹真的砸下来，一切都完蛋了，钱没有任何意义；但如果导弹没砸中，所有的钱都会在这一刻被赚走！” 泽扎斯表示，这种思维看似玩世不恭，实则蕴含深刻的投资博弈智慧： - 面对毁灭性的尾部风险（如第三次世界大战），理性的定价策略反而是赋予其极低的发生概率。 - 2017 至 2018 年朝鲜频繁试射导弹期间，韩国投资者向大摩急切索要对冲报告，研究员私下只能打趣道：“这种情况下我们能研究什么？难道建议你去挖个防空洞吗？” - 历届美国政府对经济和股市都抱有现实的敏感度。尽管白宫不会机械地设定“标普指数跌几个百分点就立刻改变政策”，但决策层与美联储显然时刻盯着 200 日均线，没人愿意仅仅因为嘴上放狠话而亲手引爆一场熊市。 - 投资者预期摩擦终会以妥协收场，加之企业盈利预期不断上调，构成了股市托底的现实支撑。 --- ### [32:45](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=1965s) 多极世界崛起与欧洲、日本的“再工业化”机遇 全球格局正步入大多数现代交易员未曾经历的新常态：多极化世界（Multipolar World）。欧洲和日本战后数十年来首次意识到，不能再把防务与产业彻底寄托于美国，也不能单方面依赖廉价的外部供应链（如中国或俄罗斯）。从俄乌冲突到中东局势升级，外部冲击正迫使这些发达经济体告别二战以来的防御与工业停滞状态。 乔希·布朗强调，德国、法国、意大利绝不能沦为纯粹的“度假旅游胜地”，它们必须重新建立自主可控的重工业与防务供应链。恐惧与危机感往往能唤醒沉睡已久的“动物精神”（animal spirits）。历史上一再证明，为应对冲突而启动的扩产备战，往往会伴随着波澜壮阔的资本开支大牛市。 ```mermaid flowchart LR accTitle: 欧洲与日本再工业化驱动逻辑 accDescr: 展示地缘安全威胁如何促使欧日摆脱二战后停滞，走向资本开支与股票重估。 subgraph Input["外部催化剂"] A["地缘冲突与安全威胁"] end subgraph Europe["欧洲重工业重估"] B["打破外部依赖"] --> C["防务与制造业回流"] C --> D["重资产企业叠加 AI 提效"] end subgraph Japan["日本内生动能"] E["公司治理改革深化"] --> F["良性通胀回归与民族意识"] F --> G["50%居民现金资产再配置"] end Input --> Europe Input --> Japan ``` 泽扎斯认为，在资本开支驱动增长的初期阶段，这在经济学上几乎是“纯代数运算”的确定性拉动——只要真金白银砸向产能，短期经济增长就会被直接推升，至于终极投资回报率（ROI）那是后期才需面对的问题。针对“除美国外的发达市场（Developed ex-US）是否具备更高权益投资价值”，大摩给出了肯定判断，并指出了欧日各自独特的重估逻辑： - **欧洲**：核心在于制造业重构与防务开支激增。摩根士丹利欧洲股票策略师玛丽娜（Marina）提出，欧洲传统企业普遍偏向重资产（heavy-asset），这使其成为“隐性的 AI 受益者”——在这类物理重工业中引入 AI 带来的边际生产率提升空间，远大于轻资产公司。 - **日本**：内生动能更为强劲。在再工业化浪潮之外，日本国内早已叠加了深度的公司治理改革、良性温和通胀的抬头，以及庞大家庭资产负债表中高达 50% 现金储备潜在的入市重配空间。 ### [36:14](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=2174s) 区域市场对比与前所未有的 AI 资本开支周期 讨论从欧洲和日本散户投资者的风险厌恶倾向切入。欧洲目前约有 7 万亿美元的资金沉淀在低息银行存款中，整整一代人都习惯了这种略显僵化的市场状态，缺乏做出改变的动力。随后，话题迅速转向全球最受关注的核心主题——人工智能（AI）的资本开支（CapEx）。主持人感慨，眼下如果不谈 AI，几乎就错过了所有人职业生涯中可能见证的最大规模资本支出周期。坊间传闻未来十年相关开支可能高达 50 万亿美元，而根据摩根士丹利（Morgan Stanley）自身的研究预测，仅在 2025 年至 2028 年这短短几年内，全球数据中心相关的资本开支就将达到惊人的 2.9 万亿美元。 ```mermaid flowchart LR accTitle: AI 资本支出传导与制约瓶颈 accDescr: 展示从超大规模云厂商的资本开支，经过关键制约瓶颈，最终驱动物理建设与未来生产力飞跃的路径。 subgraph S1["投入端"] A["超大规模厂商 CapEx<br/>(2025-2028 达 2.9 万亿美元)"] end subgraph S2["制约瓶颈"] B["电力与能源供应"] C["专业劳动力与政策"] end subgraph S3["产出阶段"] D["当前：实体建设与基建<br/>(直接拉动约 1/4 GDP 增长)"] E["未来：全行业生产力繁荣<br/>(企业 EBITDA 显著分化)"] end A --> B & C B & C --> D D --> E ``` 面对市场普遍关心的问题——“我们现在究竟处于 AI 投资周期的第几局？投资者是否已经错过了主升浪？”——摩根士丹利副全球研究主管 Michael Zezas 给出了明确判断：当前仍处于非常早期的阶段，市场上依然有大量的投资机会。他特别提到摩根士丹利的主题研究负责人 Stephen Byrd。Stephen 早期长期担任能源与公用事业分析师，原本负责一个相对沉闷的传统板块，但由于所有 AI 算力最终都需要电力驱动，他顺理成章地转型成为解决“AI 能否获得充足能源”的顶尖专家。就各大超大规模云服务商（hyperscalers）规划的算力蓝图而言，建设才刚刚起步。无论超大规模厂商最终获取的投资回报率（ROI）是高是低，他们都在不惜代价地推进建设。而电力短缺、专业劳工不足以及潜在的政策约束，反而拖慢了实际推进节奏，使得这场军备竞赛比原本预期的拉得更长、处于更早期的阶段。 --- ### [38:53](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=2333s) 数据中心的草根阻力与 AI 的公关难题 建设不仅受制于物理层面的能源瓶颈，还遭遇了社会和政策层面的阻力。弗吉尼亚州之所以成为当今全球数据中心之都，有着独特的历史渊源——早年间美国在线（AOL）在这里起步，最初的数据中心就是为了服务它而设立的。但眼下，包括弗吉尼亚、加利福尼亚和俄克拉荷马等州在内，地方社区因环保担忧或单纯出于对 AI 的排斥情绪，开始出现草根阶层的抵制声音。 Michael Zezas 直言，AI 概念目前确实面临着严重的“公关问题（PR problem）”。这种反对声浪并不受党派政治立场的局限，红州和蓝州都有选民提出反对意见，只是各自挑剔的理由有所不同。虽然目前的抵制规模尚不足以真正撼动整体基础设施的铺设，但已被列为必须密切监控的关键风险因素。 与此同时，Zezas 强调了 AI 当前对现实经济最直接的贡献：摩根士丹利宏观经济团队测算，今年美国约有四分之一的 GDP 增长直接来自当前 AI 基础设施的实体建设。这种拉动与未来的抽象技术红利无关，而是实打实的土木施工、管道铺设和机房建筑工程所带来的经济增量。 --- ### [40:28](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=2428s) 生产力飞跃的兑现路径与微观财务验证 谈及市场翘首以盼的“生产力大爆发”，Michael Zezas 认为其到来的大方向毋庸置疑，但具体节奏很难精准预测。他以自己在摩根士丹利研究部推进 AI 工具落地的亲身体验为例：摩根士丹利研究部全球主管 Katie Huberty 长期以来极具前瞻性，全力推动将 AI 深度融入日常工作流与未来研究体系。但现实表明，个人在家中使用 AI 工具或初创团队从零起步时，效率提升可能立竿见影甚至令人震撼；然而在大型机构内部，将 AI 嵌入盘根错节的既有工作流与人员协同体系中，实现工业化改造往往“说起来容易做起来难”，需要漫长的磨合周期。 针对“技术成果是否应该首先在科技巨头内部兑现为效率”的疑问，Zezas 指出微观层面的财务改善其实已经显现。摩根士丹利全球覆盖约 3,400 只股票，分析师团队定期通过多维度方法进行摸排： - 将覆盖的公司系统性划分为 AI 赋能者（Enablers）、AI 采纳者（Adopters）及最终受益者； - 调研各家公司管理层对技术应用的推进状况； - 利用自然语言技术抓取公司财报电话会议纪要（Earnings Transcripts），量化统计公开报告 AI 具体落地成果的企业比例。 调研结果显示，在 MSCI 全球指数（MSCI World）中，能拿出具体数据证明 AI 带来实质收益的公司占比已从去年的约 20% 稳步提升至 30% 以上。更为关键的是，在摩根士丹利归类为“AI 采纳者”的企业中，其 2024 至 2025 年的平均息税折旧摊销前利润（EBITDA）扩张速度，比非采纳者快了整整一倍（2x）。虽然这尚未大规模体现在宏观层面的生产力数据上，但在企业微观财务报表中，AI 赋能带来的效益已经逐步得到印证。 --- ### [43:42](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=2622s) 盈利预测的巨大鸿沟与市场广度分歧 节目现场展示了摩根士丹利提供的图表 20.2，揭示了不同板块间盈利预期的剧烈分化： | 资产类别 / 观察口径 | 2027 年一致预期每股收益（EPS）变动幅度 | |:--- |:--- | | **AI 基础设施相关股票 (AI Infrastructure Stocks)** | **+42%** | | **标普 500 指数整体 (S&P 500)** | **+10%** | | **剔除 AI 基础设施后的标普 500 指数 (S&P 500 Ex-AI Infrastructure)** | **仅 +3%** | 这种悬殊的差距意味着非 AI 板块的利润增速必须迎头赶上，否则大盘的牛市基础将显得极其脆弱。Zezas 借用摩根士丹利首席美股策略师 Mike Wilson 的视角分析称，当前市场在“实际企业盈利范围的扩大”与“股票价格广度（Broadening）的扩散”之间出现了背离。从大盘指数来看标普 500 指数年内涨幅不错，但上涨个股的占比其实仅处于历史平均水平附近。市场各方一直在讨论牛市行情的下一波浪潮是否会离开超大规模云厂商，向非 AI 板块全面扩散，但这种业绩预期的再平衡至今尚未在足够广泛的行业中落地。 ### [45:43](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=2743s) 美股行情扩散：科技股等权重指数大幅跑赢 讨论首先聚焦于当前美股涨势的驱动结构：究竟需要广大 AI 应用端企业相继超预期盈利，还是仅凭少数集中的“AI 领头羊”就能维持整个市场的繁荣？Michael Zezas 指出，摩根士丹利并不认为市场“必须”依赖某些条件才能活下去，但从其每年的年中与年度展望流程（集合所有策略师和经济学家的共识）来看，他们已经连续两个周期坚定看好美股行情的“广度扩散”（broadening out），以及海外权益资产的长期趋势。在强劲的资本支出（CapEx）驱动下，这种结构性增长周期将在未来数年成为常态，而债券则更多扮演提供票息和分散风险的角色。 Josh Brown 紧接着补充了极具冲击力的一组实时市场数据：科技巨头“壮丽七雄”（Magnificent 7）年内表现其实相当平淡，整体仅上涨约 70 个基点，Meta、微软走势停滞，英伟达在尝试突破后也遭遇打压；然而，标普 500 等权重指数年内却涨了 11%。更夸张的是纳斯达克 100 指数，在权重占比高达 40% 的 Mag 7 几乎零涨幅的情况下，整指年内依然大涨 19%。Josh 进一步调出等权重科技股 ETF（RSPT）的数据，显示其年内涨幅竟然高达 39%。“科技板块内部的扩散已经实实在在地发生了，”这种巨大的表现分化不仅佐证了行情并未熄火，而且表明市场的上涨动力正在向更广泛的企业渗透。 - **Mag 7 与等权重板块分化**：Mag 7 年内基本走平（仅微涨 70 个基点），但科技等权重 ETF（RSPT）年内暴涨 39%。 - **主要指数表现**：标普 500 等权重指数年内上涨 11%，纳斯达克 100 指数在巨头停滞的情况下依然累涨 19%。 --- ### [48:12](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=2892s) 2026 年中期选举：政策连续性与市场噪音 转入政治与宏观政策议题，主持人询问机构投资者和对冲基金对即将到来的中期选举有何担忧，以及“政策将保持连续性”的预判是否可能被颠覆。Michael Zezas 分析，除非民主党或共和党取得压倒性大胜，否则这一判断很难出错，但目前没有任何预测支持这种极端结果。预测市场数据显示，民主党有约 80% 的概率拿下众议院，而同时掌控参众两院的概率仅为 50/50 左右。参议院的归属基本取决于缅因州等少数关键州的席位争夺。 ```mermaid flowchart TD accTitle: 2026 Midterm Election Legislative Gridlock accDescr: 解释中期选举无论结果如何都会陷入立法僵局，无法形成实质性政策冲击 subgraph Outcomes["选举席位推演"] A["民主党微弱优势掌控参众两院<br/>(参院仅 50-53 席)"] B["两党分治国会"] end subgraph Constraints["宪政与程序制约"] C["缺乏打破阻挠议事(Filibuster)的超多数"] D["无法推翻总统否决权(Veto)"] end subgraph MarketImpact["市场与政策结论"] E["无法推翻过往减税法案"] F["难以收回总统关税与贸易裁量权"] G["立法陷入僵局：政策延续性占主导"] end A --> C A --> D B --> C C --> E D --> E D --> F E --> G F --> G ``` Zezas 深入拆解了立法逻辑：即使民主党在参议院拿下 50 到 53 个席位，他们也无法达到打破阻挠议事（Filibuster）所需的绝对多数，更不可能获得推翻总统否决权所需的三分之二票数。因此，诸如客户担心的“前几年通过的税改法案是否会被废除”根本无法成真——民主党即便有心改革，也缺乏足够的选票和白宫控制权。同理，国会过去授予总统的关税与贸易自由裁量权，也绝无可能在短期内被收回。对于可能出现的市场波动，Zezas 提醒道：如果选举夜出现意外结果导致小盘股或税改受益板块出现情绪化的膝跳式下挫（例如担忧三年后会加税），投资者最明智的做法就是“直接反手逆向交易”（fade that reaction），因为离下一届大选还有两年，“在金融市场里两年漫长得宛如一个世纪”。 --- ### [51:05](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=3065s) 地缘危机与霍尔木兹海峡：AI 赋能下的“三级投研” 当被问及今年机构客户提出过最棘手、最出人意料的问题时，Michael Zezas 提到了伊朗局势升级引发的地缘危机。当时投资者关心的不再是抽象的“霍尔木兹海峡会不会成为瓶颈”，而是极其硬核的技术细节：该海峡究竟每天能通行多少原油？各方船只在何种条件下可以通航？海湾周边国家的利益权衡与实际介入程度如何？大家现场打趣道，普通人甚至军方平时可能只知道海峡在地图右边，但当伊朗感到自身面临“存亡危机”时，封锁海峡便成了其极具杀伤力的底牌（上一次严重袭击油轮还要追溯到 1980 年代的“两伊油轮战”）。 为了回答这一复杂课题，摩根士丹利位于伦敦的顶级大宗商品分析师 Martin Rats 联合团队展现了极强的专业深度。他们直接利用公司内部部署的前沿 AI 工具，快速调用并处理海量开源情报（OSINT），在极短时间内精细绘制并模拟出了不同极端情境下海峡的实际通航能力与原油流量。 这一案例顺理成章地引出了关于“AI 对金融市场研究产生何种深远冲击”的讨论。Zezas 指出，过去被视作华尔街“聪明人”核心壁垒的能力，很快就会变成人人皆有的“入场标配”（table stakes）： - **传统复杂建模普及化**：以往在金融模型上叠加蒙特卡洛模拟（Monte Carlo analysis），需要专门聘请统计学家并耗费超算资源；如今任何分析师都能借助 AI 随手完成。 - **超大规模文本透视（“三级研究” / Level 3 Research）**：以往想要对比 40,000 份债券契约文件（bond indentures）的细微条款差异并将其映射到信用利差上，需要消耗庞大的人力团队；而如今通过 AI 大模型和精准的提示词（prompts），单凭一块高性能算力芯片就能彻底穿透底层文档。投研范式已从单纯拼人头、拼信息差，全面进化为驾驭机器与前沿工具的深度竞争。 ### [55:01](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=3301s) 研报垃圾化隐忧与人类判断力：番茄与水果沙拉的比喻 针对生成式 AI 是否会抹平大型研究机构的核心竞争优势，嘉宾们展开了深刻的探讨。现场指出，尽管 AI 降低了获取大量信息和初步分析的门槛，但也带来了严重的副作用——就像 Instagram 等社交媒体上充斥着低质的“AI 垃圾内容”（AI slop）一样，金融市场上也可能会泛滥成灾地出现“研报垃圾”（research slop）。Josh Brown 尖锐地剖析了算法模型的底层激励机制：如今全球的每家企业本质上都变成了争夺用户留存的“注意力业务”（engagement business），甚至连严肃的分析工具都在向 Netflix 和 TikTok 的逻辑靠拢。机器的本能是迎合人类并防止被关闭，因此当用户反复追问时，AI 最终总会过度拟合（overfit），拼凑出一个“听起来完全正确”但实际脱离现实的迎合性答案。 Michael Zezas 强调，工具的普及永远无法取代人类在实战中沉淀的认知与审慎考量。就像金融界人人都听说过蒙特卡洛模拟（Monte Carlo analysis），但听过概念与知道在现实中如何精准建模完全是两码事。谈到知识与经验的区别时，主持人和嘉宾默契地引用了一句经典谚语进行拆解：“知识是知道番茄在植物学上属于水果，而智慧是明白绝不能把它切进水果沙拉里。”在海量信息与算法噪音铺天盖地的时代，投资者依然需要依靠深厚的人类专家网络来穿透迷雾，否则极易在追逐虚假叙事的过程中犯下灾难性的判断失误。 - **AI 研报泡沫机制**：算法模型为了讨好用户并延长互动时长，倾向于持续修正输出，直到给出听觉上极其合理但可能存在严重逻辑缺陷的结论。 - **专业机构的护城河**：面对未来全员武装 AI 的市场环境，追逐错误叙事而被证伪的周期性洗牌仍会发生，具备实操“智慧”（Wisdom）与“审慎判断”（Judgment）的专家网络依然是过滤噪音的核心壁垒。 --- ### [56:52](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=3412s) 快问快答：AI 时代的就业前景是乐观还是阵痛？ 节目进入摩根士丹利专题“快问快答”环节，主持人抛出的首个核心问题聚焦于 AI 时代的劳动力市场：大摩究竟是倾向于乐观派，认为伴随新技术而来的新增就业岗位会及时出现并平稳承接；还是认为年轻一代求职者将不得不面对数年残酷的就业断档期（air pocket）？ Michael Zezas 给出的官方基调是“除非被事实证伪，否则整体保持乐观”（optimistic until proven otherwise），同时“睁大双眼保持高度警惕”（eyes wide open）。他透露，摩根士丹利的经济学家团队此前对宏观经济数据进行了极其详尽的解构，绘制了一张劳动力技能与容易被 AI 颠覆的工作岗位之间的映射图谱，并实时监控不同人群的失业率变动。从法律、会计等白领初级岗位到青年整体就业率，目前的真实数据表明，尽管如果“眯起眼睛仔细审视”可能会察觉到青年失业率存在极细微的异动，但整体而言尚无确凿证据表明结构性失业潮已经降临。因此大摩当前宏观立场依然偏向乐观，但保持密切追踪。 - **大摩研究框架**：通过构建“可替代技能与经济数据映射模型”，持续对初级专业服务（如法律、审计）和年轻从业群体的失业数据进行敏感度监控。 - **现状结论**：劳动力市场目前并未出现明显的失业断层，技术转型的平滑度暂时好于悲观预期。 --- ### [58:01](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=3481s) 软件板块遭遇“SaaS 末日”：自建代码能取代企业级软件吗？ 主持人将话题转向了当年美股表现最为惨淡、甚至被戏称为“SaaS 末日”（SaaS apocalypse）的企业级软件板块。面对众多百亿美元级 SaaS 巨头市值大幅缩水的现状，市场最大的疑问在于：这些软件公司究竟是在痛苦中沉淀长期价值，还是未来的商业变现之路注定越走越窄？ Michael Zezas 坦言两方面兼而有之：基本面上这些软件公司的日子确实会更加艰难，但市场的恐慌抛售显然存在过度反应。大摩坚决不认同“未来每家企业内部的每个人都会用 AI 替自己编写专属软件”的极端论调。大厂企业级软件之所以具备护城河，在于其能够将复杂的业务逻辑、合规规范与工作流“工业化”交付；这种标准化与体系化能力，绝非单个员工凭借几个大语言模型提示词就能自行构建并安全运行的。 ```mermaid flowchart TD accTitle: Enterprise SaaS Moat vs In-House AI accDescr: 比较企业级SaaS平台的工业化交付优势与内部员工纯自建AI工具的局限性 subgraph InternalRisks["内部个人 / AI 纯自建模式"] A["个人调用 LLM 编写独立代码"] --> B["缺乏跨部门数据打通"] B --> C["合规、风控与企业级集成薄弱"] end subgraph SaaSPlatform["企业级 SaaS 厂商护城河"] D["底层业务流工业化封装"] --> E["原生结合私有高价值数据"] E --> F["合规管控、稳定性与规模化落地"] end InternalRisks -.->|"难以取代"| SaaSPlatform ``` 谈及摩根士丹利自身的研发预算分配走向时，Josh Brown 追问未来三年大摩的研究预算会继续投向传统企业级 SaaS，还是会被算力账单以及向 Anthropic 等基础模型公司购买 Token 蚕食。Zezas 笑着承认这已经超出了快问快答的简洁范畴：未来整体技术预算在总支出中的占比必然会进一步上升，但由首席技术官 Katie 团队推进的实际方案并非单纯采购 Token，核心关键在于让正确的企业私有数据与高价值的应用场景严密咬合，Token 消费仅是底层消耗品，而非技术升级的全部。 - **市场情绪评估**：市场对“AI 杀死 SaaS”的恐慌被过度放大，成熟软件厂商的工业化封装与流程集成能力依然是坚固的壁垒。 - **IT 预算结构演变**：企业技术采购重点正在从单纯的基础模型 Token 计费，转向深度融合私有专有数据与业务场景的企业级综合解决方案。 --- ### [01:00:23](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=3623s) 凯文·沃什执掌美联储与摩根士丹利宏观展望 话题随后切入人事与货币政策：摩根士丹利的机构客户们对凯文·沃什（Kevin Warsh）走马上任掌控新一届美联储持何种态度？Zezas 用了一个看似平淡却含义深远的词来概括——“挺好的”（Fine）。他指出，虽然债券市场尚未经历重大危机考验，但客户的平稳情绪本身就是极大的利好。相比于此前坊间流传的一些推崇非正统经济学学说、或者被外界怀疑一通白宫电话就能左右货币政策的极端政治化候选人，传统出身的沃什显然让华尔街吞下了一颗定心丸。 Zezas 进一步分享了大摩首席经济学家 Seth 团队的专业视角：退一步讲，即便白宫真的安插了一位政治意图鲜明的主席，市场关于美联储独立性将彻底瓦解的恐慌也是站不住脚的。美联储利率决议是由整个联邦公开市场委员会（FOMC）集体投票决定的，绝非主席个人能够一手遮天。固然极端的沟通风格可能打乱市场预期，但若拉长历史周期，实际政策利率跟随通胀与经济增长波动的宏观规律不会发生根本性偏离。如今提名一位符合建制派预期的传统央行家，市场自然免去了这层不必要的压力测试。 在全期节目的尾声，Josh Brown 对摩根士丹利的宏观立场进行了全面总结：大摩虽然不是华尔街上盲目狂热的多头，但对绝大多数资产类别的态度依然相当积极且具有建设性。只要企业盈利增长稳步兑现、中东地缘政治局势不再出现恶性升级、基准利率路径维持温和预期，全年市场大概率将以稳健的态势收官。Zezas 对此深表赞同，并重申支撑这一判断的底层支柱正是由 AI 驱动的资本开支浪潮（CapEx）以及全球向多极化格局过渡的长期历史趋势。 在欢快的寒暄与玩笑声中（主持人们调侃嘉宾要不要明天早上 7:30 再来录一期），Zezas 向听众介绍了摩根士丹利智库（Morgan Stanley Institute），听众可以通过其官网和通讯获取汇聚了全公司顶级智慧的思想领导力研究成果，本期高密度的宏观对谈就此圆满落下帷幕。 - **美联储新架构预期**：市场对沃什的评价保持平和稳健；FOMC 的多边委员会机制是抵御单点行政干预的制度性防火墙。 - **大摩核心收官主线**：以多极化世界转型与 AI 相关资本支出（CapEx）为代表的长期结构性趋势，为宏观经济与风险资产提供了坚实的底部支撑。 ## Notable quotes > [02:30](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=150s) 责任在她那里，她说了算。 — Michael Zezas > [26:47](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=1607s) 这绝非单变量问题，除了谈判本身的巨大不确定性，市场上还存在大量相互对冲的宏观平衡力量。 — Michael Zezas > [28:41](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=1721s) 不，导弹在天上的时候你必须买入！因为如果导弹真的砸下来，一切都完蛋了，钱没有任何意义；但如果导弹没砸中，所有的钱都会在这一刻被赚走！ — Josh Brown > [28:41](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=1721s) 在这种情况下我们能研究什么？难道建议你去挖个防空洞吗？ — Michael Zezas > [45:43](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=2743s) 科技板块内部的扩散已经实实在在地发生了。 — Josh Brown > [48:12](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=2892s) 在金融市场里两年漫长得宛如一个世纪。 — Michael Zezas > [55:01](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=3301s) 知识是知道番茄在植物学上属于水果，而智慧是明白绝不能把它切进水果沙拉里。 — Josh Brown / Michael Zezas > [56:52](https://www.youtube.com/watch?v=6OMtpw-TPVs&t=3412s) 除非被事实证伪，否则整体保持乐观，同时睁大双眼保持高度警惕。 — Michael Zezas ## Mentioned Michael Batnick, Downtown Josh Brown, Michael Zezas, Katie Huberty, James Gorman, Seth Carpenter, Stephen Byrd, Mike Wilson, Kevin Warsh, Martin Rats, Meredith Whitney, Art Cashin, Morgan Stanley, Public, E*TRADE, Polymarket, Intel, Anthropic, MBIA, Ambac, CHIPS Act, Lawrence of Arabia, Disclosure Day ## Video description On episode 247 of The Compound and Friends, ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Michael Batnick⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ and ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Downtown Josh Brown⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ are joined by ⁠Michael Zezas⁠, Deputy Global Head of Research at Morgan Stanley, to discuss: AI capex, data centers, productivity gains, prediction markets, the 2026 midterms, the Fed, enterprise software, and why policy calls are so difficult to translate directly into investment outcomes. This episode is sponsored by Public. To learn more about Public, visit ⁠https://public.com/Compound ►00:00 - Cold Open ►04:34 - Intro ►05:52 - Morgan Stanley, Research, and The Markets ►16:52 - Regime Change and Market Calls ►32:45 - Investing in a Multipolar World ►36:35 - AI, Capex, and Global Projections ►48:09 - The Midterms ►56:48 - Lightning Round 👕 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. For important disclosures, stock price charts and equity rating histories regarding companies that are the subject of this public appearance, please see the Morgan Stanley Research Disclosure Website at https://www.morganstanley.com/researchdisclosures, or contact your investment representative or Morgan Stanley Research at 1585 Broadway, (Attention: Research Management), New York, NY, 10036 USA. For valuation methodology and risks associated with any price targets referenced in this research report, please contact the Client Support Team as follows: US/Canada +1 800 303-2495; Hong Kong +852 2848-5999; Latin America +1 718 754-5444 (U.S.); London +44 (0)20-7425-8169; Singapore +65 6834-6860; Sydney +61 (0)2-9770-1505; Tokyo +81 (0)3-6836-9000. Alternatively, you may contact your investment representative or Morgan Stanley Research at 1585 Broadway, (Attention: Research Management), New York, NY 10036 USA. #thecompound #investing #stockmarket #stocks [Source](https://www.youtube.com/watch?v=6OMtpw-TPVs)

[2:30] Michael Zezas: 责任在她那里，她说了算。 [26:47] Michael Zezas: 这绝非单变量问题，除了谈判本身的巨大不确定性，市场上还存在大量相互对冲的宏观平衡力量。 [28:41] Josh Brown: 不，导弹在天上的时候你必须买入！因为如果导弹真的砸下来，一切都完蛋了，钱没有任何意义；但如果导弹没砸中，所有的钱都会在这一刻被赚走！ [28:41] Michael Zezas: 在这种情况下我们能研究什么？难道建议你去挖个防空洞吗？ [45:43] Josh Brown: 科技板块内部的扩散已经实实在在地发生了。 [48:12] Michael Zezas: 在金融市场里两年漫长得宛如一个世纪。 [55:01] Josh Brown / Michael Zezas: 知识是知道番茄在植物学上属于水果，而智慧是明白绝不能把它切进水果沙拉里。 [56:52] Michael Zezas: 除非被事实证伪，否则整体保持乐观，同时睁大双眼保持高度警惕。

[14:28] when you're explaining, you're losing. [29:52] Art Cashin's former boss: You buy when the missiles are in the air. 'Cuz if they hit, who cares? If they don't hit, that's where all the money is made. [29:52] Michael Zezas: What do you want me to tell you? Dig a hole? [32:45] Downtown Josh Brown: Sometimes what certain country markets need is a little bit of fear and a little bit of paranoia for the animal spirits to get going [51:05] Michael Zezas: I literally knew it was on the map... and I knew that it's just the contingency planning playbook if Iran gets attacked. [53:43] Michael Zezas: Knowledge is knowing that a tomato is a fruit; wisdom is knowing not to put it in a fruit salad. [62:26] Michael Zezas: We're bringing together all the thought leadership across the firm for all types of financial decisions.

[0:00] First of all, how you doing? [0:01] >> We're going to have so much fun. [0:03] >> All right. [0:04] >> I promise. [clears throat] [0:04] >> I like that. [0:05] >> Thank you for coming. [0:06] >> Yeah, thanks for having me. [0:07] >> So, you know who my close friends with? [0:09] It's not Graham, it's Colin. But I knew [0:11] he came from somebody. [0:12] >> Colin Ross? [0:13] >> Yeah. [0:13] >> Oh, I love Colin Ross. How do you know [0:15] >> at college. [0:16] >> Oh, you kidding me? [0:17] >> Yeah. [0:18] But I just watched on the on the on the [0:20] train home. [0:20] >> Yeah. Lawrence of Arabia? [0:23] >> Are you going to listen to me? [0:24] >> Yes. [0:25] >> It's 4 hours. You're going to have to [0:26] watch it like a TV show. [0:27] >> I'll watch it like a TV show. I'll watch [0:28] like 40 minutes of it. [0:30] >> It's 4 hours, but I they digitally [0:32] restored it [0:33] for Netflix. [0:34] >> Is that Is that with Jake Gyllenhaal? [0:36] >> Yeah, no. [0:37] >> [laughter] [0:37] >> It's from 1962. [0:39] Everyone's dead. [0:41] Some of Some of the best actors who ever [0:43] lived. Like literally who ever lived. [0:45] And uh [0:47] I'll I'll let you watch it. I won't tell [0:48] you anymore. [0:49] >> Thank you Thank you for spoiling Freaky [0:50] Friday, too. [0:51] >> Thank you for spoiling Disclosure Day. [0:53] >> sound like something I was going to put [0:54] any time [0:55] >> Lawrence of Lawrence of Arabia, [0:56] digitally remastered for for Netflix. [0:59] >> So, Michael, I'm [0:59] >> I'm a I'm a big movie fan, and to the [1:02] extent that I kind of like to raw dog [1:03] everything. Meaning, don't I don't want [1:05] to know anything. Like you can't avoid [1:07] the You can't avoid the trailer for [1:08] Disclosure Day, but I don't want to hear [1:10] I don't want to look at the reviews or [1:11] Rotten Tomatoes. I just want to I just [1:13] want to with with clean eyes. Cuz I'm [1:15] very I'm easily manipulate or [1:17] influenced. [1:18] The bar set too high, the bar set too [1:19] low. So, Josh gave a full review [1:21] yesterday without me asking for [1:23] >> Day or it wasn't Disclosure Day. [1:26] >> He's like, "The aliens weren't there." I [1:27] don't know, I'm not going to ruin it for [1:28] anybody else. Like you did it for me. [1:29] >> I made you do that. I I assumed you saw [1:31] it. [1:32] >> You gave us a full review. [1:33] >> No, I was talking to you like you saw it [1:34] already. [1:35] >> You gave me a full review. [1:36] >> I said seven seven out of 10. [1:38] >> [clears throat] [1:38] >> Well, listen, I do most of my movie [1:40] watching on an airplane these days, so [1:42] >> Yeah, I understand. Huge airplane movie [1:43] guy. [1:44] >> Yeah. I saw I saw this in the theater. [1:46] It didn't need to be seen in the [1:47] theater. Yeah. But I figured it's [1:48] Spielberg, it's aliens. I'm going to go [1:50] in the theater. It was like, "Eh, I [1:52] could have watched it at [1:52] >> There was that whole like gorilla [1:53] marketing campaign about it, right? That [1:55] it like it secretly was actually going [1:57] to disclose that there were aliens? [1:59] >> Yes. Um [2:00] >> That didn't happen? [2:00] >> Wait, did we get did we get a did we did [2:03] we decide? [2:04] >> Oh, yes, you were right. [2:05] >> I'm right. [2:06] >> He he said like Jesus. [2:08] >> I said that's well, that's like that's [2:09] what Josh said. [2:09] >> My instinct was Jesus. [2:11] >> Yeah, that's right. [2:11] >> Yeah. [2:12] >> That's right. [2:12] >> I thought the Z might be silent. I [2:13] wasn't sure. [2:14] >> No. No, it's a Greek name and you know [2:16] what? We say it wrong. Like Jesus is the [2:18] way my family says it, but it should be [2:20] something more like Sezos or something [2:22] like that. [2:22] >> Oh, really? [2:22] >> Yeah. [2:23] >> That's too bad. [2:24] >> Well, exactly. Like I [2:25] I don't even say it the right way. [2:27] >> It's Jesus, rhymes with Jesus. [2:29] >> I had another question. [2:30] >> Yeah. [2:31] >> [applause] [2:31] >> Who are the deputy? [2:32] >> Yeah. [2:32] >> Who's the who's the head of research? [2:34] >> Katie Huberty. [2:35] >> I didn't know that. [2:36] >> Yeah. [2:36] >> All right. [2:37] >> All right. [2:37] >> Okay. [2:37] >> Buck stops with her. [2:38] >> Yeah. All right. I did I did know that. [2:40] She's great. [2:42] All right. How are we looking? We doing [2:43] good? [2:44] >> Yep. [2:44] >> Such an important man. [2:45] >> You remember the name? [2:46] >> Johnny Yogo. Only in my own mind. [2:48] >> Are you watching the World Cup? [2:49] >> What what office do you work in? [2:51] >> 1585. [2:52] >> Okay. [2:52] >> Times Square, yeah. [2:53] >> All right. [2:53] >> Yeah. [2:54] >> How long How long have you been in that [2:55] office? [2:56] >> [snorts] [2:57] >> Um [2:58] that office since 2010. I've been with [3:01] Morgan Stanley for 19 years. Part of [3:03] that was in 522 Fifth Avenue when we [3:05] started that place. [3:06] >> Okay. You interact with the wealth guys? [3:08] >> Oh, yeah. I got to start in wealth [3:10] actually. [3:10] >> Yeah. [3:11] >> Yeah, I started we had this SMA platform [3:13] in wealth management and that's how I [3:14] got started in munis and corporates [3:16] there. Yep. [3:17] >> Okay. All right. Cool. We're going to [3:18] have I promise we're going to have some [3:20] fun today. [3:20] >> Good. [3:21] >> All right. [3:21] >> Good. [3:22] >> We're good to go? [3:24] All right. Who the heck is Nicole? [3:27] >> It's 212-279-2427. [3:30] >> [music] [3:35] [music] [3:40] >> It feels like there are two types of [3:41] investing [music] platforms right now. [3:43] You've got the legacy brokerages that [3:45] look like they were designed in 1997. [3:47] [music] [3:49] And then you've got the new wave that [3:50] looked at investing and thought, "You [3:52] know what this needs? Sports betting." [3:54] Neither seemed like a great place to [3:56] build your wealth. That's where Public [3:58] comes in. It's the modern [3:59] [music and singing] investing platform [4:01] for those who take it seriously. Stocks, [4:03] options, [music] bonds, crypto, they [4:05] have it all. And the energy you're not [4:07] spending on building a casino, it's [4:09] going into AI. [4:11] >> [music] [4:11] >> Public is the only investing platform [4:13] where you can create agents that can [4:15] monitor the market, manage your cash, [4:18] [music] and execute your trades. Just [4:20] enter a prompt, approve the workflow, [4:22] and put your agent to work. Go to [4:25] public.com/compound [4:26] to get started. That's [4:28] public.com/compound. [4:30] Paid for by Public Investing. Full [4:32] disclosures in podcast description. [4:35] All right. [4:37] Ladies and gentlemen, welcome to What [4:38] did we say? 247? Episode 247 of The [4:41] Compound and Friends. [4:43] My name is Downtown Josh Brown. Here [4:46] with my co-host, Mr. Michael Batnick. [4:48] >> Hello, hello. [4:50] >> Michael, I swear this gets more serious [4:52] as it goes on. [4:54] Michael Michael Zezas is joining us [4:56] today. We're super excited about it. [4:58] First-time guest on the show. Michael is [5:00] the deputy global head of research for [5:02] Morgan Stanley, [5:03] where he joined the firm in 2007 [music] [5:06] as a credit analyst and portfolio [5:08] manager. In 2010, he became the head of [5:11] municipal credit strategy, followed by [5:13] head of US public policy research [5:16] [music] in 2016. [5:19] He has been a member of Institutional [5:20] Investor's All-America Fixed Income [5:22] Research Team since 2013, [music] [5:25] and by Smith's Research and Ratings as [5:28] an All-Star First Team Member from 2014 [5:32] to 2021. [5:34] Of course, in 2022, he fell off. [5:36] >> Yeah. [5:37] >> [cheering] [5:37] >> It was a tough year. [5:38] >> Things haven't been as good since then, [5:39] but we are going to revive his career [5:41] today. [5:42] >> Let's do it. [5:42] >> All right. Dude, thank you for being [5:44] here. I really appreciate it. [5:45] Uh so, Morgan Stanley, they seem to be [5:47] doing okay. [5:48] >> Yeah, not bad. [5:49] >> Okay. [5:49] >> Not bad. [5:50] >> [laughter] [5:50] >> As a shareholder, I'm pretty happy. [5:52] >> I think that I think Gorman was [5:54] incredible. And uh, you don't have to [5:56] agree, but I'm sure you agree, but [5:57] >> I do. [5:58] >> Okay. [5:59] Just he figured out if we're going to [6:01] build a true wealth franchise, what's [6:04] the most important thing to the people [6:05] working there? [6:06] There have to be clients. There has to [6:08] be demand. [6:09] >> Seems it seems pretty logical. [6:11] >> Seems logical, but you'd