AI Capex Tops Housing Peak as Softening Pay Squeezes US Consumer

AI资本支出超地产峰值，薪资疲软挤压美国消费者 企业在AI基础设施方面的巨额支出使美国经济保持增长势头，尽管实际工资增长滞后和企业利润份额上升正对消费需求构成威胁。 本期节目分析了美国经济当前的运行轨迹，重点指出了规模已超过GDP 7%的空前AI基础设施投资潮如何推动商业周期并扭曲传统的市场广度。讨论小组探讨了企业利润与员工薪资之间日益扩大的分化，指出尽管高端可选消费依然强劲，但由于实际收入增长乏力，潜在的信用压力正在显现。最后，主持人深入剖析了劳动力市场的韧性，反驳了关于AI摧毁就业的论调，并列举了经济放缓的潜在催化因素，例如迫在眉睫的财政拖累以及企业AI资本支出疲软。 AI 资本支出规模已突破美国 GDP 的 7%，不仅超越了 2005 年房地产泡沫时期的峰值，也高于 1990 年代末互联网繁荣水平，成为主导现代商业周期的核心引擎。 (16:37) 传统 GDP 核算因海外芯片等硬件进口直接被数学扣除而低估了 AI 的经济拉动，但巨额科技支出正通过高管与员工 RSU 套现、地方税收及高端可选消费产生广泛的经济外溢。 (19:18) 经济产出分配出现历史性分化，一季度劳动报酬在国民总收入（GDI）中的占比降至 1947 年有记录以来的最低水平 51%，而企业利润占比则飙升至 1950 年以来的最高点 12.1%。 (45:19) 美联储政策面临名义 GDP 增长约 6% 与薪资增速不足 4% 的脱节难题；若供给冲击引发的持续通胀迫使央行追加 75 至 100 个基点的紧缩加息，恐将对实体经济造成沉重打击。 (55:05)

Macroeconomists Neil Dutta and Skanda Amarnath argue that conventional recession signals are failing because a massive wave of artificial intelligence capital investment has supplanted housing as the primary driver of the business cycle, masking underlying labor softness and household credit stress. AI CapEx Displaces Traditional Leading Indicators (16:37) Skanda Amarnath points out that capital spending on AI and tech infrastructure has surged past 7% of U.S. GDP, surpassing both the 2005 housing peak and the late-1990s telecom boom. Historically, housing served as the core business cycle indicator, but residential construction is currently muted—with new home sales down 6% to 7% year-to-date—as it competes with data centers for physical resources, land, electrical hardware, and power grid access. This investment boom has transformed broader industrial order books. Heavy machinery and energy firms such as Caterpillar, GE Vernova, and Vertiv increasingly trade and behave like semiconductor suppliers due to dedicated data center demand for backup generation, turbines, and cooling systems. While critics highlight that overseas hardware purchases register as import leakages that mechanically subtract from headline GDP, Amarnath and Dutta emphasize that business cycles are driven by corporate risk appetite and capital commitments rather than GDP accounting identities. Labor Share Plunges as Corporate Profits Surge (45:19) Beneath robust headline expansion, the panelists highlight a historic decoupling between corporate earnings and worker compensation. In recent Gross Domestic Income (GDI) figures, labor's share fell to 51.0%—its lowest level since record-keeping began in 1947—while corporate profit share climbed to 12.1%, its highest point since 1950. First-quarter corporate profits expanded by 2.7%, while worker compensation grew just 0.8%. Neil Dutta pushes back against assertions of an immediate productivity boom, emphasizing that the current cycle features rising software prices and negative real compensation growth rather than the broad price deflation seen in the 1990s. Gains from current technological investment are flowing overwhelmingly into corporate profit margins instead of worker paychecks, leaving nominal wage growth trailing below 4.0% while nominal GDP expands around 6.0%. Household Polarization and Credit Balance Strains (24:38) The discussion reveals an expanding divergence between affluent asset owners and wage-dependent households. Supported by equity market gains, high-end discretionary spending remains elevated—exemplified by secondary market tickets for New York Knicks playoff games exceeding $4,500 to $7,000 and strong demand across air travel, resorts, and experiential services. Conversely, real consumption of physical goods, apparel volumes, and staple foods has softened. Dutta notes that compressed personal saving rates and decelerating aggregate wage growth coincide with rising delinquencies across credit cards, auto loans, and mortgages flagged in Federal Reserve reports. Amarnath warns that as petrochemical and input cost pressures are passed through retail consumer staple channels, real purchasing power faces further compression unless underlying labor market momentum rebounds. Monetary Policy Dilemmas and Cycle-Ending Catalysts (48:26) This macroeconomic divergence presents a dangerous setup for the Federal Reserve. Dutta and Amarnath warn that if policymakers interpret supply-driven price shocks in commodities and transportation as broad demand-side overheating, the central bank could trigger an aggressive 75 to 100 basis point rate-hiking campaign. Such blunt monetary tightening would fail to fix structural supply bottlenecks like shipping route disruptions, while actively punishing a vulnerable domestic labor market. Regarding the ultimate conclusion of this economic expansion, the panelists point to two primary structural risks: fading fiscal stimulus turning into a net economic drag by early 2026, and enterprise 'CapEx fatigue.' Amarnath draws parallels to the 2000 tech cycle, warning that if non-tech corporations treating AI infrastructure as an operating expense fail to realize clear returns on investment, corporate software and hardware budgets could face abrupt curtailment. While aggressive AI infrastructure spending and corporate liquidity have successfully averted near-term recession risks, the broader economy faces widening structural cracks. The unresolved question is whether non-tech enterprises can translate costly AI adoption into genuine worker productivity and wage gains before passive monetary tightening, fiscal drag, and household credit strains derail the expansion.

宏观研究者 Neil Dutta 与 Skanda Amarnath 认为，住房营建许可和个人储蓄率等传统衰退指标当前均已失灵，因为以数据中心和半导体为核心的人工智能（AI）资本开支已实质性取代房地产，成为驱动美国商业周期的新核心。然而，企业利润与劳动者报酬份额的历史性分化、下层消费者的信贷承压，以及非科技企业对 AI 投资回报率（ROI）预期的潜在疲劳，构成了当前经济平稳表象下的关键脆弱性。 AI资本支出取代地产成为周期核心引擎 (16:37) 传统宏观框架通常采纳经济学家 Ed Leamer“住房即商业周期”的论断，但当前这一传导机制已被严重打乱。全美新屋销售年内迄今同比下滑 6% 至 7%，住宅建筑不仅承受着高利率与投入成本的挤压，还在土地、电力及供应链资源上与 AI 数据中心建设产生直接竞争。Amarnath 指出，涵盖算力硬件、软件及电力能源设备的科技资本开支在 GDP 中的占比已突破 7%，不仅超越了 2005 年房地产泡沫时期的峰值，也高于 1990 年代末互联网与电信基础设施投资的高点，成为主导美国产出的最大周期力量。 针对部分华尔街经济学家因高带宽内存（如 SK Hynix）等零部件依赖海外制造而将巨额支出作为“进口漏损”从 GDP 扣除的算法，Dutta 与 Amarnath 提出反驳，强调商业周期本质上是“风险周期”而非国民核算账目。科技巨头的利润与限制性股票（RSU）兑现正实质性支撑地方财政与高净值人群支出（例如加州所得税代扣增长的一半以上由科技 RSU 驱动），并在高端服务与文娱消费领域产生强大的外溢乘数。只要科技巨头与超大规模计算中心持续通过债务与现金流扩张基建，以投资为导向的宏观基本面便难以迅速失速。 利润与报酬分化加剧劳动份额历史低点 (45:19) 美国经济内部正在经历罕见的分配割裂：名义 GDP 增速维持在约 6%，但名义薪资与薪酬增长已滑落至 4% 以下。最新一季度 GDP 修正数据显示，全美劳动者报酬仅微增 0.8%，而国内企业利润大幅增长 2.7%。这导致劳工在国民总收入（GDI）中的分配份额骤降至 51%，创下 1947 年有统计记录以来的最低水平；与此同时，企业利润份额升至 12.1%，触及 1950 年以来的历史峰值。Dutta 强调，不同于 1990 年代 IT 硬件通缩伴随实际家庭收入飙升的生产率繁荣，本轮 AI 带来的收益几乎全额被企业利润率吸收，普通工人反而承受着负实际薪酬增长。 这种劳动力市场的结构性分化解释了民众对整体宏观数据的不满。除数据中心营建与部分高端制造业外，传统蓝领需求偏弱；而白领岗位在经历疫情后招聘调整与技术整合期后虽有企稳，但招聘动能依然滞缓。针对 JOLTS 职位空缺跳升至 760 万的读数，Dutta 驳斥了“AI 迅速摧毁所有岗位”的“劳动总额谬误”，但指出空缺数据充斥虚假招聘，应以更为真实的入职率与离职率作为判断依据。Amarnath 则观察到，企业出于维护部署未来 AI 工作流所必需的核心技术知识与组织记忆，表现出明显的裁员犹豫，避免盲目跟风激进减员。 财富效应掩盖消费分化与信贷压力 (24:38) 美国消费市场正分化为两条平行的运行轨道。股票市场的持续繁荣在高净值群体中催生出强大的财富效应，推动高端赛事、豪华度假与航空出行等体验类服务需求居高不下，Delta 和 Hyatt 等企业股价维持强势；与此同时，GLP-1 类减肥药物的普及也在改变消费者偏好，促使资金从快餐食品流向休闲娱乐。然而，处于收入分配底层的消费群体正面临日益严峻的流动性枯竭，美联储《褐皮书》已明确记录了信用卡、汽车贷款及农业贷款违约率的攀升。 消费品行业正面临成本转嫁带来的挤压效应。随着石化原料等成本抬升，消费必需品企业正计划通过提价将压力传导给家庭终端，进一步削弱实际购买力。Dutta 研判，全美实际消费支出目前维持在约 2% 的年度温和扩张，呈现出渐进的“线性放缓”特征。在缺乏真实工资增长支撑的环境下，部分商品虽因涨价维持了名义销售额，其实际售出量已在收缩。除非发生跨行业的广泛裁员，或股市出现深幅回撤直接破坏高净值人群流动性，否则消费更倾向于向 1.0%–1.5% 的低速区间滑移，而非直接断崖式陷入衰退。 供给冲击下的联储困境与AI资本疲劳 (48:26) 美联储货币政策正面临严峻的权衡考验。Amarnath 提出了鹰派情景：若强劲的科技投资最终带动劳动力需求和工资向 6% 的 GDP 增速靠拢，央行或将被迫通过 75 至 100 个基点的连续加息以逆转先前的预防性降息。Dutta 则给出了鸽派图景，认为薪资增速回落将拖累消费降温，央行应保持耐心。更为棘手的是供给侧冲击引发的政策陷阱——霍尔木兹海峡地缘阻碍或气候波动等原油及农产品瓶颈无法通过利率调控消除；若决策层误将外部供给冲击引发的黏性通胀视作经济全面过热并施加紧缩，将对实体部门造成严重负面冲击。 关于本轮周期终结的潜在催化剂，讨论指向了财政支持退坡与非科技企业的“资本开支疲劳”。Dutta 指出，随着上一轮重大财政刺激红利出清，财政端将在年底至明年初转为净拖累。Amarnath 则警示，这与 2000 年千禧虫（Y2K）前的集中采购类似：科技巨头因护城河竞争仍在不计成本投入算力，但传统非科技企业必须将 AI 视作硬性营业成本并要求立竿见影的 ROI。一旦企业在试用后迟迟无法将技术转化为切实的生产率提升，软件与算力支出或将面临预算削减，而这一情绪传导至企业信用利差之日，便将是本轮繁荣周期面临最大考验之时。 当前美国经济由规模空前的 AI 基础设施投资与高净值财富效应强力支撑，彻底脱钩了以住房建设为主导的传统商业周期路径。然而，宏观风险已转移至微观结构层面：在企业利润高度挤占劳动所得、供给端通胀掣肘货币政策的背景下，实体经济最终能否软着陆，关键取决于传统企业对 AI 投资回报率的耐心极限何时耗尽，以及信贷市场的扩张动能能否经受住财政退坡的考验。