be amazed. So, [6:13] a lot of large firms have wealth [6:15] managers who have to go out and find [6:17] their own clients, or sitting around [6:19] waiting for something to happen. And [6:20] Morgan Stanley, very brilliantly, I [6:22] think, acquires E*TRADE. They buy the [6:25] business that's now Morgan Stanley at [6:27] Work, and they just buy these um, [6:29] funnels to push more people toward the [6:32] wealth management people. And as a [6:34] result, it's I think it's probably the [6:36] fastest-growing Wall Street wealth [6:38] management business. There I mean, do [6:40] you know if it is, or I think it is? [6:42] >> Uh, that sounds right. I don't know [6:43] exactly, but I mean [6:45] >> 20 trillion doll- [6:46] >> Yeah, that's it in terms of assets. [6:47] >> I'm confident in my [6:48] >> 20 trillion. [6:49] >> Okay. [6:49] >> All right. [6:50] >> Okay, that's right. [6:50] >> Thanks. No, and and I [6:52] >> I know you played a big part in that, [6:54] but I'm just saying [6:54] >> Yeah, well, like a little part for the [6:56] few years I was in it. But, listen, I [6:57] like to think that as someone who um, [7:00] you know, covered munis for a decade, [7:03] spent a lot of time with uh, our wealth [7:05] management field effort, talking to [7:07] clients across the spectrum. Um, listen, [7:11] I'm biased because I'm in research. I [7:12] like to think some of our intellectual [7:14] capital is something [7:15] >> do with that, too. [7:16] >> Our ability for us to leverage that and [7:17] to turn it into solutions for clients, [7:19] so. [7:19] >> Okay. [7:20] >> Still doing plenty of that. [7:21] >> Well, whatever you guys are doing, it [7:22] worked. [7:23] >> All right. [7:23] >> How did you go from muni coverage, which [7:25] is like one of the ultimate zoom-ins on [7:28] Wall Street, to like this big-picture [7:31] global macro dude? [7:32] >> Yeah. Well, um, [7:35] I think it has a lot to do with [7:37] how munis changed in yeah, like like [7:40] 2008, 2009, right? So, before that and [7:44] before the financial crisis, everything [7:46] muni land was wrapped by the monoline [7:48] insurers. And so, you basically had one [7:50] credit counterparty. It was the [7:51] monolines. Everything was triple A. Um [7:54] when that went away, all of a sudden you [7:57] needed credit analysts to start looking [7:59] at everything that was underlying, [8:01] right? And that was sort of dovetailed [8:03] with everything that started going on [8:05] with the sovereign debt crisis in [8:06] Europe. There were all sorts of [8:07] narratives that came into the market. [8:08] They were like the Meredith Whitney [8:10] >> were in the you were in the on the muni [8:12] side when that Meredith Whitney scare [8:14] >> Oh, yeah. Oh, yeah. I mean, and I I [8:16] remember talking to hedge funds at that [8:18] time trying to explain the difference [8:20] between Greece needing a rollover debt [8:23] that was like 150% of its revenue in any [8:25] given year versus California, which was [8:27] paying like 5% in P&I, right? And so, [8:31] getting into those discussions, you [8:33] realize the public policy aspect of [8:36] municipal credit was really really [8:38] important to clients at a micro level. [8:41] Um and then, you know, in terms of [8:44] defining returns, I mean, it it's fixed [8:46] income product, right? So, like you were [8:47] going to have to have a view about all [8:49] the different sort of intersection of [8:50] things that were driving treasury [8:52] markets, not just monetary policy, [8:54] fiscal policy, tax policy. So, [8:57] I sort of adopted this approach as a [9:00] muni strategist that was [9:03] that was uh I was trying to be somewhat [9:05] different than our peers, right? Like [9:06] muni strategy was this field where like [9:09] everybody else had been around for like [9:10] 20 years and I was like the low cost [9:12] option that got brought in, right? I was [9:15] like, "Okay, well, instead of doing the [9:16] thing where I'm going to like calculate [9:17] carry and roll down at every tenor and [9:19] say this looks a little bit better than [9:21] that one, like let's [9:22] >> the hospitals, not the roads. [9:24] >> Yeah. Right. [9:25] >> Right. [9:25] That's my research. [9:26] >> Yeah. Like let's look at this more from [9:28] a macro perspective. Not just like [9:30] what's going on in the world, but also [9:32] like what are the big secular trends [9:34] that govern the market, govern its [9:36] behavior, establish priors that way, [9:39] have a framework for setting up [9:40] portfolios, and then iterate on top of [9:43] that. And the heavy dose of that had to [9:45] be [9:46] tax policy [9:48] um during President Obama's first term, [9:49] a lot of his healthcare policy. [9:52] Um and so I just I got in the practice [9:54] of having to be effectively like a DC [9:57] watcher or DC analyst. [9:59] And I also have a master's degree in [10:00] public policy, so it was kind of a [10:02] natural place for me to say, "Okay, this [10:04] this may or may not become law, and if [10:05] it does, I can tell you down to like the [10:07] balance sheet and income statement level [10:08] what it's going to mean for this [10:10] municipality versus that." [10:12] In 2016, I was [10:15] uh I don't know lucky or unlucky enough [10:16] to [10:17] uh be the person that um the folks [10:20] around the research department back then [10:21] tapped to say, "Listen, [10:23] clients are asking a lot of questions [10:25] about what happens if Donald Trump wins. [10:27] He's not going to win, but like let's [10:28] try and answer these questions on a [10:30] multi-asset basis." [10:32] So I did that. It was a good stretch [10:33] assignment, and then of course when he [10:35] wins, [10:36] the questions never stopped. It became [10:37] this kind of permanent research focus [10:40] area. [10:40] >> you become like the public policy [10:41] research guy within the bank. [10:44] >> Yeah, and that not by design, by [10:46] accident. A lot like most of my good [10:47] career moves have been by accident as [10:49] opposed to design. Yeah. [10:51] >> So all right, but so everyone in the [10:52] industry is getting those questions. [10:54] Yeah. So uh but you were at that time at [10:57] one of the largest investment banks in [10:59] the world. Yeah. So I would imagine the [11:02] volume is higher. [11:03] >> Yeah. [11:03] >> But also the level at which people are [11:06] making investment decisions based on [11:08] what you say is significant. [11:10] >> Yeah. [11:11] >> You were also in uncharted waters. We [11:13] had never seen a first presidential term [11:15] like that. [11:16] >> Yeah. [11:16] >> So how did you how did you navigate that [11:18] and actually give people something that [11:20] was meaningful, but not commit to [11:22] something where you couldn't have [11:23] possibly known the outcome? [11:25] >> Well, I think the first thing is [11:27] we came into that environment with a [11:29] really good collaborative culture as a [11:31] research department. So, [11:34] I could define with a decent degree of [11:38] conviction on my own like what I thought [11:40] a Trump presidency was going to mean in [11:41] terms of this policy will happen, this [11:43] one won't, and here are kind of ones [11:45] that are uncertain. But in any case, you [11:48] know, now let's like map out the [11:49] fundamental impacts. [11:51] And we had enough people with enough [11:54] collaborative interest to do that that [11:55] we were able to put together road maps [11:57] into the election in 2016 and then [11:59] throughout 2017 or like, "Okay, if tax [12:01] policy happens this way versus that way, [12:02] it's going to mean, you know, this [12:04] fundamentally for this sector, let's [12:06] compare it to valuation and let's right? [12:07] So, there was enough [12:09] people to collaborate with in a good [12:11] culture that we could work on a [12:13] cross-asset, cross-region basis to to [12:15] map all this stuff out. So, really like [12:17] credit to my colleagues and to the [12:19] culture of Morgan Stanley. [12:21] Um [12:22] you know, from there, it did start [12:24] feeling like we were like chasing our [12:25] tail a little bit. And this is where [12:28] I always try to, if I can, like carve [12:30] out time to be like, "Can we step back [12:32] and put together [12:34] like a secular theme or like a framework [12:36] that we can we think it's going to be [12:38] durable for years?" [12:40] And so, it became pretty obvious in 2017 [12:43] that this was not business as usual. And [12:46] a lot of clients were asking us the [12:48] question, [12:49] "Hey, like when is this going back to [12:51] business as usual? Okay, like you know, [12:53] we were starting to flirt with tariffs, [12:54] those income in the plane in 2018." [12:57] Um [12:58] but, you know, uh President Trump [13:00] calling CEOs into the White House, like [13:02] all sorts of stuff that just [13:04] you know, now seems like quaint, but [13:06] then just seemed wild. [13:07] >> These were to your point, these were [13:09] bombs being dropped on the market on a [13:11] daily. He would tweet at the CEO of [13:13] Pfizer [13:14] >> Right. [13:15] >> and voice his displeasure over drug [13:16] pricing, and Pfizer would lose $50 [13:19] billion in market cap. Like these were [13:21] not trivial things. They Now we look [13:23] back and we laugh because [13:26] you know, none of those bombs actually [13:28] detonated and did any lasting damage, [13:30] but like in that moment market [13:32] participants were like shocked. Like, [13:35] oh, wait a minute. The White House is [13:37] now a player on the on the chessboard. [13:39] >> Yeah, and then [13:40] >> And is the queen, frankly. [13:41] >> Right. So, we'd be taking the those [13:43] style of questions from hedge fund [13:45] managers, but then, you know, official [13:47] institutions overseas, central banks be [13:49] like, okay, just like tell us how long [13:51] we need to like white knuckle this until [13:53] things go back to to to normal. [13:56] >> And when is it going to be 2005 again? [13:58] >> Right. [13:58] >> Yeah. [13:58] >> And [13:59] so and and my and my view was and the [14:01] way we ended up putting a lot of work [14:03] around it was like things are not going [14:04] back to normal. In fact, um what you're [14:07] seeing [14:09] is kind of what American voters have [14:12] elected. Maybe they haven't elected the [14:14] tactical application of it, but when you [14:17] kind of look back, a lot of things had [14:18] changed underneath the surface. [14:21] Um it's like like like Jake Sullivan who [14:23] was national security advisor under [14:25] Biden talks about as the new Washington [14:27] consensus. It's sort of very clear in [14:28] hindsight. But, um [14:31] the the old Washington consensus was [14:33] like, you know, whether you're [14:34] Republican or Democrat, going back to [14:36] the early '80s all the way, you know, [14:38] maybe let's call it until the financial [14:40] crisis, [14:41] you agreed on certain things. You're [14:43] like, free trade's a good idea. The [14:45] government should not be involved in the [14:47] economy. Multilateral um institutions [14:50] are good, right? Soft power in the US is [14:53] a good idea. Um [14:55] and now, if you look at like voter [14:57] attitudes, they're like, we don't really [14:59] [clears throat] see the value in any of [15:01] that. Yeah. Um now, maybe they don't see [15:03] the value in like jacking tariffs up all [15:05] the way or going to war in the Middle [15:08] East, right? Or at least that's what the [15:09] polling would suggest, but they don't [15:12] see value in supporting some of those [15:15] principles that were kind of key to this [15:18] US-led [15:19] principles-based laissez-faire economics [15:22] international economic system that like [15:24] everyone gets [15:25] >> them. [15:26] >> It's right. Or at least they didn't see [15:27] those benefits, right? Like [15:29] >> right, it benefited but not in a direct [15:31] enough way where they said, [15:33] "Yes, I care about NATO or I care about [15:37] NAFTA or I think it's good that uh [15:40] China's in the WTO." Like [15:42] >> Yeah. [15:43] >> They either were very against those [15:44] things because they hadn't seen any real [15:47] wage growth in 15 years. [15:49] Um their communities had opioid [15:52] addiction problems and they had family [15:54] members who were sent into wars of [15:56] choice and they just looked around after [15:58] 15 years of that they were like, [16:00] "This isn't helping me. I'm not getting [16:02] ahead. The economy is leaving me behind. [16:04] >> Right. [16:04] >> I want to vote for volatility. I want [16:07] somebody to break everything." [16:08] >> Yeah, and so like you could see that [16:09] right in front of your face. [16:10] >> Right. [16:11] >> the counterargument is, [16:13] "Well, let me explain to you why free [16:15] trade has been really good relative to [16:18] like a past you didn't actually [16:20] >> actually experience. [16:21] >> Yeah, what's that saying when you're [16:23] you're explaining, you're losing, right? [16:25] Like so that's not to say that [16:29] voters have elected the optimal outcome [16:31] here, but it's what they've chosen, [16:33] right? [16:33] >> So you arrived at that in the 2010 late [16:36] 2010s like not only is this very [16:38] different, but also it's not going back. [16:41] >> Yeah. [16:41] >> It's probably going further. [16:42] >> Yeah, and so this was the genesis of our [16:45] whole multipolar world thematic. [16:48] >> So so brag about that because you were [16:49] right. This is this is I'll quote you to [16:51] you and then you can react. [16:53] From 2020, our call was and is that [16:56] there's been a regime change driven by [16:58] shifting voter preferences. [17:00] And with it comes the need for financial [17:02] decision makers of all kinds to invest [17:05] with the understanding that the best [17:07] benefits of globalization were behind us [17:09] and that geopolitical fragility is on [17:11] the rise. Okay, great. We're 6 years [17:14] hence. Great call. [17:15] >> Yeah. [17:16] >> Okay. [17:16] >> Yeah, I mean, uh now [17:19] identifying this early [17:21] was I think important, but it was also [17:23] sort of complicated at the time to [17:25] understand when these trends would [17:27] actually manifest in like real, you [17:30] know, tangible investable themes, right? [17:32] Because [17:33] you know, you I think we said with a [17:36] high degree of confidence like trade [17:37] barriers are going up, they're not [17:38] coming down, right? And uh you know, [17:41] obviously Biden was president in between [17:44] Trump's two terms. And um we always run [17:48] investor survey ahead of the election. [17:50] In the investor survey ahead of 2020, um [17:53] most investors said, yeah, that Trump [17:55] winning would be better for markets than [17:57] Biden. But if you answered Biden [17:59] >> Right. [17:59] >> was better for markets, you were also [18:01] very likely to answer a separate [18:03] question on the survey that said, oh, [18:05] well, Biden's going to take tariffs [18:06] down. [18:06] >> to roll them back. [18:07] >> Right. [18:07] >> And then we didn't. [18:08] >> Which was wrong. Because it's it's First [18:11] of all, like if there's it's not as if [18:13] the Democrats and Biden wanted to go [18:15] back to some free trade norm with China. [18:18] Um in fact, they they kind of wanted to [18:20] do the opposite, but their tactics were [18:22] very different. So, of course you [18:23] wouldn't bring down tariffs proactively [18:25] if you're trying to negotiate with [18:27] China. [18:28] Um so, the point that like the direction [18:30] of travel on trade barriers, whether [18:31] they're going to be tariff or non-tariff [18:33] driven, was like was pretty obvious. The [18:35] direction of travel on industrial [18:37] policy, which is to say the government [18:39] getting more involved, was pretty [18:42] obvious. [18:43] So, you just needed time to work out in [18:45] terms of manifesting into, okay, [18:49] money being spent on, for lack of a [18:51] better term, reindustrialization. [18:53] >> Reshoring, onshoring, whatever you want [18:54] to call it. [18:55] >> Yeah, and then of course with the US [18:57] sort of exercising a different stature [18:59] on the global stage, um and some of [19:03] these geopolitical conflicts coming [19:04] together, Russia-Ukraine, now the US and [19:07] Iran sort of highlighting some very [19:09] critical choke points, [19:11] you know, corporate America C-suites [19:13] across the globe kind of get the [19:14] message. Like we we you know, we can no [19:16] longer sort of optimize for [19:17] globalization. [19:19] We need what our investment bankers call [19:22] an anti-fragile strategy. So you're [19:24] starting to get these supply chain [19:26] shifts to work around um those choke [19:28] points, to work around those other [19:30] sensitivities. Some of them self-imposed [19:32] by the US or trade barriers to say we [19:34] don't want people doing this kind of [19:35] business in China. Um [19:38] and [19:39] takes a long time and a lot of money to [19:41] retool and re-equip. [19:42] >> stock better exemplifies what you're [19:43] talking about than Intel. [19:45] So this is a company that was ridiculed [19:47] for having adopted the strategy of we're [19:50] going to be the fab that's based in [19:52] North America, not Taiwan. [19:53] >> Yeah. [19:54] >> And we are going to build uh microchip [19:57] plants in Arizona, because of course. [20:00] And the stock price [20:02] I think uh derates, I don't know, 20 [20:05] multiple points. It's a It's almost like [20:07] a single digit and almost trades like an [20:08] automaker. And then they start losing [20:10] money. [20:11] And all of a sudden [20:14] there's there's this idea driven in part [20:17] by what you're talking about, White [20:19] House directly, but also sort of the [20:21] needs of the industry post uh lockdowns [20:24] and post inflation is like we can't rely [20:27] on just chips from, you know, Taiwan. We [20:31] need to actually have stuff that gets [20:33] made here. And there's a total rethink [20:35] on Intel, and then they join the AI [20:37] parade. [20:38] >> Yeah. [20:38] >> And then you see a stock go from, I [20:40] don't know where it was, $15 to $80. [20:43] >> Yeah. [20:44] >> Um and that's a lot of market cap along [20:45] the way. You think that's like sort of [20:47] what we're what we're getting at from [20:48] the investment standpoint? [20:49] >> example here of some of the policy [20:50] continuity, right? Because actually the [20:53] government took a stake. [20:54] >> Where did that [20:55] >> Right. They took a stake in Intel, which [20:56] is a big departure. [20:57] >> And where did that come from? That came [20:58] from the CHIPS Act, which was something [21:00] that was completed on a bipartisan basis [21:02] when Joe Biden was president. [21:04] >> Mhm. [21:04] >> So there's actually a fair amount of [21:06] continuity across these industrial [21:08] policy choices. And I'd argue, [21:12] you know, like we're starting to get [21:13] questions about the midterms and midterm [21:14] elections. [21:16] I I think that's going to be a lot more [21:18] noise than signal any any outcome there. [21:20] There's not a lot of policy change that [21:21] comes out of different um midterm [21:23] election outcomes. Policy continuity is [21:26] actually far more important. Um so, that [21:30] and like we haven't even started talking [21:32] about AI yet. You you sort of have like [21:34] AI is kind of this like exogenous [21:36] technological shock driving tons of [21:38] capex. [21:40] Um and then I'd say a largely like a [21:43] policy-driven, [21:44] okay, impetus to just well, we're going [21:47] to spend a lot of money [21:48] to rewire global commerce in a way that [21:51] we think suits our values better. Is [21:54] that economically optimal or not? I [21:56] think an economist would probably say [21:57] not, but you know, almost doesn't even [21:59] matter what the answer to that question [22:00] is. But like [22:02] you know, it's good for winning [22:02] elections. [22:03] >> So, that's that's why it's the [22:05] >> policy. [22:05] >> Well, and either way we're we're going [22:07] to spend trillions of dollars trying to [22:09] figure it out. [22:11] >> Michael, I think it's it's really hard [22:12] to draw a direct line from policy to [22:14] investing [22:15] all of the time. You say it's noisy now. [22:17] I think it's always noisy. And I think [22:19] maybe this has [22:21] an inkling to do with the rise of the [22:22] prediction markets [22:24] because there is so much geopolitical [22:25] uncertainty. And even if you get a [22:28] policy call right or whatever, you then [22:31] have to nail what how investors are [22:33] going to react to that. [22:34] >> Yeah. [22:34] >> And now it's like, well, F all that. [22:36] Let's just go to straight to the source. [22:38] Like, will this happen yes or no? I [22:40] don't have to bet on this and then [22:41] industrial [22:42] >> Democrats take the house or not? I don't [22:44] need to figure out what stocks to buy. [22:45] >> Right. Will CPI come in Will CPI come in [22:48] hot or cold? And then I have to figure [22:50] out well, man, are stocks extended or [22:52] therefore gold or do I sell gold? [22:54] >> Yeah, no, listen, if you're talking [22:56] about a sort of First of all, we really [22:59] like prediction markets as another [23:02] probability assessment of what's going [23:03] to happen. I don't think I have any edge [23:05] on predicting election outcomes. Um we, [23:10] you know, we know the uncertainties [23:11] implied by polls. Okay, I think [23:13] sometimes people put a lot of false [23:14] precision on polls. If you treat them [23:16] with a healthy amount of uncertainty [23:18] that they deserve and you understand [23:19] polling errors, um none of the election [23:22] outcomes over the last, you know, few [23:24] cycles would have been all that [23:25] surprising. [23:27] Prediction markets, um I think are just [23:31] like another very useful comparison to [23:34] the polls and they help you with that. [23:36] Um yeah, I mean, to the extent you want [23:38] to express it like a pure view on a [23:39] certain outcome, [23:40] like that's that matters. I mean, the [23:42] the the world that I travel in tends to [23:45] be like, "Well, can I use prediction [23:46] markets to hedge a certain outcome?" [23:49] >> Yeah. [23:49] >> And that's far more complicated because [23:51] >> to tie tying policy [23:53] >> Yeah. [23:53] >> And sorry to cut you up, but tying [23:54] policy to investing, [23:57] that's not easy. It's intellectually [23:58] very stimulating, but that's a [23:59] >> Yeah. [24:00] >> Well, because my thought is this, like [24:02] if you get if you understand the [24:05] long-term implications of a policy [24:07] choice, [24:08] which which can't be done in a vacuum. [24:10] Like it can't. Like we, you know, if we [24:12] think, you know, [24:14] like [snorts] going into last year, [24:16] okay, we knew tariffs were going higher. [24:18] Were they going to go as high as like [24:20] the average effective rate of like 30%, [24:22] which is kind of where, you know, things [24:25] were going on liberation day. Were they [24:26] going to settle somewhere in between? [24:28] You know, we had a base case written [24:29] down of somewhere in between. I think [24:31] the average effective rate is sitting in [24:33] like the the sort of like 12 to 15% [24:36] range now. That's roughly where we were [24:38] coming into last year, though. [24:40] We had it kind of kicking in more later. [24:42] Um well, then we've got to Okay, so [24:46] we can apply that to goods. We've got to [24:48] work with our economists. We're going to [24:49] feed it through in the inflation [24:50] forecast. That's going to feed into [24:52] central bank policy. That's going to [24:54] feed into every single macro market um [24:57] uh estimate that we have. Um and [25:01] obviously to the ex- we're dealing with, [25:03] you know, the the consumer analyst and [25:05] anyone else who's affected