## Background 本期内容来自 YouTube 频道 The Compound，由主持人 Michael Batnick 于 2026 年 6 月 13 日发布。节目邀请了 Renaissance Macro Research 的 Neil Dutta 与 Employ America 的 Skanda Amarnath，围绕 AI 投资狂潮、劳动力市场真实韧性、通胀风险、消费者支出与美联储政策前景等核心宏观议题展开深度剖析。 ## TL;DR 节目深入探讨了当前美国经济在科技资本开支暴增与传统周期信号失灵下的复杂格局。嘉宾指出 AI 基础设施与工业相关投资占 GDP 比重已突破 7%，取代传统住宅地产成为商业周期的主导引擎，并拉动工业股与半导体高度趋同；但与此相对，名义薪酬放缓、个人储蓄率低迷以及国民收入向企业利润极端倾斜，正在加剧实际购买力受损与底层信贷压力。围绕美联储货币政策与凯文·沃什等潜在人事动向，嘉宾们警示了由供给冲击引发的非理性防御性加息风险，并强调终结本轮周期的关键在于非科技企业在 AI 支出上的实际投资回报率（ROI）及信用利差表现。 ## Atmosphere & style 整场对谈兼具华尔街专业智库的严谨锐利与纽约体育死忠的市井幽默。开场由麦迪逊广场花园天价球票和客场击溃马刺的兴奋闲聊切入，随即迅速转入高密度的宏观经济学辨析与数据挑刺，嘉宾们言辞辛辣、毫无外交辞令，甚至公开调侃央行官员的哲学玄谈；在深度的周期博弈、图表拆解与现实痛点碰撞中，始终穿插着对尼克斯队季后赛命运的狂喜与宿命感喟叹，听感畅快淋漓且信息量极其饱满。 ## Bulletin - [16:37](https://www.youtube.com/watch?v=swPQA7uJRgI&t=997s) 涵盖科技设备、软件与电力设施的 AI 资本支出已突破美国 GDP 的 7%，正式超越 2005 年房地产泡沫时期的住宅投资峰值及 1990 年代末互联网基建高峰，成为宏观经济的核心驱动引擎。 - [45:19](https://www.youtube.com/watch?v=swPQA7uJRgI&t=2719s) 第一季度美国国内企业利润增长 2.7%（占国民总收入比例升至 12.1%，创 1950 年来新高），而劳动者报酬仅增 0.8%（占比降至 51%，创 1947 年有记录以来新低），分配脱节加剧了社会分化。 - [37:07](https://www.youtube.com/watch?v=swPQA7uJRgI&t=2227s) 标普 500 指数中有 15 家非科技公司与半导体 ETF（SMH）日收益率相关性达 0.5 以上，合计市值达 2 万亿美元，其中 12 家为传统工业巨头，显示市场广度拓展本质上源于 AI 数据中心物理基建的拉动。 - [48:26](https://www.youtube.com/watch?v=swPQA7uJRgI&t=2906s) 名义 GDP 增速（约 6%）与劳动报酬增速（低于 4%）显著背离，导致美联储政策前景出现鹰鸽分歧：若收入随支出走高可能招致加息 75-100 个基点，若消费被弱薪资压低则央行具备按兵不动空间。 ## Key takeaways - [05:18](https://www.youtube.com/watch?v=swPQA7uJRgI&t=318s) 从纽约主场总决赛门票被炒至 7000 美元可看出，美股牛市催生的资产财富效应使得顶层高净值人群可选消费极为充沛，呈现出典型的 K 型经济分化与“财富去积累之谜”。 - [10:10](https://www.youtube.com/watch?v=swPQA7uJRgI&t=610s) 传统宏观经济风险分布发生位移，通胀黏性与价格压力靠近右尾分布，而就业市场未跌入失速左尾并展现韧性，使得金融条件整体宽裕、衰退叙事频频落空。 - [13:10](https://www.youtube.com/watch?v=swPQA7uJRgI&t=790s) 个人储蓄率低位往往是滞后指标而非预测衰退的良好择时工具；传统商业周期中“地产即周期”的定律在本轮被打破，AI 数据中心对土地、电力和建材的挤占让传统住宅周期退居次要位置。 - [18:28](https://www.youtube.com/watch?v=swPQA7uJRgI&t=1108s) 单纯的 GDP 统计低估了 AI 投资的宏观冲击，海外硬件采购构成的“进口漏损”虽压低了本土净 GDP 读数，但通过限制性股票（RSU）兑现与财富效应深刻提振了终端消费和地方税收。 - [21:41](https://www.youtube.com/watch?v=swPQA7uJRgI&t=1301s) 商业周期的本质是风险周期与支出意愿，科技巨头庞大的资产负债表与超 7000 亿美元的 Capex 承诺直接构成了当前周期的先导支撑，打破了传统设备投资落后于经济增长的规律。 - [24:38](https://www.youtube.com/watch?v=swPQA7uJRgI&t=1478s) 美联储褐皮书已显示信用卡、抵押贷款和农业贷款违约率边际抬头，底层家庭在薪资增速放缓、高物价与高利率并存的挤压下资产负债表正在承受侵蚀。 - [35:07](https://www.youtube.com/watch?v=swPQA7uJRgI&t=2107s) 本轮所谓的“生产力繁荣”存在宏观悖论：科技与软件硬件价格不跌反涨，家庭实际薪酬甚至面临负增长，技术带来的超额收益几乎全部沉淀在企业端的高利润率中。 - [37:07](https://www.youtube.com/watch?v=swPQA7uJRgI&t=2227s) 今年标普 493 跑赢科技七巨头并非各行业均衡的内生复苏，而是由于卡特彼勒、伊顿、维谛技术等传统工业股承接了海量数据中心电源、散热与发电设备订单，成为变相的 AI 概念股。 - [43:45](https://www.youtube.com/watch?v=swPQA7uJRgI&t=2625s) 消费呈现明显的由“商品”向“服务”倾斜，餐饮杂货受食材成本高企及 GLP-1 减肥药普及影响显露疲态，而酒店与航空出行等休闲体验开支持续创新高。 - [52:21](https://www.youtube.com/watch?v=swPQA7uJRgI&t=3141s) 针对潜在美联储主席候选人凯文·沃什，市场对其推销的格林斯潘式“黄金时代”生产率红利叙事存疑，且美联储委员会的集体投票机制及日益抬头的鹰派异议将对极端鸽派主张形成强力制约。 - [54:02](https://www.youtube.com/watch?v=swPQA7uJRgI&t=3242s) 当前通胀多发性源自霍尔木兹海峡地缘局势、航空燃油及大宗原料等供给冲击，货币政策抬高隔夜利率对此类瓶颈无能为力，若央行误将其视为需求过热而加息，实体经济与市场将承受严重挫伤。 - [01:00:35](https://www.youtube.com/watch?v=swPQA7uJRgI&t=3635s) JOLTS 职位空缺数据单月激增 70 万充斥着虚假招聘等统计噪音，但初请失业金回落及白领招聘企稳表明劳动力市场已自发筑底，AI 摧毁就业的担忧属于劳动力总额谬误。 - [01:02:54](https://www.youtube.com/watch?v=swPQA7uJRgI&t=3774s) 企业在 AI 时代裁员面临流失系统核心架构知识的巨大风险，大规模裁员通常是经济恶化的末期滞后指标而非先行信号。 - [01:05:08](https://www.youtube.com/watch?v=swPQA7uJRgI&t=3908s) 判定本轮周期拐点的核心变量在于非科技实体企业在 AI 上的支出意愿与投资回报率（ROI）；一旦 Token 采购等日常成本难以为继，可能引发类似 2000 年 Y2K 需求透支后的周期性回落。 ## Detailed walkthrough ### [00:00](https://www.youtube.com/watch?v=swPQA7uJRgI&t=0s) 尼克斯客场包机之旅与客场球迷的溃败 节目在一阵轻松热烈的闲聊中开场。主持人好奇嘉宾是如何搞到那趟特殊航班机票的。嘉宾透露，自己联系了券商的一位名叫 John 的朋友，直接要求搞到总决赛第三场（Game 3）的门票。对方一开始表示根本办不到，但随后带来了一个名为“尼克斯客场飞行体验”（Knicks Flyaway Experience）的难得机会。他和妻子乘坐由达美航空（Delta）客机改造的包机直飞圣安东尼奥，机上坐满了尼克斯的铁杆球迷，还安排了见面会，偶遇了克莱德·弗雷泽（Clyde Frazier）、约翰·斯塔克斯（John Starks）以及冰球名宿“国王”亨里克·伦德奎斯特（Henrik Lundqvist）等传奇球星。 更令他得意的是，尼克斯官方 Instagram 账号还专门发布了他和妻子的现场合影，被同行调侃为“帅哥靓妹”。他笑称，这比上彭博社（Bloomberg）或 CNBC 露脸还要爽，尽管还是比不上来这档播客做客。目睹尼克斯在客场彻底击碎主场观众的希望与梦想，看着圣安东尼奥马刺球迷在绝望中提前退场，给他带来了极大的满足感。他特别提到杰伦·布伦森（Jalen Brunson）不可思议的稳定发挥：当时在费城的客场比赛中，一位主场球迷眼睁睁看着布伦森连续命中，气得怒掷帽子大喊“他怎么可能一个都不投丢”，那一幕令客场尼克斯球迷心花怒放。 --- ### [02:17](https://www.youtube.com/watch?v=swPQA7uJRgI&t=137s) 天价球票行情、转卖风波与巧合 话题随即转到纽约主场的球票疯狂行情上。主持人分享了自己持有的麦迪逊广场花园季票（109 区第 20 排）的亲身经历：他与一位朋友合买季票，首轮票价一张 380 美元，随后逐轮涨至 550 美元、750 美元，到总决赛时单张票面价已高达 1,350 美元。不料朋友擅自以每张 4,500 美元的价格将第四场（Game 4）的球票转卖，获利 3,000 美元，这让他十分抓狂，因为他根本没打算卖掉如此关键的比赛门票。 此时纽约主场的门票二级市场已经彻底失控，入场门槛价（get-in price）直逼 7,000 美元，相比之下超级碗（Super Bowl）的入场门槛价不过 3,500 美元左右，总决赛票价直接翻了一倍；同区域位置甚至有人挂出 17,000 美元的天价。为了能进场，他们不得不通过票务经纪人置换位置较差的席位，勉强做到盈亏平衡。 前一晚，他在麦迪逊广场花园参加观赛派对（Watch Party），现场涌入了 18,000 名球迷。由于 10 美元的官方原价票在两分钟内被秒光（排队人数多达 33 万人），他花了 60 美元在二级市场购票入场。现场几乎全是最硬核的死忠球迷，没有无动于衷的企业赞助商包厢客，声浪震耳欲聋。戏剧性的是，赛后他与 Future Proof 的 Matt Middleton 及 Chris Cherry 一起喝酒时，Middleton 竟然发现正是自己的内弟（妹夫）买下了主持人转卖出去的那对季票。主持人当场笑骂着让对方把票还回来，令人感叹世界之小。 - **季票成本递增**：首轮 380 美元 $\to$ 次轮 550 美元 $\to$ 东决 750 美元 $\to$ 总决赛 1,350 美元 - **二级市场售价**：总决赛第四场被朋友以 4,500 美元售出 - **入场最低门槛**：纽约主场起步价高达 7,000 美元（约为超级碗 3,500 美元起步价的两倍） - **观赛派对规模**：18,000 人挤满麦迪逊广场花园，官方 10 美元门票二级市场溢价至约 60 美元 --- ### [05:18](https://www.youtube.com/watch?v=swPQA7uJRgI&t=318s) 疯狂票价背后的“K 型复苏”与财富不缩水之谜 针对如此离谱的高昂票价，嘉宾们从宏观经济角度展开了深度探讨。这种不计成本抢购天价门票的现象，本质上是美股长牛背景下高净值人群消费力极度充沛的缩影。 ```mermaid flowchart TD accTitle: K-Shaped Economy and High-End Ticket Prices accDescr: 股票市场繁荣推动顶层财富扩张，即使就业市场降温，超高净值阶层依然推高了奢侈文体消费的门槛价格。 subgraph TopTier["顶层财富端 (K型上方)"] A["股市长牛与资产升值"] --> B["超高净值群体财富充盈"] B --> C["高额非必需支出爆发\n(尼克斯总决赛、世界杯)"] C --> D["极高入场门槛价\n($7,000 纽约主场)"] end subgraph MacroDivergence["宏观分化现象"] E["就业市场逐步降温放缓"] -.->|脱钩| C F["财富去积累之谜\n(Wealth De-accumulation Puzzle)"] --> B end ``` 嘉宾指出，尽管广泛的劳动力市场数据出现放缓迹象，但处于收入与财富分配顶端的高收入阶层，其可选消费（discretionary spending）依然处在繁荣期，直接推动了典型的“K 型”分化。纽约市本身就是一个脱离普通现实的极端财富聚集地，这也就解释了为何得克萨斯州虽然在人口和经济增速上被称为全美增长引擎，但面对由纽约客蜂拥而至推高的门票与客场观赛狂潮，依然显现出老牌金融资本中心的绝对财富压制。 经济学家们经常讨论的“财富去积累之谜”（Wealth De-accumulation Puzzle）在此得到了现实印证：即极端富裕阶层往往很难随着年龄增长把财富真正花光，而此类罕见的文体盛事恰恰为他们提供了挥霍剩余流动性的出口。不仅是 NBA 总决赛，即将在北美（美国波士顿、亚特兰大、纽约/新泽西，以及加拿大多伦多、墨西哥城等地）联合举办的世界杯门票同样水涨船高。相比之下，不同场次间的二级市场票价分化极其剧烈：圣安东尼奥第五场入场门槛约 1,700 美元，而第二场则从几天前的 1,000 美元迅速滑落至 600 至 700 美元。 --- ### [07:30](https://www.youtube.com/watch?v=swPQA7uJRgI&t=450s) 赞助商插播与节目正式开场：嘉宾介绍 闲聊结束后，主持人正式拉开 The Compound and Friends 第 245 期的序幕。由于常规搭档 Josh Brown 本周缺席，节目组特意邀请了经济学家 Skanda Amarnath 与老朋友 Neil Dutta 共同坐镇。 随后插播赞助商 Nuveen 的广告：面对通胀波动与不断变化的宏观经济环境，Nuveen 强调其横跨房地产、私人信贷、基础设施和自然资本的全面私募市场平台，致力于帮助投资者的多元化投资组合跨越传统资产类别进行配置。 广告结束后，主持人难掩激动之情，庆祝尼克斯终于拿下了一场久违的总决赛胜利。随后他向听众正式介绍了本期重磅嘉宾 Neil Dutta。Neil 是 Renaissance Macro Research（文艺复兴宏观研究公司）的经济学主管，长期主导宏观经济研究，专精于美国宏观经济、美联储货币政策动向、全球宏观趋势以及跨资产投资主题分析，是该播客听众极为喜爱的常客。 ### [09:29](https://www.youtube.com/watch?v=swPQA7uJRgI&t=569s) 嘉宾介绍：当市场经济学家遇上政策倡导者 主持人 Michael Batnick 向观众正式介绍了当期的两位重磅嘉宾。首先是老朋友、文艺复兴宏观研究（Renaissance Macro Research）的经济学家 Neil Dutta。Michael 调侃他到底算不算“股市经济学家”，Neil 则纠正自己是专注剖析数据并寻找共识漏洞的“商业经济学家”。随后，Michael 欢迎了第一次做客节目的 Skanda Amarnath。作为政策研究智库 Employ America 的联合创始人兼执行总监，Skanda 带领团队深耕就业与通胀数据分析，致力于为美联储的政策沟通与战略制定全新框架。在此之前，他曾在纽约联储的研究部门从事资本市场分析。 ### [10:10](https://www.youtube.com/watch?v=swPQA7uJRgI&t=610s) 宏观经济走势分歧：是通胀右尾还是增长“湿软”？ 在进入核心的 AI 议题前，Michael 邀请两位嘉宾登高俯瞰，为当前的美国经济把脉。Skanda 指出，过去半年来宏观风险的分布已经发生了微妙而剧烈的位移。在 6 到 12 个月前，市场的主流叙事还是就业市场放缓、通胀顺利回落，除了狂热的 AI 之外，其余大部分经济部门都显得“湿软软”（soggy）且缺乏动能。然而现实却给出了截然不同的答卷：通胀重新抬头，这不仅受霍尔木兹海峡局势与原油价格的驱动，整体价格压力也更靠近分布的“右尾”；与此相对，就业市场并未掉入失速的“左尾”，反而在部分长期疲软的行业出现了微弱回暖。唯一的常数是，AI 繁荣依旧在以不可阻挡之势向前奔涌，这使得整体金融条件格外宽松、经济增速优于预期，但通胀黏性也远高于决策层所乐见的水平。 Neil 对这种“湿软”（soggy）的形容深表赞同，并坦言自己已经持续谨慎悲观了一年半之久。他回顾道，在近期伊朗相关地缘冲突爆发前，美国 10 年期国债收益率一度下探至 3.93%，市场甚至在押注年内更多的降息幅度。尽管眼下就业市场不像去年年底那么羸弱，但家庭部门的裂痕依然在悄然扩大。Neil 最担忧的核心隐患在于整体薪资增长持续放缓，而在联邦基金利率依然高企、物价居高不下的双重夹击下，美国家庭正在过度透支储蓄来支撑日常开销。他强调，靠消耗储蓄拉动的消费终有极限，家庭资产负债表正在积聚越来越大的压力。 ### [13:10](https://www.youtube.com/watch?v=swPQA7uJRgI&t=790s) 储蓄率暴跌与地产冰封：老指标为何频频失灵？ 面对高利率下陷入“绝对冰河期”（absolute ice age）的房地产市场，经济何以展现出超预期的韧性？Skanda 对消费者的坚挺程度感到惊讶，目前个人储蓄率已经接近崩溃，说明消费支出的增长明显跑赢了工薪阶层的实际到手薪水。不过 Michael 提出了另一种市场心理学视角，他援引自己在《Animal Spirits》播客中与 Ben Carlson 的讨论反驳道：消费者往往在对未来感到极度乐观时，才敢于放心大胆地把储蓄花光，因此储蓄率走低不一定意味着灾难即将降临。 Neil 从经济学计量方法上印证了 Michael 的直觉。他指出，储蓄率向来是一个极其糟糕的择时指标： - 早在 2005 年，官方公布的储蓄率甚至一度跌为负值，引来大批空头高呼缓冲垫已彻底耗尽、经济崩溃在即，但随后美国经济依然持续增长了整整两年； - 统计局日后往往会通过数据修正“凭空发现”更多被遗漏的收入，从而大幅推高此前被低估的储蓄率； - 官方储蓄率本质上只是国民收入账户中“收入减支出除以收入”的会计残差，极易失真。 针对营建许可等传统衰退领先指标，Neil 指出著名经济学家 Edward Leamer 曾提出“住房就是商业周期”（Housing is the Business Cycle）的经典论断，但在本轮周期中，传统的新屋销售与住宅建筑面临建筑商利润承压、同比下滑 6% 至 7% 的窘境，更关键的是，住宅开发正在土地、电力与建筑资源上遭到 AI 数据中心的正面挤占与争夺。 ```mermaid flowchart TD accTitle: AI投资与传统地产周期的资源竞争关系 accDescr: 流程图展示了高利率环境下传统地产走弱与AI资本开支暴增如何重塑宏观经济周期。 subgraph MacroEnvironment["当前宏观环境"] Rates["高利率约束与黏性通胀"] Savings["个人储蓄率低位徘徊"] end subgraph DivergentSectors["分化的经济动能"] Housing["传统住宅地产<br>新屋销售下滑 6%-7%<br>利润承压且陷入冰河期"] AICapex["AI 基础设施浪潮<br>数据中心 / 电力 / 科技装备<br>占 GDP 比重突破 7%"] end subgraph CycleImpact["打破传统商业周期规律"] ResourceCompetition["资源挤占：土地、电力与资本竞争"] NewRegime["打破'地产即周期'定律<br>科技周期直接主导宏观走势与 r* 争论"] end MacroEnvironment --> DivergentSectors Housing -.-> ResourceCompetition AICapex --> ResourceCompetition ResourceCompetition --> NewRegime ``` ### [16:37](https://www.youtube.com/watch?v=swPQA7uJRgI&t=997s) 狂飙的 AI 资本开支：正式超越 2005 年地产泡沫峰值 随后，Michael 翻出了一段令人惊叹的预言记录。他在做功课时搜到了 Skanda 于 2025 年 1 月撰写并发表的观点（当时曾与彭博 Odd Lots 节目的 Joe Weisenthal 和 Tracy Alloway 共同探讨）： > “我们现在正处于科技周期与商业周期深度融合的阶段。从经济衰退动态，到美联储关于潜在产出增长率与中性利率（$r^*$）的激烈辩论，AI 必将在未来几年的宏观经济讨论中烙下深刻印记。到 2025 年底，AI 繁荣占美国经济的比重，完全有可能比肩 2005 年住宅投资在房地产泡沫巅峰时占 GDP 近 7% 的历史水平。” Skanda 兴奋地确认，这一判断不仅得到了完美验证，甚至进程更加迅猛。如今，涵盖科技设备、软件开发以及为海量数据中心供电供能的工业设备等 AI 相关投资总和，占美国 GDP 的比重已经强力“撕裂并冲破”（ripped past）了 7% 的大关。这意味着，当下的 AI 资本开支不仅一举超越了 2000 年代中期美国房地产泡沫顶峰时期的住宅投资总额，更是将 1990 年代末互联网科技泡沫（Dot-com boom）以及世纪之交电信光纤基建投资的历史峰值远远甩在了身后，科技投资已正式接管全美宏观周期的主驾驶座。 ### [18:28](https://www.youtube.com/watch?v=swPQA7uJRgI&t=1108s) AI 投资繁荣与传统 GDP 核算的盲区 嘉宾们深入探讨了当前由科技主导的经济扩张。Skanda Amarnath 指出，储蓄率下降等特征往往出现在周期顶点，这表明一旦繁荣发生逆转，下行风险同样巨大。然而，眼下完全没有任何迹象表明支出正在放缓；相反，资本市场正涌现出越来越多的融资、发债和资本支出，全都在围绕这一轮 AI 技术与投资浪潮进行押注，整个宏观环境处于典型的“全面繁荣期”（go-go time）。 ```mermaid flowchart TD accTitle: AI 投资支出对宏观经济的传导路径 accDescr: 展现从美国本土科技支出到海外供应链进口漏损，以及财富效应推动消费和财政收入的机制。 subgraph TechSpend["科技巨头投资与支出"] A["美国本土 AI 资本开支 (Capex)"] --> B["海外供应链 (SK海力士 / 日韩硬件)"] A --> C["企业盈利增长与股价上涨"] end subgraph MacroImpact["宏观经济扩散效应"] B -.->|"进口漏损 (Import Leakage)"| D["压低美国纯 GDP 读数"] C --> E["限制性股票 (RSU) 解锁兑现"] E --> F["州政府所得税扣缴激增 (如加州)"] C --> G["资产财富效应 (Wealth Effect)"] G --> H["强劲的终端消费支出 (如尼克斯门票)"] end ``` 针对华尔街部分经济学家认为“AI 对宏观经济影响相当微弱”的观点，Neil Dutta 表达了强烈质疑。他解释道，这主要是传统国民经济核算（GDP Accounting）带来的错觉： - **进口漏损（Import Leakage）**：虽然信息处理与软件仅占 GDP 的 6% 至 7% 左右，但大量 AI 支出实际上流向了海外硬件（如 SK 海力士的存储芯片、日韩高端硬件与能源系统）。在会计上，这些进口算作 GDP 的减项，导致巨额本土支出未能体现在美国 GDP 读数中，但这些支出依然是真实存在的经济驱动力。 - **被忽略的连锁反应**：简单的核算公式忽略了复杂的次生经济联系。企业通过 AI 繁荣实现巨额盈利，推高了员工薪酬体系中的限制性股票（RSU）兑现规模。例如，加州立法分析办公室（LAO）的数据显示，该州个人所得税代扣代缴增长的一半以上都来自于 RSU 的解禁兑现。 - **财富效应传导**：强劲的企业盈利与股市上涨通过财富效应反哺终端消费支出（从日常消费一路延伸到昂贵的纽约尼克斯队门票）。Neil 形象地反驳称，如果只死抠会计公式，那也可以说“住宅投资只占 GDP 的几个百分点，而且还要扣除从加拿大进口的石膏板和木材”，这种逻辑显然脱离了经济现实。 --- ### [21:41](https://www.youtube.com/watch?v=swPQA7uJRgI&t=1301s) 商业周期本质：风险偏好、支出与衰退前瞻 Skanda 赞同 Neil 的看法，并认为死守 GDP 核算完全误解了经济繁荣的本质。他强调：“商业周期本质上就是风险周期”（Business cycles are risk cycles）。繁荣的核心在于支出意愿——无论是招聘劳动力的支出还是资本开支。这些支出沉淀在各大科技巨头的资产负债表上，体现着当下极高的风险偏好；虽然这些巨头的资产负债表在历史上属于最稳健的一批，但支出规模之庞大、增速之迅猛，本身就蕴含着巨大波动。 谈及投资者应如何研判经济拐点时，Michael Batnick 提出疑问：传统的先行指标（leading indicators）是否已经失效？未来的衰退是否会像暴风雨般突如其来，伴随着股市暴跌与开支骤停，几乎不留任何预警信号？ - **历史镜鉴（2000 年互联网泡沫）**：Neil Dutta 回顾指出，2000 年泡沫破裂前同样没有任何单一的宏观指标（如房屋开工率断崖）提前发出警报，真正的转折点在于当年 4 月左右企业盈利开始不及预期，随后引发连锁崩塌。 - **提前透支的支出需求**：当年的繁荣伴随着千年虫（Y2K）和 IT 系统建设的严重需求前置（spending pulled forward）。每当指数级增长曲线在现实中被过度外推，一旦支出难以为继，就会迎来清算时刻。 - **美联储加息的局限性**：在互联网泡沫中，即便美联储持续加息，股市依然逆势上涨。Neil 直言，当下指望美联储通过强硬加息来“刺破泡沫”是不现实的，这不仅难以降伏科技支出，反而会进一步恶化房地产和信贷敏感部门的生存压力。 --- ### [24:38](https://www.youtube.com/watch?v=swPQA7uJRgI&t=1478s) 劳动力市场隐忧、违约率微升与顽固通胀 嘉宾们进一步将当前的宏观环境与 1990 年代末进行对比，梳理出两大关键差异： | 宏观维度 | 1990 年代末（Dot-com 繁荣期） | 当前周期（AI 投资浪潮） | |:--- |:--- |:--- | | **劳动力市场** | 极度过热，出现广泛的工资上涨压力 | 处于稳定但偏弱状态，未见全行业工资螺旋上升 | | **通胀环境** | 整体处于温和受控状态 | 面临关税、地缘政治（霍尔木兹海峡/航运成本）、高昂航油以及 AI 硬件涨价等多重外生压力 | | **信贷违约表现** | 后期逐步显现压力 | 信用卡、抵押贷款与农业贷款违约率边际抬升 | 虽然家庭部门整体流动性依然相对充裕，但底层断层已经浮现。Neil Dutta 援引最新的美联储《褐皮书》（Beige Book）数据指出，抵押贷款、信用卡和农业贷款的违约率正在边际恶化。如果劳动力市场无法重新加速、薪资增速持续放缓，这些滞后暴露的信贷问题将会像“惯性滑行的物体”一样持续发酵，银行业未来一年势必需要计提更多贷款损失准备金。 Skanda 总结道，虽然缺乏薪资通胀压力在表面上减少了美联储进一步紧缩的借口，但本轮周期的棘手之处在于通胀具有多发性——关税政策的不确定性、中东航运受阻推高的航空燃油与机票价格，乃至 AI 繁荣自身推高的电子硬件成本（例如现在购买一台新笔记本电脑的价格已今非昔比）。若未来半年劳动力市场无法重现活力，宏观经济将重新面临严重的左尾下行风险。 ### [26:21](https://www.youtube.com/watch?v=swPQA7uJRgI&t=1581s) 行业通胀痛点与消费者财富效应 Skanda Amarnath 和 Neil Dutta 深入探讨了硬件短缺带来的连锁反应。内存及各类电脑硬件短缺已产生实质影响，这种各行业层面的摩擦让人联想到 2000 年代初——当时各种突发性通胀悄然抬头，最终买单的只能是消费者。对普通人来说，这不仅没能提升生活水平，某些层面甚至导致了生活质量下滑。与 1990 年代那种劳动力市场强劲、通胀几乎没有抬头的“极乐净土（nirvana）”相比，当前的经济环境在劳动力市场表现上稍逊一筹，而在通胀端则显得更加严峻和刺手。 随后，Michael Batnick 将话题转向消费支出与财富效应（wealth effect）。他指出，自己和 Josh Brown 在节目里经常讨论究竟是什么驱使人们花更多钱。Batnick 直言他并不认同“房产升值能直接拉动日常消费”的观点，因为房价波动并不会天天直接显示在屏幕上；相反，投资账户的走势会直接滋生过度自信：“你从投资账户里提走一笔钱，拿去装修房子或者买个大玩具，结果过了 10 天，这个水桶又被填满了——因为半导体股票一直在源源不断地给你发大富翁式的游戏币。” 股市带来的财富效应毫无疑问支撑了当下巨额的终端消费。 ```mermaid flowchart TD accTitle: 财富效应与资本开支驱动逻辑 accDescr: 展示半导体股票收益刺激消费以及科技巨头资本开支维系商业周期的双重循环结构。 subgraph 消费者端 A["半导体/科技股大幅上涨"] --> B["投资账户账面浮盈激增"] B --> C["消费信心膨胀，提取资金用于大额支出"] C --> D["资产价格持续上涨填补消费缺口"] end subgraph 周期驱动端 E["超大规模云厂商资本开支<br/>(约7000亿美元)"] --> F["非住宅商业固定投资激增"] F --> G["主导商业周期与GDP波动"] end ``` ### [27:36](https://www.youtube.com/watch?v=swPQA7uJRgI&t=1656s) 2022年熊市的对比与消费放缓的非线性风险 回顾 2022 年的市场回调，当时出现了一轮惨烈的熊市，虽然没有持续整整五年，但大家青睐的核心科技股大多腰斩——Amazon、Facebook（Meta）以及暴跌了近三分之二的 Nvidia，大量账面财富灰飞烟灭。但令人意外的是，消费并没有戛然而止。Neil Dutta 解释道，根本原因在于当时拥有真实的收入增长（income growth），而且居民手中还握有疫情期间积累的大量超额储蓄，这与当前的格局完全不同。 当被问及“究竟需要发生什么才会让消费真正放缓”时，Neil Dutta 分析指出，经济学家关注的核心在于这种放缓究竟是线性的还是非线性的： - **当前的线性放缓**：目前消费支出年化增长率在 2% 左右窄幅波动，随着收入增速变得疲软，如果储蓄率维持不变，消费自然会略微回落至 1% 到 1.5%，但绝非断崖式下跌。 - **潜在的非线性冲击**：如果科技行业之外出现大规模实质性裁员，或者市场出现剧烈的大幅调整，消费端可能会迅速遭到重创。 ### [29:14](https://www.youtube.com/watch?v=swPQA7uJRgI&t=1754s) 超越传统周期的科技投资繁荣 面对科技巨头（Hyperscalers）和其他参与者高达 7000 亿美元的资本开支，消费者即便稍微缩减开支，整体大局也不会受到致命影响。