by it to make [25:07] sure that like okay, we we think we [25:09] understand the policy. Here's multiple [25:10] scenarios for it. [25:12] Now, you have to contextualize that [25:14] within all the other sort of variables [25:17] driving the price of your asset class. [25:19] So, in that sense like there are [25:21] we don't really put together playbooks [25:23] like this policy equals, you know, X or [25:26] Y. [25:26] >> cuz there's other things happening. [25:28] >> It's it's a it's a multi-variable [25:30] problem. [25:30] >> a great example. So, last So, last year [25:33] if if in January I had told you 2 months [25:36] from now the president's going to sit in [25:37] the Rose Garden [25:38] and put this [25:39] um poster board up with specific tariffs [25:43] for every country and some of them are [25:44] going to be insane levels that nobody [25:46] has in their estimates. [25:48] >> Yeah. [25:49] >> And then he's going to reverse because [25:50] the stock market's going to have a [25:51] negative reaction, but still [25:53] they're going to be 15% let's say. [25:55] >> Yeah. [25:56] >> And in some industries higher and [25:58] they're still going to be like all sorts [25:59] of rhetoric about tariffs for the rest [26:01] of the year. So, I told you this last [26:03] January, it comes to pass in March. [26:07] But had I given not you, the colloquial [26:09] you, had I given the street [26:12] definitive this is what's going to [26:13] happen. [26:14] >> Right. The tariffs would be 5x higher, [26:15] yeah. [26:16] >> So, what but what would everyone do? [26:18] They would immediately drop their [26:19] earnings estimates, especially for [26:20] industrials. [26:22] Um and they would probably lower the [26:24] multiple that we'd be willing to pay for [26:26] stocks. And both in both cases they [26:28] would have been drastically wrong [26:30] directionally. [26:31] >> Yeah. [26:31] >> So, you had huge earnings earnings [26:33] expansion, which is the AI story. And we [26:35] had I think multiples rising throughout [26:38] the course of last year. [26:39] >> Yeah. [26:39] >> So, [26:40] that's why it's to Michael's question [26:42] like [26:44] even if you nail the outcome of a [26:47] policy. [26:47] >> Yeah. It has to be considered, yeah, [26:50] it's a multi-variable problem. [26:51] >> other things. [26:52] >> And like in that like last year like [26:54] much credit to Seth Carpenter and the [26:57] economics team. [26:59] A lot of our competitors went straight [27:00] to recession call and [27:02] >> Within within like a day. [27:03] >> Yeah, and like we're not there. There's [27:05] a lot of other things counterbalancing [27:07] this in in addition to the uncertainty [27:09] around the negotiations themselves. So [27:12] it has to be dealt with with in that way [27:14] and and listen like [27:16] I like to think that we [27:18] there's a reason that we try to have [27:20] such a collaborative culture and such a [27:22] like cross asset cross region culture. [27:24] It's because [27:25] any of these problems affecting any [27:27] individual asset class coverage area is [27:30] a multi-variable problem. And that's how [27:33] we put ourselves out to clients. I'll [27:35] say based on like our readership data, [27:37] our follow on engagement data. That's [27:39] what they like. Like if you're if you're [27:40] doing collaborative work, you're going [27:42] to get more than one standard deviation [27:43] engagement than other folks. [27:45] And [27:47] um [27:47] it's it's it's you know, it's also by [27:50] the way like what LLMs have a hard time [27:52] doing. So we're just going to lean more [27:54] and more into that. [27:55] >> what LLMs have a hard time doing? [27:56] >> Doing basically like with a simple [27:59] >> Cross asset or or cross discipline. [28:01] >> Cross discipline, cross asset and like [28:03] listen the the maybe there are some [28:05] agent orchestrations in the future which [28:06] might help with that but [28:08] you know, certainly right now giving a [28:11] generalist [28:13] a big thematic question that has a lot [28:15] of variable inputs into it and saying [28:17] figure this out working with an LLM [28:20] We we get really low success rate of [28:22] that. We've been testing that a lot [28:23] ourselves. We talked to LLM optimization [28:25] specialist. That seems to be low success [28:28] rate. So it's something that even coming [28:29] into the AI environment, we leaned into [28:32] culturally cuz we thought there was a [28:33] lot of alpha there. [28:34] Um [28:36] but now we think it's just increasingly [28:39] a part of the strategy of how we have to [28:40] serve clients. [28:41] >> So we'll get to the stuff obviously. [28:43] But we mentioned the tower stuff and [28:46] even if you could have predicted [28:47] perfectly, not to believe at this point, [28:49] but the markets move so fast these days. [28:51] >> Yeah. [28:52] >> So, so fast. So, you could be thinking [28:53] about the implications and the market [28:54] could be pricing in, you know, whatever [28:56] is going to eventually happen. John Key [28:58] phone chart four, please. [28:59] We're looking at the policy uncertainty [29:01] index versus the VIX and there is a [29:04] gigantic spread between the two. [29:06] Uh and usually they're they're [29:08] directionally correlated. But there [29:10] seems to be such a heightened level of [29:13] geopolitical uncertainty that is not [29:15] really translating into equity markets. [29:17] And you saw this during during the war [29:20] where everyone's scratching their head. [29:22] It's like, wait a minute. This was [29:23] supposed to be the black swan event. [29:25] Everybody was supposed to was expecting, [29:27] if this happens, all right, S&P limit [29:29] down, crude up, whatever, doubled. And [29:32] [snorts] [29:32] >> Like for 30 years, someday we're going [29:34] to bomb Iran and they're going to close [29:35] the Strait of Hormuz and it's going to [29:36] be lights out for the global economy. [29:39] >> And I I think the VIX got to 50. [29:41] >> Not maybe. I don't [29:42] >> Maybe for 5 minutes. [29:43] >> transmission mechanism between elevated [29:47] geopolitical uncertainty and equities [29:50] and fixed income and other instruments, [29:52] it's a really weird disconnect. [29:54] >> It is and [29:56] I've got theories. I don't have any [29:57] great explanations, but I think one of [29:59] them is I think when you're [30:01] when you're dealing with something that [30:03] taken to its logical end, it it is [30:06] potentially existential for the economy [30:08] and markets, maybe sometimes your [30:11] optimal strategy is to put a really low [30:13] probability on it actually going all the [30:15] way through. You know, it's World War [30:17] III. [30:17] >> Right. Right. [30:19] >> What's the story that Art Cashin said [30:20] about about the bombs? [30:22] >> Yeah, uh Cashin from the New York Stock [30:23] Exchange told the story about his first [30:25] week at work. It's uh early '60s [30:28] and the rumor going around the floor was [30:30] the [30:30] the missiles from Cuba were in the air. [30:33] So, he's running around trying to get [30:34] sell orders off and his boss grabs him [30:36] by the collar and says, "Art, what are [30:38] you doing?" He said, "The The missiles [30:40] are in the air. I'm I'm selling." [30:41] >> Yeah. [30:41] >> He said, "No, no, no, no, no. [30:43] You You buy when the missiles are in the [30:44] air. Cuz if they hit, who cares? If they [30:47] don't hit, that's where all the money is [30:49] made. [30:49] >> That I mean that's And there were no [30:50] missiles, thankfully. [30:51] >> Yeah, I mean I hate to be so like [30:53] cavalier about it, but like but there's [30:56] a lot of wisdom in that. I mean we [30:59] I think it was like [31:00] I can't remember it was 2017 or 2018 [31:02] when North Korea was testing a bunch of [31:04] missiles and there was a lot of rhetoric [31:06] between the US and North Korea going on [31:08] and [31:08] um you know, our Korean investors wanted [31:12] some research around that to help them [31:14] navigate and hedging and you're like [31:16] >> What do you want me to tell you? Like [31:18] research? [31:19] >> Yeah, right. Like yeah, so um [31:22] >> Dig a hole. [31:22] >> Right, right. Like it's it's Yeah, so I [31:25] I mean not to suggest that Strait of [31:27] Hormuz closure is nearly as existential [31:29] as that, but um [31:32] it's not irrational, in my view, to [31:34] presume that an administration which has [31:38] shown itself to be sensitive to economic [31:40] uh not perfectly sensitive, right? I'm [31:42] not one of those people who's like, oh [31:43] like [31:44] the White House is never is going to [31:46] react like with some of you get the [31:48] question like, what percentage down does [31:50] SPX have to be before the White House [31:52] reverses course? I'm like, it's not that [31:53] simple. But at least there is some [31:55] sensitivity [31:56] >> far off, either. [31:58] I think the market I think the White [32:00] House is aware of the 200-day moving [32:02] average. [32:03] >> Yeah. [32:03] >> I think the Fed is, too. [32:04] >> Yeah. [32:05] >> And I think anybody that suggests [32:06] otherwise is just not really paying that [32:09] close attention. But we seem to pull [32:12] rabbits out of hats. We get in the [32:14] vicinity of like an actual [32:16] down I don't think anyone wants to cause [32:18] a bear market. [32:19] >> Yeah. [32:20] >> Especially if they're just giving [32:21] speeches. [32:22] >> Yeah, I mean that's that's fair. I I I [32:24] guess the only thing I would say is it's [32:25] not irrational for for investors to [32:30] see positive outcomes around some of [32:33] these friction points. And then of [32:35] course they could be proven wrong over [32:37] time. [32:37] >> Well, they were right. I mean, the other [32:38] thing is earnings estimates kept rising [32:40] and 100% right. [32:42] >> Yeah. [32:42] >> Before we do the AI story, which we'll [32:45] finish with, I do want to get to uh [32:48] investing in a multipolar world world [32:50] and um [32:52] This is So, this is new for most people [32:54] who are in the market. They haven't They [32:56] haven't been alive at a time where all [32:59] of a sudden the Europeans and the [33:00] Japanese decide it's in their best [33:03] interests to not just leave everything [33:04] to America [33:05] and to start thinking about their own [33:08] ability to build [33:09] and mechanize and defend and to me, like [33:14] that's one of the bigger changes of this [33:16] era and it sort of I guess started with [33:19] the Ukrainian uh the the the conflict in [33:22] Ukraine [33:24] and then I guess Israel and Iran and it [33:26] just like every step forward seems to [33:30] make the numbers go higher. Let's put [33:31] this chart up. Total new investment in [33:33] re-industrialization in Europe. And he [33:35] was Not that this is all military, [33:37] obviously, but just like just this idea [33:41] everybody needs their own supply chains. [33:42] >> Yeah. [33:43] >> Um Germany, France, Italy can't just be [33:46] tourist destinations. Like, they have to [33:47] be industrialized countries again. And [33:50] they have to get serious about not [33:52] relying on China or the US. And that I [33:55] think there are huge investment [33:57] implications and I think they're [33:59] actually positive. [34:00] >> Yeah. [34:00] >> So, look at overseas stocks last year [34:03] had a great year and maybe sometimes [34:05] what certain country markets need is a [34:07] little bit of fear and a little bit of [34:09] paranoia for the animal spirits to get [34:12] going. [34:12] >> Yeah. [34:12] >> And uh it's I mean, it's not ahistorical [34:15] to suggest we've had some great bull [34:17] markets as a result of building for war. [34:21] It's just reality. [34:22] >> No, that's that's historically accurate. [34:25] There's no other way to say it, but I [34:26] mean [34:27] I'll say if you're going to spend on [34:29] capital and productive capacity whatever [34:32] it's catalyzed by uh you know at least [34:35] in the early stages it's basically [34:37] algebra in terms of how it drives [34:39] growth. You know there's always the [34:41] lingering questions and the the [34:42] questions will linger for AI also about [34:44] like what's the ultimate ROI on this but [34:47] I just don't think we're [34:49] we're not there yet. [34:49] >> these countries more inv- Does that make [34:52] Japan Does that make developed ex-US [34:55] more investable from an equity [34:57] perspective? [34:59] The fact that these countries are now [35:00] going through this sort of cathartic [35:02] moment where they're shaking off the [35:04] post-World War [35:06] II stasis and they're getting serious [35:08] now about re-industrializing. [35:10] >> We think so [35:11] but would like nuance it a bit [35:13] differently for like Europe versus Japan [35:15] or like they Europe is [35:18] about industrialization, defense and you [35:22] know Marina who's our our European [35:24] equity strategist has a good theory of [35:26] the case around [35:29] Europe being kind of a sneaky AI [35:32] beneficiary, right? Just because of how [35:34] you know asset heavy those businesses [35:36] are and and how like there could be some [35:38] some good productivity gains there. [35:41] >> It's not asset heavy, it's heavy asset. [35:43] He [35:43] he had an acronym [35:45] >> I'm sorry about [laughter] that. [35:46] >> We'll allow it. [35:47] >> Sorry. [35:47] >> We'll allow it. [35:48] >> Halo halo I got I know I know. [35:52] In Japan [35:54] um [35:55] you know there was already some stuff [35:56] going on before all this came into play, [35:58] right? There there was a lot of [35:59] corporate governance reform [36:01] >> um a little bit more nationalism than [36:04] we've seen in the last few decades. [36:06] >> like some healthy inflation creeping in [36:08] and like by the way you've got a [36:11] household base there that like 50% is [36:14] invested in deposits. I think the number [36:16] is about like $7 trillion [36:18] so um [36:19] >> risk risk averse [36:21] >> Right. [36:21] >> sort of sclerotic markets [36:23] >> there was [36:24] for a generation there was not really [36:27] a reason they had to think differently. [36:29] >> Right. So, um the short answer is yes, [36:32] so just like nuanced a little bit based [36:34] on each region. [36:35] >> Okay. Um I want to ask you about I want [36:38] to ask you about AI because [36:40] like [36:41] it's almost [clears throat] like to [36:42] discuss anything else misses probably [36:46] the biggest cycle of [36:48] CapEx that any of us will ever see. [36:50] >> Yeah. [36:51] >> If the numbers even get close to the [36:52] projections. [36:53] >> Yeah. [36:54] >> Is it like $50 trillion over the next 10 [36:56] years or whatever people are saying? [36:58] I mean, who knows, but [36:59] >> You guys have an estimate for $2.9 [37:01] trillion in global CapEx. [37:02] >> That's just for [37:03] >> 25 to 28. [37:04] >> Till 28. And that's just for the data [37:05] that's just for the data centers. [37:07] >> And that's before the robots even show [37:08] up. And that's before everything goes [37:10] autonomous. And that's before [37:12] uh [37:13] Mars. Like there are a lot of other [37:15] things that Okay. So, are we like in the [37:18] fourth inning of AI spend or is this [37:21] like Has the game even started yet? How [37:24] do you How do you talk to in groups of [37:26] investors [37:27] about whether or not they've missed a [37:29] lot of the the gains or like what's [37:31] still to come? Is there Is there room to [37:34] make new investments today? [37:35] >> Yeah, I think there's plenty of room. I [37:38] don't know if it's first inning or [37:39] fourth inning. It's early. [37:41] >> nobody really knows. [37:42] >> Right. Yeah. And I mean, so Steven Bird [37:44] runs our thematic research team. And [37:47] talk about people started their career [37:49] in one area and then they become [37:50] specialized in in really valuable in [37:52] another area. Um he was our energy and [37:55] utilities analyst for years. It's all [37:57] this very sleepy area, but gets pulled [37:59] into becoming an AI expert because well, [38:02] all this stuff's got to get powered. [38:04] Yeah. And like no one knows more than [38:05] him about whether or not we can power [38:06] this and how it's going to happen and [38:08] the step function scenario. [38:09] >> it. That's interesting. [38:10] >> Yeah, yeah, yeah. And he And he's great. [38:11] He runs all of our thematic research [38:12] now. Um [38:15] So, like the the the short answer is in [38:17] terms of the build-out, still pretty [38:19] early stages in terms of the compute [38:22] that the hyperscalers say that they want [38:24] to develop, right? And so, whether or [38:27] not you think the hyperscalers are going [38:28] to get, you know, super high ROI or [38:31] something that's good, but either way, [38:33] like they're they're they're going to [38:34] pay to develop a lot more of it. And the [38:37] sort of governors, the constraints [38:39] around that build out, which which are [38:41] energy and labor, and maybe a little bit [38:43] of policy, [38:44] um if anything sort of like is making [38:47] that a bit earlier stage, things are [38:48] going maybe slower than the people who [38:51] want to invest in this, um but actually [38:53] get [38:53] >> back into your area of expertise. [38:56] Um [38:56] >> Yeah. [38:57] >> You know there's big concentration of [38:58] data centers in Virginia for [39:01] reasons that are really interesting. Uh [39:02] it's where AOL started. Yeah. So, I [39:04] think the first data center was there to [39:06] serve them. [39:07] >> Yeah. [39:07] >> And then as a result, it's like the data [39:09] center capital of the world. Yeah. Um [39:11] but [39:12] you are starting to hear stories about [39:14] communities that either for [39:15] environmental reasons or [39:17] for the just the reasons of we don't [39:20] like AI, [39:20] >> Yeah. [39:21] >> but you're starting to see sort of [39:22] grassroots. It's not very loud. I don't [39:25] know if it's stopped anything of [39:26] consequence, but [39:27] >> But they [39:28] you know, the concept of AI definitely [39:30] has like a PR problem. [39:32] >> Big time. [39:32] >> And and yeah, you're seeing it manifest [39:35] in [39:36] kind of a non-partisan way, right? Since [39:38] Virginia, California, Oklahoma, like you [39:41] know, it it doesn't [39:42] >> red or blue. [39:43] >> Yeah. It's different people objecting to [39:45] it for different reasons. [39:46] >> Right. Um [39:47] it's not Listen, it's a risk we have to [39:49] watch. It's not yet big enough relative [39:52] to the build out and where build outs [39:53] can happen around data centers and [39:55] other, you know, [39:57] like like other parts of the [39:58] infrastructure for it to like matter all [39:59] that much, but we have to watch it. Um [40:02] but the other thing I'd say cuz you're [40:03] like, is it early is it not? [40:05] You know, our economics team has about a [40:08] quarter of GDP growth this year coming [40:10] from just like the build out of the [40:12] current build out of AI. Like really [40:14] very little baked in from a productivity [40:16] perspective. So, that that's the other [40:18] aspect of it. [40:19] >> the construction aspect of it is driving [40:21] that world. [40:22] >> the like the the Yeah, just like the [40:23] That's literally let laying laying down [40:26] pipes and building structures. [40:27] >> Mhm. Okay. Do you think a productivity [40:29] boom is coming? [40:30] >> Yeah, we think so. I I think it's hard [40:32] to [40:32] time it exactly, right? I mean, listen, [40:34] as speaking to someone who's [40:37] actively trying to put AI into all of [40:39] our processes in the department and you [40:41] know, my boss Katie Huberty, he's been [40:43] like very forward-looking on this and [40:45] even before I stepped into this position [40:47] sort of driving [40:49] um [40:49] this in all sorts of like innovative [40:51] takes on how on the future of research [40:53] that require AI being linked into the [40:56] process. [40:57] Um yeah, industrializing this stuff in [41:00] large organizations and unlocking that [41:02] productivity is you know, sometimes [41:04] easier said than done. The individual [41:05] productivity that somebody might ex- be [41:07] experiencing at home or if they're [41:10] starting a business from scratch using [41:12] some of these tools, [41:13] uh that could be mind-boggling you, but [41:15] inside of a large organization where [41:17] you're linking workflows and people [41:18] together, it could take time. [41:20] My only point being that like [41:23] it Do you see major productivity gains, [41:25] you know, this year, next year, the year [41:26] It wouldn't surprise me if it still [41:28] takes a little, but like the direction [41:29] of travel is pretty clear. [41:31] >> Don't you have to see it in tech [41:32] companies themselves first? [41:34] They're already very productive [41:35] companies, but I just [41:38] it struck me that like we're selling all [41:40] this AI and we expect the the recipients [41:43] of the technology to be able to run [41:45] their businesses more efficiently. [41:46] >> Yeah. [41:47] >> And of course, we want to hear about uh [41:49] restaurants with rising profit margins [41:51] and we want to hear about uh hotel [41:54] companies and and insurance companies. [41:56] Of course. [41:56] >> Yeah. [41:57] >> But like wouldn't the first place we see [41:59] it be at the technology companies [42:02] themselves who are actually developing [42:05] >> the tech? [42:05] >> And I mean, and and listen, you just [42:07] going by what they report about [42:08] themselves, you're seeing that. And I'd [42:11] say that our team does a lot of primary [42:14] research on this. So, we you know, we we [42:17] cover about 3,400 stocks globally. [42:20] Steve Bird and Katie [42:22] will will do this thing periodically [42:24] where every analyst has to sort of map [42:27] their coverage into whether you're an AI [42:30] enabler, an AI adopter, beneficiary and [42:32] to what degree based on [42:34] your assessment of the company. And then [42:36] we'll do surveys of the management of [42:40] those companies and we'll also scrape [42:43] the transcripts, the earning transcripts [42:44] of those companies to see like who's [42:46] actually reporting something tangible [42:48] about AI. [42:49] >> of the S&P seeing tangible benefits in [42:52] their financial statements from the use [42:54] of AI? [42:55] >> Yeah, so it's so the [42:56] MSCI world [43:00] the numbers exactly right, but like it's [43:02] I think that the beginning of last year [43:04] was like 20%. We're up into like the 30s [43:06] now. So, it's it's climbing. And then [43:09] for for companies that we mapped as AI [43:12] adopters [43:14] from 2024 to 2025 the average EBITDA [43:17] expansion was was 2x faster than the [43:20] non-adopters. So, [43:22] the benefits are there. They're starting [43:24] to rise. [43:25] Um [43:26] but [43:28] you know, to the point of it like [43:30] specifically translating into like GDP, [43:33] like that like that's a bit of a [43:35] different measurement problem. But in [43:37] the in any case, you're starting to see [43:39] the tangible benefits show up in [43:40] earnings for sure. [43:42] >> So, right now it is [43:44] um [43:45] a lot of charts in here. John, let's [43:46] pull up chart [43:49] My God, I sound like my friend Sessa. [43:52] What chart are we looking at? [43:53] Chart 20. [43:54] >> Wish you wish. [43:55] >> Chart 20.2. [43:57] Um okay, we're looking at the consensus [44:00] 2027 EPS change for AI infrastructure [44:03] stocks. [44:05] Insane, 42%. The S&P 500 at 10% and the [44:10] S&P 500 XAI infrastructure [44:13] just 3%. [44:15] Man, this better go up. [44:17] For this to work. [44:18] >> Yeah, I mean [44:18] >> And I think I mean I think it will. [44:20] >> No, I mean we think it will too, but [44:21] you're right. There's been this at least [44:22] in terms of like the broadening theme. [44:24] Okay, and this is not