Skanda 指出，大众常说“美国消费者就是美国经济”，但从商业周期角度来看，**投资才是商业周期的核心**（"investment is the business cycle"）。消费通常是平滑稳定的，唯有投资具备高度波动性。目前市场明显处于 Risk-on（追逐风险）状态，大量电网能源扩容和数据中心建设的规划资金已承诺投入，只要资本市场确信这个叙事在延续，就能牢牢托住市场情绪与走向。 Neil Dutta 补充道，这一轮周期的反常之处在于，驱动力完全来自设备投资和非住宅固定投资。在传统商业周期中，最先破裂并引发转折的通常是住宅投资，而资本开支设备投资通常是落后于增长的滞后指标（即所谓的加速数效应 accelerator effect，企业预期经济增长才会追加投资）。但在本轮周期中，整体宏观经济增长预期并没有大幅起飞，却爆发了一场惊人的资本支出狂潮。科技行业不再像传统认知那样独立于周期之外，它如今已经取代了房地产和传统制造业，成为驱动 GDP 波动和美股行情的最大周期性力量。 - **超预期资本开支**：云服务商及科技企业约 7000 亿美元的 Capex 承诺为硬件、能源和数据中心注入强大动能。 - **反常的周期先行指标**：非住宅商业固定投资违背历史滞后性规律，在宏观增速平淡的背景下独立领跑。 - **缺乏前瞻指标**：除极少数科技供应链的专属订单外，传统宏观指标几乎无法提前预警科技驱动型周期的拐点。 ### [31:25](https://www.youtube.com/watch?v=swPQA7uJRgI&t=1885s) S曲线拐点与“伪生产力繁荣”的质疑 市场情绪的轮动极其迅速。Batnick 回忆道，就在前不久的秋季，市场还在大举抛售“美股七巨头（Mag 7）”，Oracle 的股价一度因市场怀疑 Sam Altman 无法兑现 5 年 3000 亿美元的巨额合同而大跌 60%，随后却又被迅速推向泡沫狂欢。Batnick 强调，如今一切都被这场投资与基建浪潮拉动，哪怕终端大众根本还没广泛用上智能体（Agentic AI）这类前沿工具。在利率与通胀波动的背景下，Anthropic 的营收从 900 万美元直接飙升至 4500 万美元级别，算力严重供不应求。他认为指责周期见顶还为时过早，当下的繁荣可能才刚刚开始。Skanda 对此表示赞同，指出虽然所有指数级曲线（exponential curves）最终都会蜕变为 S 曲线（S-curves），但眼下拐点显然尚未到来。 然而，Neil Dutta 对这一乐观叙事提出了尖锐质疑。他认为在华尔街看来，极高的数据中心建设支出支撑了亮眼的短期盈利，但仅仅因为估值倍数（PE）没有显得极端离谱，并不意味着不存在泡沫风险——一旦数据中心基建狂潮暂停，盈利也将面临考验。更为关键的是宏观经济层面的悖论：投资的终极目的是提高家庭的生活水平，但在本轮所谓的“生产力繁荣”中，IT 产品、软件和笔记本电脑的价格不但没有像 1990 年代或 2000 年代那样持续通缩，反而出现上涨；与此同时，强劲的生产力没有转化为实际薪酬的增长，今年实际薪酬甚至出现了负增长，所有超额收益完全流向了企业的利润率。“当增长全归于利润率而居民实际报酬为负时，我们真的能称这是一场生产力繁荣吗？” ### [35:07](https://www.youtube.com/watch?v=swPQA7uJRgI&t=2107s) 生产率神话与现实：AI 落地受阻与宏观供给冲击 当前市场上对生产率的过度乐观叙述，往往忽视了现实宏观数据的疲态。回顾第一季度的数据，生产率表现并不理想，甚至可以说是疲软的。与此同时，经济体系正受到一系列现实供给冲击的夹击：汽油与各类能源价格再度抬头、机票价格攀升、甚至活牛与牛肉等大宗消费原材料价格的持续上涨，都会不可避免地传导至下游。这些供给约束不仅会对通胀形成推力，更会直接拖累整体生产率增长。相比 2023 年至 2024 年期间的乐观预期，眼下宏观层面的生产率图景已经明显蒙上了阴影。 关于 AI 究竟在多大程度上提振了实体经济，嘉宾们指出了技术扩散（Diffusion）中的巨大鸿沟。虽然科技圈与播客听众对生成式 AI 趋之若鹜，但在真实的工商业世界中，大规模采用依然步履蹒跚。中小企业因为船小好掉头，可以无所顾忌地尝试各种新工具；但对于大型跨国公司而言，数据合规、隐私与安全壁垒高耸，企业内部通常被严格设限，员工“只能使用微软的 Copilot 这类指定产品，其他第三方工具一概被防火墙拦截”。 数据追踪的失真也加剧了误解。例如市场热衷于跟踪商业支出平台 Ramp 的指数，将其作为衡量创新的风向标，但诸如“OpenAI 份额下滑、大家都在转用 Claude”这类数据并不能说明实体经济是否发生了根本变革。市场普遍预设 AI 必然带来全行业的利润率扩张，但如果成本红利无法真正转化为家庭生活水平的改善与消费品价格的下降，所谓的“生产率繁荣”就只是一场局限于企业内部与技术极客圈子的自嗨。 --- ### [37:07](https://www.youtube.com/watch?v=swPQA7uJRgI&t=2227s) 工业股变身“算力底座”：市场广度背后的资本开支真相 美股表面上的“百花齐放”，实际上深嵌在 AI 产业链的资本支出之中。根据尼尔·杜塔（Neil Dutta）的一项量化分析，将标普 500 指数所有成分股与半导体 ETF（SMH）进行一年期日收益率相关性测试，结果令人震惊：有 15 家非科技公司与半导体指数的相关系数达到了 0.5 甚至更高，这 15 家公司总市值高达 2 万亿美元，其中 12 家是纯粹的传统工业制造企业。 这一名单涵盖了维谛技术（Vertiv）、伊顿（Eaton）、卡特彼勒（Caterpillar）、康明斯（Cummins）、哈贝尔（Hubbell）以及舒适系统（Comfort Systems）等老牌公司。这些传统工业股之所以走势神似半导体板块，是因为它们的在手订单已经彻底转型为“AI 资本开支订单”：卡特彼勒向超大规模云服务商的数据中心销售备用发电机组与重型引擎，维谛技术提供关键的数据中心冷却与电源管理设备，伊顿供应核心电气组件，而通用电气 Vernova（GE Vernova）则为耗电巨大的数据中心供应燃气轮机。传统的 GICS（全球行业分类标准）严重滞后于底层真实的经济敞口。 ```mermaid flowchart TD accTitle: AI 资本开支向传统工业板块的传导机制 accDescr: 超大规模数据中心需求拉动传统重工业设备供应商，导致工业股与半导体走势高度趋同 subgraph AI_Core["AI 资本开支核心"] A["科技巨头 / 超大规模云厂商 (Hyperscalers)"] end subgraph Infrastructure_Demand["数据中心关键物理基建"] B["备用电力保障"] C["制冷与电源管理"] D["电网与重型发电"] end subgraph Industrial_Suppliers["转型'AI概念股'的传统工业巨头"] E["卡特彼勒 (Caterpillar) / 康明斯 (Cummins)"] F["维谛技术 (Vertiv) / 伊顿 (Eaton)"] G["通用电气 Vernova (GE Vernova)"] end A --> B & C & D B --> E C --> F D --> G ``` 这种产业渗透彻底改变了市场广度的逻辑。从年初至今的表现来看，“科技七巨头”（Mag 7）上涨了 7.3%，而标普 500 其余的 493 只成分股（S&P 493）却上涨了 12.6%。然而，标普等权重指数走强并不意味着广泛行业的全面内生复苏，更多是由于 AI 数据中心建设在整个供应链中引发了强烈的连锁反应。例如 Generac 依靠数据中心备用电源迎风翻盘，福特（Ford）甚至试图通过销售电池切入电力热潮，连带着全国范围内的工业物流与货运需求也被动提升。这不是传统意义上经济各领域均衡发展的广度，而是“整个资本开支极度集中在单一领域所激发的系统性联动”，在风险偏好高企的牛市环境中，资产相关性正迅速趋近于 1。 - **半导体高相关非科技股**：标普 500 中 15 家非科技企业与 SMH 收益率相关性 $\ge 0.5$，合计市值达 2 万亿美元，其中 12 家属于工业行业。 - **美股分化数据逆转**：年初至今，“科技七巨头”涨幅为 7.3%，而“标普 493”涨幅达到 12.6%，等权重指数在工业股拉动下跑赢。 --- ### [40:52](https://www.youtube.com/watch?v=swPQA7uJRgI&t=2452s) 股市是否在计入经济放缓？非必需消费股与通胀挤压 面对客户频繁提问“美股是否正在计入经济放缓预期”，市场信号展现出极大的复杂性。一些分析师以非必需消费板块（Consumer Discretionary）相对必需消费品（Consumer Staples）的走弱为由，担忧经济降温；特别是餐饮股表现疲软、同店销售额普遍受压，市场上甚至出现了“大家都在打减肥针（GLP-1），导致餐饮消费下降”的段子。 然而，迈克尔·巴特尼克（Michael Batnick）认为，不能简单将可选消费股的下挫与宏观衰退划等号。很多连锁餐饮股（如 Cava 或 Shake Shack）之前被过度炒作，IPO 估值过高，如今的大跌更多是估值泡沫破裂和消费者偏好改变的结果——“现在没人中午还去吃汉堡了”。此外，一元店（Dollar Tree 和 Dollar General）等平价零售商业绩的波动同样极难解读：股价上涨究竟意味着底层人群收入改善，还是中产阶级被迫“降级消费”寻找折扣？单凭选股层面的杂音很难断定宏观放缓。当前大盘市盈率轻微收缩，更像是市场内生的健康调节器，而非经济衰退的前兆。 斯坎达·阿马拉纳特（Skanda Amarnath）则从成本端给出了更深层的隐忧：居民总名义支出虽仍在健康区间，但消费者面临的“二次通胀挤压”正日益加剧。不仅能源和牛肉等日常食品居高不下，必需消费品企业在业绩指引中明确表示，将把上游石化原材料（用于生产各类家用清洁剂、纸制品等生活消耗品）的涨价完全转嫁出去。企业的目标是誓死捍卫自身的营业利润率，这意味转嫁的成本将沿着零售商一路向下，最终无情地落在了终端消费者身上。虽然名义资金仍在经济循环中流动，但家庭部门正遭受实际购买力的真正挤压，“资金虽在不断转手，但我们并没有因此变得更富有”。 ### [43:45](https://www.youtube.com/watch?v=swPQA7uJRgI&t=2625s) 消费分化：商品消费降温、服务出行火爆与 GLP-1 冲击 嘉宾们围绕不同板块的股票走势展开探讨，指出当前许多市场叙事实际上取决于投资者的偏好而非宏观经济全貌。以高乐氏（Clorox）为例，其股价表现惨淡，可能面临自身独特的竞争与经营压力；但与此同时，凯悦酒店（Hyatt）和达美航空（Delta Air Lines）等股票却创下历史新高。因此，达美乐披萨（Domino's Pizza）等单家企业的股价波动只是特例，不能代表整体消费者的健康状况。如果达美乐或奇波雷（Chipotle）的 CEO 告诉分析师宏观消费承压，那只是管理层惯用的借口，市场上能轻易找到上百万个截然相反的反例来推翻这一说法。 真实数据清晰地反映出从“商品”向“服务”的结构性轮转。商品端和日用杂货店的消费确实显得疲软，但服务业中的航空出行极为强劲——尽管机票价格大幅上涨，实际客运量依然居高不下。餐饮业表现不佳，一方面是承受着严重的食品成本挤压（例如 Shake Shack 面临的高昂食材成本），另一方面则是市场长期低估了 GLP-1 类减肥药对实体经济产生的实质性冲击。消费者开始减少在杂货店和特定餐饮体验上的开支，将预算转向度假村、酒店住宿和娱乐服务等休闲领域。 - 航空与酒店股（如达美航空、凯悦酒店）频创历史新高，显示服务消费极具韧性。 - 食品成本上升使餐饮连锁店（如 Shake Shack）遭遇利润挤压，同时 GLP-1 药物正在分流传统食品消费。 --- ### [45:19](https://www.youtube.com/watch?v=swPQA7uJRgI&t=2719s) 企业利润与劳动报酬脱节：劳动力份额跌至历史新低 Michael 援引《华尔街日报》记者 Greg Ip 的报道指出，美国企业利润与劳动报酬之间出现了创纪录的分化。前一周嘉宾 Denise Chisholm 曾提及单位劳动力成本下降对股市是利好，但这对整个社会而言却非常糟糕。GDP 报告揭示出公众对经济感到痛苦的根源不在于 GDP 增长本身，而在于增长成果的分配：第一季度劳动者报酬（薪资与福利）仅增长 0.8%，而国内企业利润却暴增 2.7%。 这一差距导致劳动力在国民总收入（GDI）中的份额下滑至 51%，为 1947 年有记录以来的最低水平；与此相对，企业利润份额则攀升至 12.1%，创下 1950 年以来的最高纪录。企业利润创历史新高，而工人的收入份额却被持续压缩，这种严重的分配失衡正在加剧社会矛盾并深刻影响政治格局与选民倾向。Skanda 解释道，造成这一现象的原因有两个：一是劳动力市场本身的增速放缓，相比 2010 年代更为疲软；二是税收制度等结构性倾斜，引导人们避开 W-2 薪资收入转向股权或资本收入。此外，拥有极高利润率的“超级明星企业”（Superstar firms）在经济中的比重提升，也天然推高了利润在总收入中的占比。 ```mermaid flowchart TD accTitle: 劳动力与企业利润份额分化机制 accDescr: 展示劳动力市场疲软与税收制度如何导致国民收入分配出现极端分化 subgraph MacroFactors["驱动因素"] A["税收体系倾斜<br/>(偏向资本/自雇，抑制W-2薪资)"] B["超级明星企业主导<br/>(高利润率企业占比扩大)"] C["劳动力市场疲软<br/>(招聘放缓，薪资议价力下降)"] end subgraph DivergenceOutcome["分配结果 (第一季度数据)"] D["国内企业利润激增 2.7%<br/>(占GDI比例升至 12.1%，创1950年以来最高)"] E["劳动者报酬仅增 0.8%<br/>(占GDI比例跌至 51%，创1947年以来新低)"] end MacroFactors --> DivergenceOutcome ``` - 第一季度国内企业利润增长 2.7%，劳动者薪资与福利仅增长 0.8%。 - 劳动力在 GDI 中的份额跌至 51%（1947 年以来最低点），利润份额升至 12.1%（1950 年以来最高点）。 - 就业结构呈现分化：AI 尚未压制白领就业，白领雇佣反而有所回暖；蓝领就业（除数据中心建设、波音等少数领域外）表现普遍平淡。 --- ### [48:26](https://www.youtube.com/watch?v=swPQA7uJRgI&t=2906s) 名义 GDP 与薪资增速背离：美联储政策前景的鹰鸽之争 Neil 指出，当前名义 GDP 增速约为 6%，而名义劳动报酬（工资与薪水）增速则低于 4%，这种脱节在历史上极为罕见。两者的未来走向将直接决定美联储的货币政策路径。如果市场认为当前的高支出会最终推动劳动力收入大幅反弹、引发就业市场收紧和薪资走高（即劳动力数据向 GDP 靠拢），美联储将变得极其鹰派。在这种情景下，美联储不会只加息一次，而是可能加息 75 到 100 个基点，完全收回此前的降息。 相反，如果名义薪资放缓逐步拖累实际消费（将消费支出增速从 2% 压制到 1% 至 1.5%），加上住宅投资和建筑投资持续低迷，名义 GDP 最终向薪资水平靠拢，美联储便有空间按兵不动、等待局势平稳。Skanda 目前的立场偏向鹰派，甚至预期未来一年内有加息可能；而 Neil 则更偏向鸽派，他在分析中更看重劳动力和房地产指标。Neil 举例指出，过去一年服装价格上涨迅猛，但服装的实际销售量却在萎缩——如果劳动者收入无法跟上，企业将极难维持这种提价。 --- ### [51:10](https://www.youtube.com/watch?v=swPQA7uJRgI&t=3070s) 论美联储人事：Kevin Warsh 能否改变央行体系？ 当话题转向备受热议的 Kevin Warsh 时，现场气氛充满调侃与讽刺。Neil 直言华尔街上绝大多数人都不敢公开表达对他的真实看法，而自己身在小型机构，没有外交辞令的负担，愿意出来做这个“脏活”。 面对外界对其担任要职的传闻与所谓“政权更迭”（regime change）的期待，Neil 认为美联储同化他的概率，远大于他成功改造美联储的概率。Skanda 将这种情况比作“中国指套”（Chinese finger trap）——外来者越是用力推销鸽派诉求或试图打破规则，越会激起美联储同事和专业研究团队的强烈反弹。Neil 赞同这一观点并指出，如果一个人在美联储内部抛出漏洞百出的糟糕论点，最终只会被专业的参谋团队和同僚们无情驳回。 ### [52:21](https://www.youtube.com/watch?v=swPQA7uJRgI&t=3141s) 凯文·沃什的转向与“黄金时代”叙事 嘉宾们针对潜在美联储主席候选人凯文·沃什（Kevin Warsh）的公开言论展开了激烈的剖析。Neil Dutta 直言不讳地指出，沃什近期试图把自己包装成一个紧盯“截尾均值（Trimmed Mean）通胀率”的数据控，这与他过往的履历格格不入。回顾沃什担任美联储理事的岁月，他几乎从不发表探讨经济前景或具体政策走势的演讲——不像理事沃勒（Christopher Waller）或达拉斯联储主席洛里·洛根（Lorie Logan）那样专注于政策框架——相反，他的演讲总是充斥着高深空洞的哲学玄谈，甚至搬出伊曼努尔·康德（Immanuel Kant）等大学一年级文学课上才会听到的名字，在央行官员中显得极其古怪。 对于主持人 Michael Batnick“人难道就不能改变吗？”的提问，Neil 回应称，这种转变发生在沃什积极谋求职务的阶段，动机令人起疑。更关键的是，这套策略在 FOMC 内部很难奏效。沃什之所以大谈偏低的截尾均值通胀率，本质上是在推销他的“黄金时代”（Golden Age）宏观叙事——声称美国正处于生产率大爆发的开端，经济中存在巨大的闲置产能，因此美联储有更充裕的降息空间，试图效仿艾伦·格林斯潘（Alan Greenspan）在 20 世纪 90 年代末的做法。然而 Neil 吐槽道，人们总爱提学格林斯潘，却完全无视了格林斯潘在 90 年代后期同样经历过激进加息。 ```mermaid flowchart TD accTitle: 凯文·沃什的政策逻辑与 FOMC 实际阻力 accDescr: 展示凯文·沃什试图推行的生产率红利降息逻辑，以及他在美联储内部面临的机制与鹰派阻力。 subgraph WarshNarrative ["沃什的政策论点"] A["推崇截尾均值通胀率\nTrimmed Mean"] --> B["预判生产率激增与闲置产能\n黄金时代叙事"] B --> C["美联储具备更宽裕的降息空间"] end subgraph FOMCReality ["美联储现实机制与阻力"] D["委员会一人一票机制\n投票权分散"] E["鹰派异议成员增多\n主张转向紧缩倾向"] F["伯南克等元老公开批评\n缩表主张被斥为无意义"] end C -.->|"试图游说"| D D --- E D --- F ``` ### [54:02](https://www.youtube.com/watch?v=swPQA7uJRgI&t=3242s) 美联储内部博弈与应对供给冲击的歧途 讨论转向沃什若上任究竟能有多大影响力。Skanda Amarnath 分析认为，美联储主席并非一言堂，无法直接拍板决定加息或降息。尽管主席拥有制定会议议程和利用讲坛发声（Bully Pulpit）的议程设置权，精明的操作者能借此施加重大影响，但最终决策依然依赖一人一票的委员会投票机制。在上一次美联储利率决议中，就已经出现了四张异议票（一张偏鸽派，三张偏鹰派）；随着通胀降温停滞，鹰派成员很可能会进一步施压，要求摒弃当前的宽松倾向，甚至重新引入加息倾向。 关于美联储“应该做什么”与“实际会做什么”，Skanda 表达了悲观态度。他认为理想状态下，美联储应当保持定力，认识到当前物价压力很大程度上源自复杂的“供给冲击”。供应链冲击在产业链上下游传导极为缓慢，企业调整利润率的节奏也各不相同，其实际规模和持续时间往往被模型严重低估——正如美联储在 2021 至 2022 年间所犯的误判一样。隔夜借贷利率根本解决不了供给侧的瓶颈问题，但由于通胀数据持续高烧且波及面广，美联储官员极易陷入焦虑，担心长期通胀预期脱锚，进而被迫再度加息。 - **FOMC 异议票格局**：上一次议息会议已出现 4 张异议票（1 张鸽派，3 张鹰派），鹰派势力正在积聚反扑力量。 - **供给冲击的传导特性**：供应链中断和成本传导具有长期性与滞后性，货币政策抬高隔夜利率对此类问题完全不对症。 - **美联储决策偏差**：当预测模型屡次低估通胀水平时，央行倾向于误将其视为需求过热，从而引发非理性的防御性加息。 ### [56:30](https://www.youtube.com/watch?v=swPQA7uJRgI&t=3390s) 市场加息预期重燃与不合逻辑的紧缩风险 节目展示了利率市场走势图表（Chart 13），反映至 2027 年 9 月的联邦基金隐含利率曲线。Skanda 指出，市场目前的走势是一种折中定价：它在“按兵不动”与“重启连环加息”之间摇摆。如果美联储真的破戒启动加息，绝不会只是浅尝辄止的一次行动，极大概率会是 75 到 100 个基点的连续收紧（例如连续加息三次，每次 25 个基点）。当前金融环境整体仍然宽松，基准利率仅处于中性或温和限制性区间，且劳动力市场的左尾衰退风险已明显减退。 Neil 从宏观经济学家的角度对这种前景发出了严厉警告：如果美联储仅仅因为供给端扰动而加息，对资本市场和实体经济将是极其糟糕的打击。传统上股市能够消化加息，是因为加息往往意味着终端需求旺盛、企业盈利增长；然而眼下的美国经济表现极不均衡，本周刚公布的美联储《褐皮书》（Beige Book）便清晰描绘了一幅各领域分化严重、增长缺乏普惠性的经济图景。Neil 讽刺道，央行难道要因为原油价格、活牛价格甚至厄尔尼诺现象来加息吗？美联储既管不了霍尔木兹海峡的油轮通行，也左右不了天气。Neil 还特别提到，前主席本·伯南克（Ben Bernanke）最近公开评价沃什关于资产负债表的言论是“毫无意义的陈述”（meaningless statement），直接击碎了沃什长期执迷的缩表议题。 ### [59:17](https://www.youtube.com/watch?v=swPQA7uJRgI&t=3557s) 劳动力市场谜题：760 万空缺岗位与 AI 就业恐慌 节目的最后部分转向了听众来信。一名听众询问：长期以来牛市的质疑者总是抱怨知识密集型行业缺乏高薪工作，但劳工统计局（BLS）最新的职位空缺及劳动力流动调查（JOLTS）显示，全美职位空缺数升至 760 万个的两年新高，其中专业与商业服务业更是单月暴增 66.8 万个岗位；叠加过去两个月超预期的非农就业数据（NFP），这是否意味着市场情绪即将逆转，人们终于意识到 AI 并不会吞噬所有的就业岗位，反而在为劳动者创造新的机遇？ Neil 对此明确反驳了所谓的“AI 毁灭就业论”。他指出这种担忧本质上陷入了经典的“劳动力总额谬误”（Lump of Labor Fallacy）。在全美失业率仍处于 4.3% 的历史低位、黄金年龄段（Prime-Age）就业率保持稳健的背景下，大肆鼓吹 AI 摧毁就业完全缺乏现实依据。不过他也提醒投资者，不要被单一维度的宏观指标过度绑架，需要结合更广泛的劳动力市场结构来客观看待这波数据变动。 ### [01:00:35](https://www.youtube.com/watch?v=swPQA7uJRgI&t=3635s) 职位空缺数据的噪音与劳动力市场的真实韧性 Michael Batnick 与嘉宾们就近期劳动力市场的数据质量展开讨论。Michael 提出，他不相信专业和商业服务（Professional and Business Services）门类的职位空缺单月能暴增 70 万个，直言其中充斥着“虚假招聘信息（fake postings）”。Neil Dutta 开玩笑地借用特朗普的名言调侃称：“数据是真的，但新闻是假的（The data are real. The news is fake）。”他指出，从不同维度来看，例如 Indeed 的职位招聘数据在过去数周内整体是走弱的，因此 JOLTS 报告中的职位空缺指标（Openings）根本不能提供可靠的宏观信号，大家甚至认为“职位空缺是最假的数据”，真正具备参考价值的是聘用（Hires）和离职（Quits）数据。 随后话题转向更具实感的初请失业金人数（Initial Jobless Claims）。Michael 回忆起去年夏天当申请人数抬头时，Josh Brown 曾警告劳动力市场的失业往往具有“运动中的物体倾向于保持运动（Object in motion stays in motion）”的惯性特征，但事实证明初请失业金在冲高后见顶回落并重新正常化。Skanda Amarnath 指出，就业增长在经历大幅放缓后自发企稳确实令人意外，大约在 12 月的非农报告中就已展现出这种稳定性，证明就业增长并未掉落悬崖。特别是在专业与商业服务中的白领领域（除去临时派遣工等低薪岗位），招聘动能正在回升。过去三年白领就业一直疲软——先是 2022 年的技术行业衰退（tech-cession），紧接着美联储激进加息引发衰退担忧导致企业克制招人、消化疫情期间的过度招聘，再到后来对 AI 技术风险的观望——但实际上低企的资金成本信号早已表明企业完全有空间扩张。 ```mermaid flowchart TD accTitle: 劳动力市场信号传导机制 accDescr: 对比职位空缺假象与失业金及白领招聘的真实反馈链条 subgraph NoisyData["嘈杂/虚假信号"] A["JOLTS 职位空缺 (Openings)<br>包含虚假挂职与噪音"] end subgraph ReliableData["真实动能信号"] B["初请失业金 (Jobless Claims)<br>高位见顶后回落正常化"] C["聘用 (Hires) 与 离职 (Quits)"] D["核心白领与专业服务招聘回暖"] end subgraph MacroResult["宏观现实"] E["劳动力市场未坠落悬崖<br>在放缓后实现内生性企稳"] end A -.->|低可靠度| E B --> E C --> E D --> E ``` - 专业与商业服务行业单月职位空缺激增 70 万被质疑含有大量“幽灵招聘”。 - Indeed 平台职位发布指数在近期净读数上呈现边际走弱。 - 白领劳动力市场在经历近三年的压制（2022 技术业衰退、加息防御、疫情过度招聘出清）后出现企稳迹象。 --- ### [01:02:54](https://www.youtube.com/watch?v=swPQA7uJRgI&t=3774s) 企业裁员的隐性成本与 AI 时代的知识保留 Michael 询问，如果类似摩根大通（JPMorgan Chase）这样的大型机构进行大规模裁员，是否会因引发员工甚至政治力量的报复而面临极大的社会声誉压力，从而让企业在裁员上犹豫不决。Skanda Amarnath 从企业运营与组织架构的角度进行反驳，认为企业不敢轻举妄动的主要原因在于“流失知识资产的巨大风险（loss of knowledge）”。一旦裁掉负责核心系统架构的关键人员，公司未来将失去协调新技术和部署解决方案的能力。 大家普遍认为，目前根本无法精确判断哪些工种最容易被 AI 替代。被视为高危的白领岗位，往往恰恰是最终负责落实、整合和驾驭 AI 工具的执行者。Neil Dutta 补充道，在整个宏观经济周期中，大规模裁员往往是“最后才会出现的滞后指标”。用他的原话来说：“等到大规模裁员在数据上发出实质信号时，游戏早就已经结束了，兄弟（By the time layoffs are showing you something in terms of a signal, like it's over, buddy）。” --- ### [01:04:08](https://www.youtube.com/watch?v=swPQA7uJRgI&t=3848s) 嘉宾主页关注与推特大裁员的回顾 在对宏观指标进行总结前，主持人让两位嘉宾介绍各自的研究机构和关注渠道： - **Skanda Amarnath**：建议受众访问 `employamerica.org` 查阅公共研究报告，其团队持续跟踪拆解就业数据、CPI、PCE 及美联储政策动向；他在 Twitter/X 上除宏观分析外，最近主要在为纽约尼克斯队（New York Knicks）的季后赛呐喊助威。 - **Neil Dutta**：观众可通过屏幕上的二维码关注其研究，他本人更倾向于在 LinkedIn 上发布长篇分析，并在 Twitter/X 上分享图表，同时调侃自己也能在尼克斯队的官方 Instagram 账号里被找到。 在谈到社交媒体 X（原 Twitter）时，Neil 提及埃隆·马斯克对推特进行的“削肉见骨（cut to the bone）”式大裁员。他庆幸其他主流科技大厂当时没有盲目效仿这一极端举动。尽管当时外界恐慌大裁员会导致系统崩溃，但 Twitter 在精简至极后大体上仍维持了运转。 --- ### [01:05:08](https://www.youtube.com/watch?v=swPQA7uJRgI&t=3908s) 周期何时终结？非科技企业 AI 支出与信用利差的终极检验 关于本轮经济扩张周期何时逆转，Michael 提问：“每个人都想知道它何时结束。面对这个极其独特的周期，你们会盯紧哪一两项实质性数据？” Neil Dutta 首先指出，到今年年底至明年年初，宏观环境可能不再像当前这般顺风顺水： 1. **财政推力退坡**：被称为“伟大美丽法案”的财政刺激尾红利到年底将逐步消退，从顺风转变为逆风。 2. **劳动力市场的隐性紧缩**：市场常因通胀下降而讨论美联储的“被动宽松”，但如果名义薪酬增长持续疲软，而在美联储维持高利率不变的背景下，对劳动力市场而言这实质上构成了“被动紧缩（passive tightening）”。 ```mermaid flowchart LR accTitle: 经济周期拐点观测指标 accDescr: 描绘终结本轮周期的潜在催化机制及观测变量 subgraph PolicyAndFiscal["政策与财政边际转折"] P1["财政刺激红利耗尽<br>顺风转为逆风"] P2["名义薪酬放缓 + 利率维持高位<br>劳动力市场遭遇被动紧缩"] end subgraph BalanceSheet["资产负债表与支出意愿"] B1["非科技企业 AI 资本开支<br>Tokens 成本 vs. 投资回报率 (ROI)"] B2["企业信用利差 (Credit Spreads)<br>违约与融资风险爆发"] end subgraph Outcome["周期拐点"] EndCycle["资本开支缩减 / 类似 2000 年 Y2K 透支后回落<br>经济动能逆转"] end PolicyAndFiscal --> Outcome BalanceSheet --> Outcome ``` Skanda 强调“商业周期本质上是风险周期（Business cycles are risk cycles）”，核心在于企业的支出意愿。Michael 提出他的核心观察指标是**信用利差（Credit Spreads）**，并回忆在解放日（Liberation Day）期间，正是因为信贷市场没有恐慌，才支撑他始终坚定站在看多经济的一边；尽管信用利差恶化时往往为时已晚，但它是一个绝佳的风险试金石。 最后，Skanda 明确指出最需要警惕的资产负债表约束来自**非科技企业在 AI 上的支出意愿**。超大规模云厂商（Hyperscalers）有护城河诉求，无论如何都必须投资以维持竞争力；但对于普通实体企业，AI 是真金白银的日常费用（如 Token 采购成本），企业最终必须看到清晰的投资回报率（ROI）。这非常类似 2000 年互联网泡沫前夕对 Y2K 问题的提前开支透支——一旦非科技企业发现 ROI 不及预期而感到疲惫，决定停止追加开支维持现有水平，整个繁荣链条就会骤然紧绷。好在英伟达（NVIDIA）目前已经在财报中拆分了来自超大规模云厂商与其他企业的营收占比，为市场提供了监测 AI 真实渗透深度的关键切口。对话在对尼克斯队昨夜力克对手、夺得系列赛晋级优势的轻松调侃中进入尾声。 ### [01:08:22](https://www.youtube.com/watch?v=swPQA7uJRgI&t=4102s) 闲聊 NBA 季后赛与纽约尼克斯队的宿命回忆 节目临近尾声，话题转向了激烈的 NBA 季后赛讨论。嘉宾们复盘了眼前的对决，认为对手是一支韧性极强的队伍，绝非像骑士队（Cavs）那样容易对付。有人特别提到对方球员骨子里有股狠劲、拼劲十足（"Castle's a dog"），绝不会轻易缴械投降。不过，讨论者依然对纽约尼克斯队（New York Knicks）抱有信心，认为尼克斯整体上是一支更为成熟的球队，完全有能力稳住局面。大家也分析了上一场比赛的心理与体能动态，直言对手当时对胜利的渴求度显然高过己方，而且经历过前一轮激烈系列赛的消耗后，球员在最后几分钟明显体力不支、气力耗尽。看着尼克斯如今的表现，大家甚至有种恍如隔世的感觉，感叹球迷已经苦苦等待了太久，至今仍然不敢相信眼前发生的一切。 随后，对话勾起了在场各位童年时期作为尼克斯球迷的集体记忆。有人提到了 P.J. Brown 以及 1994 年那场刻骨铭心的总决赛。