my area of [44:25] expertise, but you know, [44:28] so I'm going to I'm going to do my best [44:29] Mike Wilson for a second here. Um [44:32] There's been a divergence between [44:35] the you know, actual earnings broadening [44:38] and like the price broadening out and [44:40] and we expect those things to reconcile, [44:42] but you're right. Like it hasn't hasn't [44:44] shown up in enough places quite yet. [44:47] >> Meaning stocks are front running that. [44:49] Like are you saying this is [44:49] >> No, that would [44:50] >> Or backwards? [44:52] >> The other way. [44:53] >> Okay. [44:53] >> Yeah. Yeah. [44:54] >> Do you think uh [44:56] Do you think there's at at any point in [44:59] time if we're going to see a broadening [45:01] trade? I'm going to put up the next [45:03] chart John after that. So it's a good [45:05] year for the S&P on a on a headline [45:07] basis. [45:08] Um But percentage of stocks positive uh [45:12] is below average not substantially so. [45:14] It's average. [45:15] >> It's it's about it's about average. [45:16] >> Yeah. [45:17] >> We keep hearing that the next leg of [45:18] this bull market will happen away from [45:20] the AI [45:21] uh hyperscalers [45:23] >> Yeah. [45:23] >> and will broaden out to the rest of the [45:25] market and like obviously every sector [45:28] has its own reasons for rallying or not [45:30] rallying. Energy is obvious, housing [45:32] related is obvious like what the drivers [45:34] are there. Um but do you need do you [45:37] guys as a house view? [45:38] >> Yeah. [45:39] >> Do you guys think we actually even need [45:41] this broadening out and we need to see [45:43] all these AI users start beating [45:46] earnings or can we just continue on the [45:49] way we've we've gone [45:51] with this sort of concentrated subset of [45:53] AI champions driving all the growth [45:56] >> Yeah. [45:56] >> and the market hold up? [45:58] >> It's an interesting way to phrase the [45:58] question. I I I don't know if we'd say [46:01] you need it, but that that's the way we [46:03] think it's going to happen. [46:04] >> You think it'll play out that way? [46:05] >> Yeah. [46:05] >> That'd be very positive for a lot of [46:07] investors. [46:08] >> Yeah. [46:08] >> Okay. There's people that are not [46:10] exposed enough to AI or looking at the [46:12] tape and they're saying, "Why is [46:14] everyone having so much fun?" [46:15] >> Yeah. Yeah, I mean so we you know, we do [46:19] a sort of big collaborative mid-year and [46:21] year-ahead outlook process every single [46:23] year where you bring together all the [46:25] strategist, the economist, and um uh for [46:29] the last two cycles pretty consistent on [46:33] broadening out of equities in the US [46:36] plus some of the secular trends we [46:37] talked about overseas and equities [46:40] kind of leading things. And it's not [46:42] that bonds are a bad place to be, but [46:45] they're going to serve more as a [46:46] diversifier, coupon-like returns, etc. [46:49] Um cuz you're in a [46:52] you know, secular enough growth [46:53] environment driven by a lot of these [46:57] kind of [46:58] major CapEx trends [47:00] that [47:01] you know, that's going to be the norm [47:02] for a couple years. [47:03] >> You know what you know what's [47:04] interesting? We're So we're seeing this [47:05] right now. The Mag 7 [47:08] like Meta and Microsoft, I know it's [47:10] just two, and [47:11] they they're acting like And [47:13] Nvidia got smacked down after trying to [47:14] break out. It's not acting awesome. The [47:16] Mag 7 are up 70 basis points year to [47:18] date. [47:19] The equal weight S&P is up 11%. [47:22] >> Yeah. [47:22] >> The Nasdaq 100, this is wild. You have [47:24] the Mag 7 are flat, which is 40% of the [47:28] Qs. [47:29] And the Nasdaq 100 is up 19% year to [47:32] date. [47:33] I don't know what the equal weight [47:33] version is. It's obviously got to be [47:35] more than that. So the Nasdaq equal [47:36] weight is up Huh. [47:38] Delete this, John. It's not up more than [47:40] that. [laughter] Um but whatever. The [47:42] The Nasdaq 100 is up 19% year to date. [47:45] >> Without its large without its largest [47:46] stocks doing all that much as a group. [47:48] So all right, so maybe that broadening [47:50] thing is already [47:50] >> Okay, here it is. Here it is. Here it [47:51] is. This is what I was looking for. [47:53] Equal weight tech. [47:56] Again, the Mag 7 are a flat [47:59] equal weight tech, RSPT is the ticker, [48:02] is up 39% year-to-date. [48:04] >> Mhm. [48:05] >> Yeah. [48:06] >> So, we got the burning We got the [48:07] burning in tech. [48:08] >> Yeah. [48:09] >> It's good. [48:10] >> That's a crazy spread. [48:11] >> to start somewhere. [48:12] >> Um I want to ask you about the midterms. [48:14] Are you getting a lot of questions from [48:16] uh investors, hedge funds, whatever? You [48:19] said earlier in the show you think it's [48:22] mostly going to be noise that comes out [48:23] of that and there will be policy [48:25] continuity. Is there a way where that [48:26] goes wrong or um what are people what [48:29] are people asking about this cycle? [48:31] >> I suppose the way that view is wrong [48:34] is if either Democrats or Republicans [48:37] win [48:39] so big [48:40] that they [48:41] >> Which is not the forecast. [48:42] >> No. Well, and you asked about prediction [48:45] markets earlier. I think coming in here [48:47] uh prediction markets had Dems at like [48:50] an 80% chance to take the House and to [48:53] take both the House and the Senate it's [48:54] kind of a kind of a toss-up about 50/50. [48:57] But if you map out [48:58] >> Senate comes down to Maine basically or [49:00] is [49:00] >> Right, exactly. And it's like And the [49:02] thing is right, so the point is if if it [49:04] comes down to one of those states, [49:06] you're talking about the Dems getting to [49:08] like 50, 51 seats. But like even if they [49:10] got the 52, 53 seats [49:12] um you don't have filibuster-proof [49:15] majority to push anything meaningful [49:16] through. You definitely don't have the [49:18] majority to [49:20] um overcome any veto. So, if you know, [49:24] like if clients question or like oh, [49:26] well, is the the tax cuts I got passed a [49:28] couple years ago were those at risk? [49:30] No, right? I mean, because the you know, [49:33] I believe the Democrats are sincere in [49:35] wanting to reform that but they don't [49:37] they're not going to have the votes and [49:38] they certainly aren't going to have the [49:39] White House for another um couple of [49:41] years. [49:42] >> Okay. [49:42] >> Um so, it's more And like can they wrest [49:45] back the power on like trade and [49:47] tariffs? Again, practically speaking [49:49] not. Like it is their you know, Congress [49:51] over years delegated that authority to [49:53] the president, they can take it right [49:55] back, but you need the votes in order to [49:56] do that. [49:56] >> So, there's no political shock that we [49:58] don't think there's a political shock [49:59] coming at some point this fall [50:02] based on anything that we're seeing [50:03] currently. [50:04] >> I don't think so. They [50:05] >> And you're not And you're not giving [50:06] people the impression that there's going [50:07] to be some sort of fireworks. [50:08] >> Yeah, I mean, I here's the caveat. [50:11] Could there be [50:13] knee-jerk reaction [50:14] in the markets [50:16] if you get a result that is surprising [50:20] on election night in favor of a change [50:23] of power, right? So, that you know, the [50:25] Democrats effectively take control both [50:28] chambers, do better than expected, and [50:31] the next day, the next week, you know, [50:33] anything that's associated with like a [50:35] positive, right? So, you know, the the [50:38] tax benefits from the O BBBA, you know, [50:42] the the the sort of sectoral [50:43] distribution of those, a lot of that's [50:45] been sort of more small cap friendly. [50:46] Like, could you see some temporary [50:48] weakness there on like a knee-jerk that [50:50] oh, this presages that 3 years down the [50:51] line there's going to be a major tax [50:53] change. [50:54] >> No one's going to trade on that. [50:55] >> yeah, right. And if they did, I would [50:57] say [50:58] fade that, right? That's that's not the [50:59] right way to think about it cuz in [51:01] two two years till the next election [51:03] cycle, might as well be 100 years. [51:05] >> What's the most surprising question that [51:06] you've gotten from an institutional [51:08] investor or a hedge fund manager this [51:09] year? [51:11] What made you think the hardest or what [51:13] made you really have to go to the [51:15] drawing board? [51:17] >> Um [51:18] well, I'll tell you what, like, we got a [51:21] lot what when the Iran stuff started, [51:23] there wasn't questions about whether or [51:26] not straight-of-hormones would become a [51:28] choke point, but there were a lot of [51:30] like very technical questions about [51:33] like, [51:34] you know, [snorts] [51:34] how much oil's going through, who can [51:36] get it through there. Um you know, [51:40] what are sort of the like puts and takes [51:43] around like the Gulf Coast countries and [51:44] their involvement. And that's another [51:48] thing where we just had to like pull in [51:49] a team effort and [51:51] >> I'm not a big Hormuz expert going into [51:53] that. [51:54] >> Going into it, no. I will say like I [51:56] I literally I I knew it was on the map. [51:59] I knew that if [52:00] >> It's on the right side. [52:02] >> All right. I I I knew that like well, [52:04] it's just the contingency planning [52:06] playbook if Iran gets attacked. That's [52:08] sort of the if if they view it as [52:10] existential. [52:10] >> sound so obvious. I don't think our [52:12] military knew that. [52:13] >> Well, it's I mean [52:14] >> It seems that's what happened. They had [52:15] a plan for that, so [52:16] >> It would be [52:18] I think our military probably knew that. [52:19] Like it's it's it's a question of [52:21] whether or not you thought it would [52:23] happen. And then well, and I'll say this [52:24] in in fairness to that debate. Um [52:28] It it it you have to go back I think to [52:30] I'm I'm not going to get this exactly [52:32] right, but there was there was something [52:33] called the Tanker Wars back in the '80s. [52:34] It was like the last time Iran ever [52:37] really shot at some ships. And it was it [52:39] was widely thought it that it would be [52:41] um not in their self-interest to do that [52:44] unless they felt like an existential [52:46] crisis because otherwise it would sort [52:48] of turn the world against them. Well, [52:50] clearly this feels pretty existential to [52:53] them. In any case [52:54] >> that threshold where they said, "Yeah, [52:56] we're doing this." [52:57] Right. [52:57] >> So right I mean I I think it's [52:59] self-evident. So [53:01] Yeah, so the questions we start getting [53:03] are like really all about the technical [53:04] aspects. How many ships can go through? [53:06] Under what conditions? Etcetera. Yeah, [53:08] credit to um Martin Rats, who's our [53:10] commodity analyst out of London, who I [53:13] just know he puts in the case of the oil [53:15] market as well as anyone that I know. Um [53:18] and to his ability to we lean into using [53:21] some of these AI tools that we now have [53:24] for people is able to like wire things [53:26] up pretty quickly using open-source [53:27] information to like map out right? Like [53:30] exactly what could flow through at what [53:32] rate under which conditions. And so that [53:35] was a hard question I got. I didn't [53:37] answer it. But but I knew that you know, [53:39] this is I think a testament to to to the [53:41] firm like we can get you the right [53:43] person. [53:43] >> Michael, what what impact do you think [53:45] all this AI and the ability for [53:47] everybody to know everything really fast [53:49] >> Yeah. [53:49] >> is going to have on markets? That's such [53:51] a fascinating thing to think about. [53:53] >> Um [53:55] it's Yeah, that's