嘉宾回忆自己当年才 11 岁，至今仍清楚记得那轮系列赛——尼克斯曾以 3-2 领先，本该在麦迪逊广场花园（Madison Square Garden）锁定胜局，却最终痛失总冠军。另一个人则笑着表示当时自己只有 9 岁，记忆已经有些模糊。面对熬过了詹姆斯·多兰（James Dolan）长达二十年混乱治理的尼克斯队史，有人坦言自己早已变得愤世嫉俗，难免怀疑眼前的美好局面到底是怎么拼凑出来的；而同伴则打趣并打断道，现在可没有时间再去悲观怀疑（"no time for cynicism"），球队已经彻底翻过了那一页。 - 讨论中提到了对手核心球员的强硬拼搏特质，直言他们绝不会轻易放弃。 - 嘉宾们回顾了 1994 年尼克斯在大比分 3-2 领先局面下错失冠军的童年惨痛记忆。 - 经历了詹姆斯·多兰（James Dolan）执掌球队长达二十年的低谷与折磨后，老球迷如今在狂喜中依然夹杂着难以置信的复杂心境。 ### [01:09:23](https://www.youtube.com/watch?v=swPQA7uJRgI&t=4163s) 节目收尾与致谢 在轻松幽默的球迷怀旧气氛中，主持人向收听和收看本期节目的所有观众表达了由衷的感谢，盛赞听众们的支持让节目充满乐趣。主持人特别向嘉宾 Skanda Amarnath 发出再访邀请，表示这次探讨宏观经济与市场的对话非常精彩，期待他下次再次做客。最后，节目在主持人的道别声与轻快的片尾音乐中圆满结束。 ## Notable quotes > [16:37](https://www.youtube.com/watch?v=swPQA7uJRgI&t=997s) 我们现在正处于科技周期与商业周期深度融合的阶段。从经济衰退动态，到美联储关于潜在产出增长率与中性利率（r*）的激烈辩论，AI 必将在未来几年的宏观经济讨论中烙下深刻印记。到 2025 年底，AI 繁荣占美国经济的比重，完全有可能比肩 2005 年住宅投资在房地产泡沫巅峰时占 GDP 近 7% 的历史水平。 — Skanda Amarnath > [21:41](https://www.youtube.com/watch?v=swPQA7uJRgI&t=1301s) 商业周期本质上就是风险周期（Business cycles are risk cycles）。 — Skanda Amarnath > [26:21](https://www.youtube.com/watch?v=swPQA7uJRgI&t=1581s) 你从投资账户里提走一笔钱，拿去装修房子或者买个大玩具，结果过了 10 天，这个水桶又被填满了——因为半导体股票一直在源源不断地给你发大富翁式的游戏币。 — Michael Batnick > [31:25](https://www.youtube.com/watch?v=swPQA7uJRgI&t=1885s) 当增长全归于利润率而居民实际报酬为负时，我们真的能称这是一场生产力繁荣吗？ — Neil Dutta > [40:52](https://www.youtube.com/watch?v=swPQA7uJRgI&t=2452s) 资金虽在不断转手，但我们并没有因此变得更富有。 — Skanda Amarnath > [01:00:35](https://www.youtube.com/watch?v=swPQA7uJRgI&t=3635s) 数据是真的，但新闻是假的（The data are real. The news is fake）。 — Neil Dutta > [01:02:54](https://www.youtube.com/watch?v=swPQA7uJRgI&t=3774s) 等到大规模裁员在数据上发出实质信号时，游戏早就已经结束了，兄弟（By the time layoffs are showing you something in terms of a signal, like it's over, buddy）。 — Neil Dutta > [01:08:22](https://www.youtube.com/watch?v=swPQA7uJRgI&t=4102s) Castle's a dog ## Mentioned Michael Batnick, Neil Dutta, Skanda Amarnath, Josh Brown, Ben Carlson, Joe Weisenthal, Tracy Alloway, Edward Leamer, Greg Ip, Denise Chisholm, Kevin Warsh, Christopher Waller, Lorie Logan, Alan Greenspan, Ben Bernanke, Elon Musk, James Dolan, Jalen Brunson, Clyde Frazier, John Starks, Henrik Lundqvist, Matt Middleton, Chris Cherry, Sam Altman, P.J. Brown, The Compound, Animal Spirits, Odd Lots, Employ America, Renaissance Macro Research, Federal Reserve, Federal Reserve Bank of New York, Nuveen, New York Knicks, San Antonio Spurs, Cleveland Cavaliers, Madison Square Garden, Delta Air Lines, SK Hynix, NVIDIA, Amazon, Meta, Oracle, Anthropic, Microsoft Copilot, Ramp, Claude, OpenAI, Vertiv, Eaton, Caterpillar, Cummins, Hubbell, Comfort Systems, GE Vernova, Generac, Ford, Clorox, Hyatt, Domino's Pizza, Shake Shack, Cava, Dollar Tree, Dollar General, Chipotle, JPMorgan Chase, Indeed, Twitter / X, VanEck Semiconductor ETF (SMH), S&P 500 ## Video description On episode 245 of The Compound and Friends, ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Michael Batnick⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ is joined by ⁠Neil Dutta⁠ and ⁠Skanda Amarnath⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ to discuss: the AI investment boom, the state of the labor market, inflation risks, consumer spending, Fed policy, market breadth, whether this cycle is as unusual as it feels, and much more! This episode is sponsored by Nuveen. Learn more about Nuveen’s comprehensive private markets platform at⁠ https://www.nuveen.com/en-us/insights/alternatives ►00:00 - Cold Open ►08:45 - Intro ►10:04 - How is the Economy? ►16:37 - AI’s Impact on The Economy ►28:06 - What would cause a slowdown? ►45:17 - The Divide between Corporate Profits and Worker Pay ►51:08 - Kevin Warsh and The Fed ►59:14 - The Labor Market 👕 Check out The Compound shop: https://www.idontshop.com 🎙️Listen to our podcasts: TCAF: https://podcasts.thecompoundnews.com/show/TCAF/ Animal Spirits: https://podcasts.thecompoundnews.com/show/animalspirits/ Ask The Compound: https://podcasts.thecompoundnews.com/show/ATC/ The Compound Newsletter: https://www.thecompoundnews.com/subscribe Linkedin: https://www.linkedin.com/company/the-compound-media/ Instagram: https://instagram.com/thecompoundnews Twitter: https://twitter.com/thecompoundnews Tik Tok: https://www.tiktok.com/@thecompoundnews Investing involves the risk of loss. This podcast is for informational purposes only and should not be or regarded as personalized investment advice or relied upon for investment decisions. Josh Brown and Michael Batnick are employees of Ritholtz Wealth Management and may maintain positions in the securities discussed in this video. All opinions expressed by them are solely their own opinion and do not reflect the opinion of Ritholtz Wealth Management. See our disclosures here: https://ritholtzwealth.com/podcast-youtube-disclosures/ The Compound Media, Incorporated, an affiliate of Ritholtz Wealth Management, receives payment from various entities for advertisements in affiliated podcasts, blogs and emails. Inclusion of such advertisements does not constitute or imply endorsement, sponsorship or recommendation thereof, or any affiliation therewith, by the Content Creator or by Ritholtz Wealth Management or any of its employees. For additional advertisement disclaimers see here https://ritholtzwealth.com/advertising-disclaimers. #thecompound #investing #stockmarket #stocks [Source](https://www.youtube.com/watch?v=swPQA7uJRgI)

[16:37] Skanda Amarnath: 我们现在正处于科技周期与商业周期深度融合的阶段。从经济衰退动态，到美联储关于潜在产出增长率与中性利率（r*）的激烈辩论，AI 必将在未来几年的宏观经济讨论中烙下深刻印记。到 2025 年底，AI 繁荣占美国经济的比重，完全有可能比肩 2005 年住宅投资在房地产泡沫巅峰时占 GDP 近 7% 的历史水平。 [21:41] Skanda Amarnath: 商业周期本质上就是风险周期（Business cycles are risk cycles）。 [26:21] Michael Batnick: 你从投资账户里提走一笔钱，拿去装修房子或者买个大玩具，结果过了 10 天，这个水桶又被填满了——因为半导体股票一直在源源不断地给你发大富翁式的游戏币。 [31:25] Neil Dutta: 当增长全归于利润率而居民实际报酬为负时，我们真的能称这是一场生产力繁荣吗？ [40:52] Skanda Amarnath: 资金虽在不断转手，但我们并没有因此变得更富有。 [60:35] Neil Dutta: 数据是真的，但新闻是假的（The data are real. The news is fake）。 [62:54] Neil Dutta: 等到大规模裁员在数据上发出实质信号时，游戏早就已经结束了，兄弟（By the time layoffs are showing you something in terms of a signal, like it's over, buddy）。 [68:22] Castle's a dog

[19:38] Skanda Amarnath: When we talk about recession, we're not really talking about GDP. We're talking about whether the stock market goes down a bunch and whether you lose your job. [45:19] Greg Ip: To understand why people are so miserable about the economy, look no further than Thursday's report on gross domestic product—not how much GDP grew, but how it was divvied up. [57:43] Neil Dutta: Can the Fed control the flow of oil through the Strait of Hormuz? Absolutely not. Does it have any power over El Niño? Like no, it does not. [59:14] Neil Dutta: I sort of reject the premise that AI will actually eat all the jobs. That's kind of like the lump of labor fallacy... Openings are like the fakest thing; it's hires and quits that matter a lot more. [61:01] Skanda Amarnath: By the time layoffs are showing you something in terms of a signal... it's over, buddy. [64:08] Neil Dutta: They cut everybody at Twitter and it still works... I'm actually glad that more tech people didn't follow his lead. [65:09] Skanda Amarnath: For non-tech companies, AI expenditure is an expense and they need to see a payoff... Scenario we think about in 2000 where it was a lot of frontloading around Y2K and at some point they just got tired of it.

[0:00] How the hell did you get on that flight? [0:01] Like what was the deal? [0:04] >> So, um, uh, I called my broker and I [0:08] said, [0:08] >> "Is your broker's name Jason?" [0:10] >> No. No. [clears throat] [0:11] >> His name is John. Um, I'm not going to, [0:15] uh, tell the firm out of respect. Uh, [0:18] but anyway, um, [0:20] uh, I I go to him, I'm like, I want [0:22] tickets to game three, so make that [0:25] happen. I know you guys can do it, so [0:28] hook me up. [0:30] and and he goes, "There's no way we can [0:33] do that." And he's like, "Give me a [0:35] minute." So, he comes back with this [0:37] opportunity, which is like they call it [0:38] the Nicks fly away experience basically. [0:41] So, um my wife and I were able to go um [0:47] they put us in a chartered plane, like a [0:50] commercial like Delta plane. They flew [0:53] us to San Antonio. They fed us. [0:55] >> Was it all It was all Knicks fans. [0:57] >> All Nick Knicks fans. [0:58] >> Unbelievable. meet and greet. I met [1:00] Clyde, John Starks. [1:04] King Henrik was there. [1:05] >> Amazing. On the on your flight. [1:07] >> Yeah. [1:09] >> Um and you know, they all spoke. Um then [1:13] we went to the game [1:14] >> and the Knicks put you on their [1:15] Instagram. [1:16] >> That was that was a Yeah, that was a [1:19] moment for your wife and your wife. Look [1:20] at Look at this handsome devil. [1:21] [laughter] [1:22] Look at this cute couple. What a what a [1:24] freaking photo. [1:25] >> Um Yeah, that was great. No, I I I mean [1:27] Yeah. got your kids yet? [1:29] >> Oh, yeah. They saw I mean they were they [1:30] were we were all I was this is like [1:32] better than going on Bloomberg or CNBC, [1:33] you know. Um not as fun as going on this [1:35] podcast, though. [1:36] >> That's right. Um [1:37] >> but uh yeah. No, and then you know we [1:39] watched the uh the hope and and joy and [1:42] dreams just evaporate out of that [1:43] building. [1:44] >> How'd that feel? [1:45] >> Wonderful. Like it was just you know the [1:47] the best is when they know they've lost [1:50] and then they all just start leaving [1:51] early. That is the best feeling. I had [1:53] that feeling in Philly for game three [1:55] and [clears throat] [1:56] I don't take Jaylen for granted but [1:58] watching him every day. Um he he's [2:02] special obviously [2:04] seeing the frustration through the eyes [2:06] of the opposing fans. This one guy threw [2:09] his hat AND SAID HE DOESN'T MISS and it [2:13] just like melted my heart to see No, [2:16] it's great to see that. All right, so [2:17] here's my story. I have season tickets. [2:20] I share it with a friend of mine and uh [2:25] 109 row 20, that's where I sit. [2:29] He sent me a picture [2:32] and I said with a ticket sale. I said, [2:34] "What is this?" He goes, "That's a [2:37] $3,000 profit for game four like we [2:38] spoke about." Like, "Dude, I I don't [2:41] remember speaking about this. I I would [2:42] not have been cool with selling game [2:44] four to the finals." At least I don't [2:45] think I don't think I don't think I said [2:47] that. Um All right. All right. So, the [2:49] situation that we're in now, the [2:51] tickets, the get in tickets for the New [2:54] York City based games are higher than [2:57] Super Bowl prices. The last time I [2:59] checked for game four, it's like $7,000. [3:02] All right. So, here's here's my tickets. [3:04] Round one was 380, then it went to like [3:07] 550, 750, $1350 for the finals for for [3:11] each ticket. So, 1350. We sold them for [3:14] $4,500. [3:16] You can't get into the building for less [3:17] than $7,000. The Super Bowl get in [3:19] price, right, was like 35. So, it's [3:21] double. Now, prices will probably come [3:23] down a little bit, but the cheapest [3:24] ticket in my section was listed, I just [3:26] checked, for $17,000. Now, I don't know [3:28] that anybody's paying $17,000, but [3:30] that's literally, can I get $15,000 for [3:32] the tickets? Perhaps. So, where we are [3:34] now is we're using we're working with my [3:36] ticket broker [3:38] to basically [3:40] to basically like have games three and [3:43] four and break even. So, basically [3:46] whatever our tickets would have cost [3:47] 1350 a piece to have like worse seats. [3:51] So, I'm still going to games three and [3:52] four, but in worse seats than my [3:54] original seat. So, we did like all of [3:55] this. Now, it's not my friend's fault. [3:57] Like, we couldn't have possibly known. [3:58] But, here's the thing that really hurts. [4:00] Last night I was uh I was uh having [4:03] drinks with Matt Middleton and Chris [4:06] Cherry from Future Proof. They got they [4:08] went to the watch party with me. So even [4:10] the wa so the watch party was a [4:11] phenomenal experience. Unbelievable. [4:13] 18,000 people in the garden. [4:18] [cheering] [4:22] It was like basically like a home game, [4:24] but it was so freaking loud. You know [4:26] why? There was like no corporate seats. [4:28] It was only dieh hard fans. It was so [4:31] loud. [4:32] >> So they released the tickets and in two [4:34] minutes they're gone. Like I was number [4:36] 330,000 in line. So $10 tickets. I [4:38] bought it on Ticket Master or Tickpick [4:40] for like 60 bucks, whatever. Well worth [4:42] it to get in there. It was such a great [4:43] experience. So we're at dinner having [4:45] drinks and um Matt was like, "Oh, I [4:47] didn't know you had season tickets." Uh [4:49] I said, "Yeah, where do you sit?" "109." [4:52] He goes, "What row?" Like row 20. He [4:56] goes, "Dude, [4:58] my brother-in-law bought your tickets." [5:00] And I'm like, "You motherucker, GIVE IT [5:02] BACK. I WANT IT BACK. I'VE CANCELLED the [5:04] transaction." How crazy is that? Out of [5:06] all the [5:07] >> a small world. [5:08] >> So Matt Middleton bought bought my [5:10] tickets. [5:11] >> For well below what I could have sold [5:12] them for, [5:13] >> man. [laughter] [5:15] >> Well, at least you're still going. [5:17] >> At least I'm still going. So, what do [5:18] you guys Do you guys think there's any [5:19] like uh economic justification? Do you [5:22] think this is like this like the whole [5:24] story? It's like, yeah, the stock [5:25] market, people are rich and they're [5:26] buying tickets at any price. [5:29] >> I think that's that's basically the [5:30] story, right? We got we have the World [5:32] Cup tickets that are obviously selling [5:34] for I mean not as much as mea wants but [5:36] like we just have like so much uh [5:38] discretionary spending is actually a [5:40] pretty solid maybe it's for the upper [5:41] income part of the distribution but it's [5:43] still like uh it's a go-go time in terms [5:45] of consumer spending that may not be [5:47] backed up by all of the sort of job [5:50] market itself but it's good enough at [5:53] the higher end for NYX tickets right [5:55] >> I think that but the higher end or the [5:57] bigger shape of the K is like a very big [6:00] it's big because if how is the get in [6:03] price $7,000 I don't understand like [6:06] literally where is are people just [6:07] throwing out their credit cards and [6:09] saying like whatever like [6:09] >> well New York City isn't real life I [6:11] mean I it's um so [6:12] >> yeah it's$7,000 to get in [6:15] >> it does feel kind of ironic that you [6:16] have Texas versus New York obviously the [6:18] story of Texas is like it's the booming [6:19] state it's the state where everything's [6:20] growing but like New York City's got the [6:22] wealth that's that's still what's [6:24] overwhelming why you have New York [6:26] Knicks take fans taking over San Antonio [6:27] but you probably won't have Spurs fans [6:29] >> there's this research about like you [6:31] know like economists talk about the [6:32] wealth de accumulation puzzle right like [6:34] why why is it that like super wealthy [6:36] people they don't like actually spend [6:37] down their wealth over time even though [6:39] everyone expects them to like now that [6:40] you've finally gotten that opportunity [6:41] to spend some of it down right [6:43] especially because a lot of the wealth [6:44] is in places like New York so I just [6:46] think it's interesting I will say [6:48] something about the World Cup I am [6:49] taking my boys to the World Cup and [6:51] we're going down to Atlanta to do it [6:53] [snorts] [6:53] >> um [6:53] >> wait where is it I thought it was in [6:55] Philly [6:55] >> no there's games in Atlanta as well [6:58] >> there's games in New York too well then [6:59] New New Jersey life. [7:01] >> Obviously, I don't follow soccer. It [7:02] travels. [7:03] >> Toronto. [7:04] >> It's in a bunch of sites. [7:05] >> Oh, the World Cup's everywhere. [7:07] >> Mexico City. [7:08] >> It's like North America's hosting. [7:10] >> Yeah. Toronto, Boston, Foxboro, not [7:12] >> but even even tickets in San Antonio. [7:14] So, I just bought tickets for game five [7:16] and they're not cheap. Like the the get [7:18] in price right now is $1,700 for game [7:20] five. [7:21] >> For game two, it's it's sinking like [7:23] stone. It's like 700 now. 600. It was [7:26] like a thousand a couple days ago. [7:27] >> Unbelievable. All right. Uh, well, [7:30] excited to have you guys. You know, [7:32] Scanda, I asked Neil who should be in [7:34] the third seat because Josh is away this [7:36] week and you were the first person. I [7:38] said, "Fantastic." [7:39] >> Ben wanted to get you in here, so we're [7:41] excited. [7:42] >> Appreciate you having me. [7:43] >> It's the Compound and Friends episode [7:45] 245. [7:57] >> [music] [7:57] >> This message is brought to you by New [8:00] Private Markets are paving the way [8:02] forward as future [music] focused [8:03] investors look beyond traditional [8:04] markets to help their portfolios adapt [8:06] to an evolving environment. With [8:08] inflation, volatility, and shifting [8:10] economic conditions, reshaping [music] [8:11] the investment landscape, the case for [8:13] private markets has never been stronger. [8:15] New's comprehensive private markets [8:16] platform spans the full risk return [8:18] spectrum offering strategies [music] [8:19] across real estate, credit, [8:21] infrastructure, and natural capital. [8:22] Designed to work alongside traditional [8:24] investments and support diversified [8:25] portfolio construction. Backed by [8:27] decades of [music] experience and deep [8:29] asset class expertise, New brings the [8:31] scale and breath needed to navigate [8:33] today's complex markets. New unlocking [8:35] opportunity [music] in private markets. [8:37] Visit new.com to learn more. Investing [8:40] involves risk. Principal loss is [8:42] possible. Private market investments may [8:44] not [music] be suitable for all [8:45] investors. [8:48] [applause] [8:50] All right. All right. All right. I am [8:54] feeling elated, boys. I am feeling [8:57] really good. It has been a long time. We [8:59] won a finals game. The first Finals [9:01] game. Unbelievable. Um, and I'm so [9:04] excited to have you guys here today. So, [9:06] uh, fan of the show, friend of the show, [9:10] fan favorite of the show. That's what I [9:11] meant to I'm not I I hope you're a fan [9:13] of the show, but you're a fan favorite. [9:14] [music] [9:14] >> I'm definitely a fan. [9:15] >> All right, Neil Dodd, everybody. Neil is [9:17] the head of economics at Renaissance [9:19] Macro Research. Neil leads macroeconomic [9:22] research efforts with an emphasis on [9:23] analyzing the US [music] economy, [9:25] Federal Reserve, global trends, and [9:27] cross market investment themes. Neil is [9:29] considered a stock market economist. Do [9:30] you stock market economist? Market [9:32] >> business economist. Okay. Um Neil looks [9:34] at the economic data and tries to see [9:36] and highlight the risks to the consensus [9:38] as he sees them. and first time guest [9:41] Scanda Ammerath. Scanda is the [9:43] co-founder and executive director of [9:45] Employee America, leading the firm's [9:47] economic policy advocacy. Employee [9:50] America runs regular analyses of price [9:52] and jobs data, [music] [9:54] interprets and forecast market [9:55] conditions, and develops new frameworks [9:57] for Federal Reserve policy, strategy, [9:59] and communications. Previously, he [10:01] served as an analyst within the capital [10:03] markets function of the research group [10:05] [music] at the Federal Reserve Bank of [10:07] New York. Welcome. [10:08] >> Thanks for having me. [10:10] All right. Um, we are going to talk [10:13] about how AI is impacting the economy. [10:17] Um, but before we get into that and [10:20] other topics, let's just start here, [10:23] take a step back, zoom out as they say, [10:26] how do you guys think the economy is [10:27] doing? What's your read? Scanda, go [10:28] ahead. [10:29] >> Yeah. So, the distribution of risks has [10:32] shifted from where it was, call it, 6, [10:35] 9, 12 months ago. You had a lot of [10:37] people talking about, well, there's a [10:39] slowdown in the job market. Inflation's [10:41] coming down. Yes, there's an AI boom, [10:43] but everything else might be kind of [10:44] soggy. What have we learned since then? [10:47] Inflation is picking up. It's not just [10:49] about oil prices, although that's [10:51] obviously a big part of it in terms of [10:52] the closure of straight horses. And yet, [10:55] we've also seen the labor market is not [10:57] signaling red right now. It's signaling [10:59] something that's at least solid. Maybe [11:01] there's been some pickup actually in [11:02] some se sectors that have been soggy for [11:04] a long time. So the labor market's not [11:07] at the left tail and the inflation [11:08] stuff's at closer to the right tail. [11:10] That's a little bit different macro [11:12] picture. The one constant has been the [11:13] AI boom is continuing to boom. And that [11:16] kind of tells you look financial [11:18] conditions pretty supportive growth [11:21] better than expected. Inflation higher [11:23] than we want. That's that's a different [11:25] picture than the one we were telling [11:26] ourselves maybe 12 months ago. [11:28] >> I've never heard the economy described [11:30] as soggy. I like that. It's like a it's [11:31] an ugly word, but maybe the right word [11:34] to describe it. Yeah, you're you're [11:35] you're feeling a little soggy these [11:36] days. I I get the drift that you're not [11:38] as optimistic as you were. [11:39] >> Uh yeah, I'm not as optimistic. I mean, [11:42] I haven't been optimistic now for it [11:44] feels like what almost a year and a [11:45] half. Um so that's been, you know, a [11:47] shift for me. Um I would tend to agree [11:49] with most of what what Scandanda said. I [11:51] mean, you know, look, I mean, I will [11:53] point out though that right before all [11:56] this Iran stuff started, [11:59] 10ear yields were at [12:02] 3.93%. [12:04] Right? right? I mean, we're kind of [12:05] right there for um additional cuts for [12:08] this year. So, um you know, I think [12:11] since then, a couple things have [12:12] changed. Obviously, to Scandanda's [12:14] point, the distribution of risks around [12:15] inflation have changed. Um but the labor [12:18] markets aren't as the the labor markets [12:21] aren't as bad as they were at the end of [12:22] last year, right? I mean, they're not [12:23] nearly as strong as they were a few [12:25] years ago, but it's not nearly as bad as [12:27] it was last year. And that's that's [12:28] obviously important. Um the one thing I [12:31] would say that has continued um is [12:35] income growth continues to slide. So um [12:38] you know at least the the growth in uh [12:40] in total wages and salaries remains [12:42] pretty sluggish. Uh and in the context [12:45] of [snorts] um an overnight fed funds [12:48] rate where it is and price inflation [12:50] where it is. I mean I do think there are [12:51] stresses building for the household [12:53] sector. Um, and I do think there's [12:55] probably a limit uh to how much people [12:57] can spend by drawing down savings and [13:00] and so forth. So, uh, that to me is is [13:02] is my is my concern. Um, it's really [13:06] around the outlook for consumer [13:08] spending, which I don't think is is [13:09] particularly good. [13:10] >> Are you guys surprised that the economy [13:13] has stayed as strong, the soggginess [13:16] notwithstanding, [13:18] um, with the housing sector in an [13:21] absolute ice age? [13:23] Yeah, I'm very surprised that the [13:25] consumer has been as strong as it has [13:27] despite a housing market that's been [13:29] pretty frozen for a while. I mean, we're [13:31] seeing maybe some relief on inventories [13:33] finally. Inventory is starting to come [13:35] down and maybe you'll start to see some [13:36] home price appreciation, but the labor [13:39] market has been soggy for a while as you [13:41] said in terms of job growth, in terms of [13:42] wage growth. Those are generally slowing [13:44] throughout 2024 and 2025. Um, and some [13:47] of that's obviously supply side, [13:49] immigration, but it's still less money [13:51] earned through your paycheck, and that [13:53] tends to matter for consumer spending [13:55] over time. So, where are you funding [13:57] this consumer spending? If you look at [13:58] the personal saving rate right now, it's [14:00] kind of collapsed. Um, so it tells you [14:03] that on some level, consumer spending [14:05] growth is just outperforming what you're [14:07] earning through your paycheck. [14:08] >> But don't don't I was talking about this [14:11] with Ben on Animal Spirits. People pull [14:14] in their savings when they're feeling [14:16] optimistic. [14:17] >> That is true. [14:18] >> That's right. [14:18] >> Right. So, it's not necessarily like, oh [14:20] no, personal savings is collapsing. [14:21] What's going on? [14:22] >> Yeah. So, I mean, I think that there's [14:24] Right. Like, so to me, it's like let's [14:26] try to link like the economic data to [14:28] like markets, right? [14:29] >> Looking at the savings rate is an actual [14:31] like it's not a