a great question [53:57] because I think a lot of [54:01] what we know or or or a lot of what we [54:04] think about as being very smart people [54:05] in markets, like that level of [54:07] intelligence I think has become table [54:09] stakes pretty soon, right? Um the [54:12] ability to [54:14] layer Monte Carlo analysis on a [54:16] financial model. Again, it's not like [54:18] any we're the third first people who [54:20] have ever thought of Monte Carlo [54:21] analysis, but like now you can just do [54:23] it really easily, right? And without [54:24] having to like hire a statistician and [54:27] have a supercomputer operating next to [54:28] you. Um or do like large-scale document [54:32] analysis, right? Like um you know, you [54:36] know, let's let's score bond indenture [54:37] documents. Let's do 40,000 of them at a [54:40] time. Let's map them on the spreads and [54:41] let's see the opportunities. Like stuff [54:43] like that which you know I I call like [54:45] like level three research. [54:47] It was stuff that you just couldn't have [54:48] done before. [54:49] >> You needed an army of bodies and now you [54:52] don't. [54:52] >> Right. Now it's a it's a powerful chip. [54:55] >> Right. [54:55] >> And somebody that knows how to get the [54:58] best out of the machine with the right [55:00] prompt. [55:00] >> Yeah. [55:01] >> Which is a skill, I suppose. [55:03] >> It's a It's a In in [55:04] >> remove a big competitive advantage from [55:06] gigantic research organizations. [55:09] >> Well, I guess but [55:09] >> So you you need the human judgment there [55:11] also at the end of the day. [55:12] >> Right. So you you you Yeah, you you [55:14] remove the need for to have a giant [55:16] research. However, you also increase the [55:19] possibility and the probability that [55:21] people like just just like there's AI [55:23] slop in, you know, like video and the [55:25] stuff you see on like Instagram, [55:28] there could be tons of like research [55:31] slop, right? [55:31] >> It'll get overfit. Like the machine [55:33] wants to please us. [55:34] >> Right. [55:35] >> It wants to give us what we seem to want [55:37] because it's [55:39] every business on earth is now an [55:41] engagement business. [55:42] >> Yeah. [55:42] >> Every business is Netflix and TikTok. [55:44] >> Yeah. [55:45] >> So, the machine doesn't want you to turn [55:47] it off. [55:47] >> Yeah. [55:47] >> So, it will to your point [55:50] keep asking the question, it'll [55:51] eventually give you something that [55:52] sounds right. [55:54] >> Yeah, and even like yeah, like like [55:56] before I said we know, you know, it's [55:57] not like you know one's ever heard of [55:58] Monte Carlo analysis before. [56:01] But just hearing [clears throat] about [56:01] it and knowing how to do it the right [56:03] way are two different things, right? So, [56:05] you you said I don't know [56:06] what you used was wisdom or experience, [56:08] but like I actually think that's an [56:10] important difference. [56:11] >> Judgment. [56:11] >> Judgment, right? Like what's the where [56:13] that saying what's the difference [56:14] between wisdom and knowledge? [56:15] >> Like a tomato is a fruit. [56:17] >> You got but [56:18] >> You don't put it in a fruit salad. [56:19] >> wisdom is you don't put tomato in a [56:21] fruit salad. [56:22] >> Right, so like that's still really, [56:25] really important. [56:26] And I think you need a like network of [56:29] experts to really sort through that [56:32] noise or you could start making some big [56:33] mistakes in judgment. So, [56:36] what happens in the future when [56:37] everyone's enabled like this? It might [56:39] not look all that different than now [56:40] where there's, you know, people chasing [56:43] narratives and that aren't right and [56:45] ultimately they get proven wrong and [56:47] people are wiser and and, you know, [56:49] operate better more sustainably over [56:51] time. [56:52] >> Before we let you out of here, can we do [56:53] a like a little Morgan Stanley research [56:56] lightning round? [56:57] >> Sure. [56:57] >> Okay. What's the house view on [57:00] uh [57:01] employment in the AI era? Are you guys [57:05] optimistic that the new jobs will come [57:08] along on time so that we don't have like [57:11] a big air pocket or do you guys think [57:13] there are going to be some tough years [57:15] for young workers before these new [57:18] opportunities present themselves or [57:20] >> Yeah. [57:20] >> maybe some other option that I'm not [57:21] thinking of. [57:22] >> I mean, I would say optimistic until [57:24] proven otherwise kind of like I eyes [57:26] wide open. So, our economist has done a [57:29] lot of work breaking down [57:31] um the sort of mapping the types of work [57:34] in the economic data that map over to [57:38] the skills that are considered [57:40] disruptable. I'm trying to see where is [57:42] the employment unemployment cohorts [57:43] higher and lower than others and like [57:45] >> You squint legal and accounting. [57:47] >> Yeah, and like you you squint, you see [57:49] some difference in like youth [57:51] unemployment, but like you got to squint [57:53] at the moment. Um [57:56] and so I'd say [57:58] I'd say mostly optimistic but vigilant. [58:01] >> Not that you're going to make stock [58:02] picks, but enterprise SaaS, the SaaS [58:04] apocalypse [58:05] >> Yeah. [58:05] >> What's the Morgan Stanley outlook for I [58:07] mean, this is [58:09] I think the standout negative group of [58:11] the year. [58:12] >> Yeah. [58:12] >> Uh or and these are some very large [58:14] companies. [58:16] Uh companies that you guys are probably [58:18] covering all of them. [58:19] >> Yeah. [58:19] >> Uh what what is Morgan Stanley saying to [58:21] people that ask about are values being [58:24] created there or is it going to be [58:26] harder than maybe the the maybe harder [58:30] >> Yeah. [58:30] >> to to make money there going forward. [58:32] >> Well, I mean, I think [58:34] both. It's it's going to be harder there [58:36] fundamentally and there's value being [58:39] created. And um we think a lot of what [58:42] happened in the markets was overblown. [58:45] The ability for these companies to [58:47] industrialize in a way that individuals [58:50] in a large organization [58:52] uh you know, might not necessarily be [58:54] able to do things for themselves. It [58:56] means that there's there's still a moat [58:57] there. [58:58] >> You think that's a enough of a a moat [59:00] that would justify investment in these [59:03] companies or [59:04] Yeah, I mean [59:05] >> It's probably too early to tell. [59:05] >> I'm not an expert enough to say on those [59:07] specific companies, but I just like [59:08] conceptually [59:10] um [59:11] I would say like we're not a buyer of [59:13] the idea that everyone's just going to [59:15] write software for themselves. [59:16] >> Stanley research budget include as much [59:18] money going to enterprise SaaS 3 years [59:21] from now as it does today [59:24] or will uh compute and anthropic bills [59:28] eat a bigger chunk of that expenditure. [59:31] Like you obviously don't know for sure, [59:33] but if you had to guess directionally [59:35] where it seems to be headed. [59:36] >> it's a good question cuz it's hard to [59:37] like distinguish between just like [59:40] buying the tokens and then some of the [59:42] other service providers that plug into [59:44] them. I mean listen, I think our [59:47] our budget going towards tech solutions [59:48] probably going to be pointed higher, at [59:51] least as a percentage of the budget. But [59:52] like a lot of the solutions that um you [59:55] know that that Katie and her team are [59:57] putting together [59:58] really boil down to making sure it's not [60:01] just about like token usage, it's also [60:03] about having the right data plugged in [60:05] and the right sort of, you know, um um [60:07] like use cases that that kind of come [60:09] from that. So [60:10] uh I'm not trying to avoid your [60:12] question. I just think it's it's it's [60:13] not as simple and like we don't [60:16] we don't know [60:16] >> lightning round ever. [60:17] >> Yeah, sorry. Well, too it's hard. Yeah, [60:20] how about yes? I'll [60:21] >> [laughter] [60:22] >> Here's an easier Here's an easier [60:23] question. [60:24] Uh how do you Morgan Stanley clients [60:27] feel about uh Kevin Warsh and the new [60:29] Fed? [60:30] >> Yeah, um [60:32] I would say fine. [60:33] >> Bond market seems okay. [60:34] >> Yeah, right. I think [60:35] >> So far he hasn't had to do much, but [60:38] >> So [60:39] yeah, I think I think the answer is [60:40] fine, which is a boring answer, but um [60:44] there would be I think far more anxiety [60:47] if you'd gone with one of these less [60:49] traditional names that have been thrown [60:51] out there. People are really embracing [60:53] unorthodox economics [60:55] >> asset or some somebody coming from pure [60:57] politics. [60:58] >> or someone who was perceived to be [61:01] just somebody who's going to take a [61:02] phone call from the White House. And [61:04] like the [61:05] And and I would say this, even in that [61:06] situation we probably would have leaned [61:08] against the concept that um the Fed [61:11] substantially compromised. As our Seth [61:14] and our econ team point out all the [61:15] time, like there's there's a whole [61:16] committee there. It's not just one [61:17] person. And so sure, communication could [61:20] get scrambled. There's all It could get [61:22] confusing. [61:24] But [61:25] it wouldn't necessarily if you just [61:27] looked back on it in history on like a [61:28] chart with like what what the policy [61:30] rate was relative to inflation and [61:32] growth, it might not look all that [61:33] crazy. Now, we don't have to deal with [61:35] that hypothetical debate really. I mean, [61:37] we we got to you know, we we we still [61:39] have to pay attention. [61:40] >> He's traditional enough. [61:42] >> Yeah, now now let's let's see if we're [61:44] right. I mean, we're we're going to [61:45] we're about to learn. But yeah, so the [61:46] the short answer is fine. [61:48] >> Okay. So, on balance uh Morgan Stanley [61:51] is like not the most bullish on the [61:53] street, but you guys are pretty [61:55] constructive on most asset classes. And [61:58] uh sort of like if we if the earnings [62:00] growth delivers, we don't get a worse [62:02] geopolitical situation in the Middle [62:04] East. And the interest rate picture is [62:08] probably just more of the same. Like if [62:10] those things hold constant, we should [62:12] finish out the year in pretty good [62:14] shape. [62:14] >> Yeah, I think that's right. And it's [62:15] largely buoyed by the secular trends and [62:18] themes that are driving this around [62:20] CapEx and [62:22] you know, AI and dealing with this [62:24] transition to a multipolar world. [62:26] >> I want to tell you uh we really learned [62:27] a lot from you today and thank you so [62:29] much for coming to the show. Did you [62:30] have fun today? [62:31] >> I I did and I really appreciate you guys [62:33] giving me the opportunity. [62:33] >> Want to come back sometime? [62:35] >> Of course. [62:35] >> Are you busy tomorrow? [62:37] >> Uh no. Yeah, what time? Like 7:30 a.m. [62:40] sharp? [62:41] >> Uh dude, thank you so much for being [62:43] part of the show. Um where can people [62:45] find more from you and from your team? [62:47] If somebody says, "Michael Zezas, this [62:49] is great. I want to I want to hear more [62:51] of what he has to say. What's the right [62:52] place to go?" [62:53] >> Morgan Stanley Institute. Go to the [62:55] webpage, sign up for the Morgan Stanley [62:56] Institute newsletter. We're bringing [62:58] together all [music] the thought [62:59] leadership across the firm for all types [63:01] of financial decisions. [63:02] >> The Morgan Stanley Institute. This has [63:03] been Michael Zezas. Thank you guys so [63:05] much for listening. Thanks for watching. [63:08] Have a great weekend. We'll talk to you [63:09] soon. [63:10] >> [music] [63:15] [applause] [63:28] >> Woo!