good timing mechanism, [14:33] right? It's it's horri I I remember back [14:35] in like 2005 the savings rate was [14:38] actually printing negative at the time [14:40] and there was like a chorus of like [14:42] bears telling you like oh my god there's [14:44] no more cushion like it's about to fall [14:45] over. We kept going growing for another [14:47] two years before I mean it didn't matter [14:48] in the end but you had to wait a while. [14:51] Um and I think the other thing of course [14:54] is that incomes tend to get revised up [14:57] over time. So the government always like [14:58] magically finds income uh and the [15:01] savings rate doesn't look as low as [15:02] people think. I mean it's just one of [15:03] the things that happens in the data. Um [15:05] but you know it's like anything in econ [15:07] economics right you have like two [15:09] different we have multiple ways of [15:10] looking at the same concept right [15:12] there's household employment and payroll [15:14] employment there's CPI and PCE and [15:17] that's also true for the savings rate [15:19] right like there's a the savings rate [15:21] that Scandanda is talking about which is [15:22] really just I think you would agree like [15:24] an income statement residual and then um [15:27] you know all they're doing is taking [15:28] income less spending over income. What [15:30] about like um something that is a m [15:32] maybe like a a better timing signal um [15:35] construction permits like Warren P has [15:38] showed a lot of that data on our show [15:40] that when that starts to roll it has [15:43] typically like a recession wasn't that [15:44] far behind [15:45] >> I mean I don't know like the Ed Lemur [15:47] thing is like was something that people [15:49] were talk have been talking about for [15:50] years right like Ed Leur was the guy [15:51] that wrote the paper like housing is the [15:53] business cycle right like normally I [15:54] would agree with him [15:55] >> was [15:56] >> yeah I mean that's the thing it's sort [15:57] of it's competing for seat resources [16:00] with the AI thing, right? Like so [16:02] they're competing with land with AI data [16:04] centers and [16:05] >> yeah, housing I mean the resale market [16:07] might be getting a little better. Like I [16:09] don't think there's much improvement in [16:11] uh in new home sales. I mean if you look [16:12] at new home sales so far this year, [16:14] they're down about 6 7% against the same [16:16] period last year. Like there's builder [16:18] margins are still under pressure. Like [16:19] there's there's not a whole lot going on [16:21] with respect to residential construction [16:23] and I don't think that's going to happen [16:24] this year. Um, but you just have [16:30] I mean I think we're probably going to [16:31] talk about this, but the AI boom is [16:33] quite spectacular in how much it's I [16:35] think helping lift uh growth. [16:37] >> All right, enough enough uh clearing of [16:39] the throat setting up the table. Scand [16:41] in preparation for this um I was looking [16:44] at I I think I googled scan I think I [16:46] googled your name LinkedIn um [laughter] [16:49] which is a weird thing to Google but I [16:50] don't know how else I got to this [16:51] >> there's only one [16:53] >> I don't know how else I got to this [16:54] result. There she is. Um, so the first [16:58] result, somebody tweeted, um, [17:01] expenditures on AI is about to surpass [17:03] the peak spending on housing at its 2005 [17:06] peak spend. Um, and then they quote you, [17:09] "At close to 7% of the economy at the [17:11] end of 2025, it's plausible the AI boom [17:13] would be on par with the share of the US [17:14] economy. Housing investment represented [17:16] at its 2005 peak." So I clicked on the [17:18] link and it said, "First heard on the [17:21] compound." [17:21] >> That's right. [17:22] >> So like, holy [ __ ] So, we were talking [17:23] about this piece that you wrote back in [17:25] January 2025, and you wrote at the time, [17:28] I think you wrote with Joe and Tracy. [17:29] Um, you either wrote or said, "We are [17:31] now at the stage where the tech cycle [17:34] and the business cycle are poised to [17:36] blur. From recession dynamics to the [17:38] Federal Reserve's debate about potential [17:40] growth and the neutral rate of interest, [17:41] known as our star, AI will leave its [17:44] mark on the next few years of macro [17:45] discussion." So, [17:47] >> there you go. [17:47] >> Voila. [laughter] [17:48] >> Boom. I I that that's one thing that's [17:52] aged quite well, I'd say. Um I think [17:54] we're clearly seeing we've we've ripped [17:56] past 7%. Right? So we we've ripped past [17:58] the the peak for housing investment in [18:01] the 2000s. So the housing bubble and the [18:03] residential investment that's happening [18:04] there as a share of GDP. I think you can [18:06] look at the relevant components now for [18:08] this AI boom. So I'm talking about tech [18:10] equipment, software, industrial [18:12] equipment that powers all the data [18:14] centers. That's all past uh the peak of [18:17] the the 2000s boom. It's past the peak [18:19] of the ve very well past the peak of the [18:22] 90s tech boom, the dotcom boom, the [18:24] telecom structures boom. So we're at a [18:26] point where this very investment [18:28] intensive part of the expansion that is [18:31] a to your point about also the saving [18:32] rate going down. You see the stuff at [18:34] the peak, it's also a way of saying if [18:37] this turns obviously it's going to leave [18:40] a mark also on the downside. I can't [18:42] tell you that that's going to happen [18:43] right now, right? I don't think that's [18:44] like anything that suggests that oh my [18:46] god all the spending is about to stop or [18:47] all the spending is about to slow down. [18:49] If anything, we're just seeing more and [18:50] more um financing, expenditure, issuance [18:54] all tied to this boom. And that just [18:57] tells you this is the go-go time and [18:58] everything that is we're all wrapping [19:01] every single part of markets, the [19:03] economy around this big technological [19:06] and investment boom. Neil, it sounds [19:08] like you have a bone to pick with [19:10] economists trying to properly account [19:13] for the impact of AI on the economy and [19:16] you're rubbing your face. So, [19:18] >> you know, I I um [19:21] >> Well, I mean, there's I mean, there are [19:22] there's a I think a decent contingent of [19:24] people on the street that think that [19:25] this is having like a fairly modest [19:27] effect on the economy, right? Like, so [19:28] it's um [19:29] >> why because they're looking at the wrong [19:30] things? They have no common sense. What [19:31] do you think? [19:32] >> Well, no. I mean, it's sort of like GDP [19:33] accounting, right? So one of the reasons [19:35] why I mean it's true that information [19:37] processing and software is what 6 7% of [19:40] GDP. Um it's also true that a lot of the [19:43] growth from AI leaks abroad in the form [19:45] of import. So that actually counts [19:47] against GDP. [19:49] >> Hold on. All right. So I was going to [19:51] say like what the hell is uh leakage? [19:53] What do you call it? [19:54] >> Import leakage. [19:55] >> Import. Okay. Is that a is that a it's [19:57] new to me. Is that a thing that's been [19:59] >> in the e uh economist lexicon or [20:04] Go ahead. You want to take it? [20:05] >> Yeah. So, it's it's basically so there's [20:06] all this spending happening [20:07] domestically, but that spending could be [20:09] for products that are produced outside [20:11] of the US. And so, you think about like [20:12] what SKHEX produces or what like all [20:15] sorts of Japanese and Korean [20:17] manufacturers are producing for all your [20:19] tech hardware or the energy systems [20:21] behind it. Um, that's all stuff that's [20:23] coming that that's value ad. That's GDP. [20:25] That's really not show like it's not US [20:27] GDP, but it is US spending. [20:29] >> So, why not just look at global GDP? [20:30] Surely it must show up there. Well, I [20:31] just think that there's a little bit. It [20:34] does. I mean, in in a in a sense, I [20:36] mean, just because the dollars don't [20:37] show up in GDP, don't don't doesn't mean [20:39] that it's not like it's vanished or [20:42] something. You know what I mean? Like, [20:43] so I think what economists miss about [20:45] this is that there are lots of uh kind [20:47] of um linkages to other areas of the [20:50] economy that are not neatly captured, I [20:52] would say, by GDP, like simple GDP [20:54] accounting, [20:55] >> like tickets, [20:56] >> like NYX tickets, like consumer [20:57] spending, like municipal government [21:00] finances, right? Like the California [21:01] Legislative Office talks about how over [21:03] half the growth in income tax [21:04] withholding is a function of all the [21:05] RSUs that are vesting, right? Like it's [21:08] like that's your meal ticket, right? If [21:10] you're a worker in one of those big [21:11] companies, um it's probably it's juicing [21:15] obviously global growth. Um but more [21:17] importantly, it's helping corporate [21:18] earnings and that matters because it [21:21] finances a wealth effect through [21:23] consumer spending. And again, that's not [21:25] something that is neatly captured by a [21:26] simple accounting identity. So, it's [21:28] almost like saying, "Yeah, I mean, [21:31] residential investment 6 or 7% of US [21:33] GDP, but not really because we import [21:36] all the drywall and lumber from Canada." [21:38] Like, it's a it's a bit I don't know. I [21:40] think it's like one of these arguments. [21:41] >> Neil, you're being too kind. I think I I [21:43] think actually like this this whole like [21:44] uh it's not really counting to GDP [21:46] because it's imports totally misses the [21:48] point of what a boom is all about. It's [21:49] about the spending. The spending is [21:51] sitting on some company's balance sheet. [21:53] It's affecting their risk. It's showing [21:54] their risk appetite currently and it [21:56] might affect their risk appetite in the [21:57] future, right? So like business cycles [21:59] are risk cycles. They're about the [22:00] willingness to spend to spend on labor, [22:02] spend on capital, and when the spending [22:04] stops, that's when you get the [22:04] recession. And when we talk about [22:05] recession, we're not really talking [22:07] about GDP. We're talking about whether [22:08] the stock market goes down a bunch and [22:10] whether you lose your job. Like that's [22:11] really what we're talking about. So all [22:12] of this like GDP accounting stuff can [22:14] really miss the point. You have a huge [22:16] volume of spending relative to the size [22:18] of the economy. It's going to sit on a [22:20] bunch of balance sheets. They might be [22:21] able to handle it. we're talking about [22:22] some of the best balance sheets [22:23] historically um in in terms of a lot of [22:25] the the big mega mega cap tech. But it's [22:28] still quite remarkable that like we've [22:30] had the scale of spending. It's growing [22:32] so fast and with that comes risks and [22:35] also obviously just growth in the [22:37] present. [22:38] >> So it's all about spending which is [22:39] driven by the labor market and of course [22:41] the equity market. Duh. But if people [22:42] have their job, they're going to spend [22:43] their money. Do you think that there's [22:45] anything that the average investor can [22:48] look at whether it's on the economic [22:49] side? Like I kind of feel like the [22:51] traditional playbook is like not really [22:53] useful here in terms of like what [22:54] leading indicators used to say. Is it [22:56] going to be like a concurrent slowdown [22:57] where it's like stocks get killed um [23:00] spending pulls back and it's just going [23:01] to happen when it's going to happen. [23:03] There will be very little warning. Feel [23:05] like people have been looking for [23:05] warning signs just for the last 15 [23:07] years. [23:08] >> It's true. I mean we you always have to [23:09] look behind your your your shoulder for [23:11] something. But I think if you look back [23:13] to the dot boom, right? Was there like a [23:15] macro signal that told you this was over [23:17] or was it that earnings started to miss [23:19] around April 2000 until you had a lot of [23:22] these sort of events through 200? [23:23] >> That's a lot of vibes too, right? [23:26] >> Yeah, there's a lot of vibes but it's [23:27] like I I don't think I could point to [23:28] aha it was housing starts turned in 1999 [23:30] or 2000 and that caused the cyclical [23:33] slowdown. It's like I don't I think it's [23:34] really about the stock market. It is [23:36] about it's about tech. [23:37] >> Some people will point to like I think [23:38] there was a Baron's article that came [23:39] out of it one weekend and the next week [23:41] the stock market killed and then it just [23:42] started to unravel. Like was it the [23:44] article that causes something? I mean it [23:46] was obviously going to burst either way [23:48] but [23:48] >> it was a bunch of spending that got [23:49] pulled forward in terms of IT systems [23:51] Y2K a lot of that stuff and then you [23:53] keep spending and keep spending and it's [23:54] like every exponential curve is always [23:56] underestimated in real time and at some [23:58] point there is [23:58] >> I mean the other thing of course is that [23:59] the Fed was hiking and and the and but [24:02] the stock market was still going up [24:04] concurrently as the Fed was hiking. So [24:05] it wasn't ne I mean this whole notion [24:07] that like oh the Fed should step in and [24:08] like [24:09] >> hike to blow this whole thing up. I [24:11] mean, I'm not, first of all, I'm not [24:12] sure that they really can. Um, but all [24:14] you'd be doing at that point is [24:17] exacerbating the stresses in the in the [24:19] areas of the economy that you were just [24:20] talking about, like housing, like some [24:21] of these credit sensitive areas. And the [24:23] key distinction between now and then is [24:26] that the labor markets were genuinely [24:27] overheating back then. Um, there's [24:29] really no evidence that I mean, we could [24:31] talk about, you know, things getting [24:32] more stable relative to where where they [24:34] were 6 months ago, but it's not like [24:36] you're seeing like broad-based wage [24:37] pressure. I mean, [24:38] >> are you guys worried about um credit [24:40] card delinquency? That's been a topic [24:41] that's that's come up recently. [24:43] >> Not yet. I say I would say there's [24:45] probably some signs that the household [24:46] sector will eventually face stresses if [24:48] like the job market isn't good. Um right [24:51] now we're seeing signs that the [24:52] opposite, but if we are like this is [24:54] still not yet at a point to me where the [24:56] household sector is pretty flush in [24:57] terms of liquidity. Um it may be [24:59] disproportionate in terms of [25:01] distribution, but [25:02] >> so a lot of a lot of the spending is [25:03] >> it's getting worse at the margin. I mean [25:05] I agree. I mean it is getting worse. I [25:06] mean even the Beebook talked about it [25:07] yesterday, right? They talked about [25:09] we're seeing uptick in mortgage [25:11] delinquencies and credit card [25:12] delinquencies and agricultural [25:14] delinquencies. I mean these are I mean I [25:16] guess you know the thing is like do you [25:17] think it's a body in motion that stays [25:19] in motion and if you do then [25:22] you know you better pray that the labor [25:24] markets start to accelerate because if [25:26] they don't and wages continue to slow [25:28] then those problems are going to just [25:30] get worse. And so, yeah, I mean, I agree [25:33] that it's it's very low, but if you're a [25:35] bank, I would probably want to provision [25:37] for more loan losses over the next year. [25:39] >> If you think about the like the '9s comp [25:41] or the late '90s comp, labor market's [25:44] definitely not as strong as it was then. [25:45] And so, there's clearly like a sense of [25:47] if this labor market does not show any [25:48] sort of real pickup over the next six [25:50] months, we will be talking about sort of [25:52] all the left tail stuff again. [25:54] >> But can that be good that it's not a [25:55] strong like because it's not causing [25:56] wage pressure? [25:57] >> It could be. The other side of it is [25:59] inflation is a lot more of a problem [26:00] this time than it is than it was in 1999 [26:02] or 2000. And so we have inflation that's [26:05] tied to tariffs, inflation that's tied [26:06] to the AI boom itself. There's inflation [26:08] that's tied to obviously the closure of [26:10] straightfor. So if you think about [26:11] airfares are a lot higher now and big [26:13] part of the jet fuel. If you think about [26:14] um AI boom and its impact on basically [26:17] if you want to go buy a laptop now you [26:18] can see the price right it's not it's [26:20] not what it used to be and there's all [26:21] sorts of computer hardware memory is [26:23] memory shortage having its impact [26:26] >> and so this is all a lot of sectoral [26:27] stuff. it stinks. Kind of like the 2000s [26:29] in a lot of ways. If you remember, like [26:30] a lot there's a lot of random inflation [26:31] that kind of creeped up around then. So [26:34] that's not great for the consumer, [26:35] right? Consumer's got to pay the pay the [26:37] bill on that. It means they're either [26:38] enjoying not as much of a standard of [26:40] living improvement or in some cases a [26:41] standard of living reduction. Um so [26:44] that's the downside here. So, we have [26:45] like not as good on the labor market, [26:47] harder, harsher on the inflation side um [26:50] this time around relative to what was [26:51] the '9s, which was kind of nirvana in [26:53] terms of labor market being pretty [26:55] strong, but inflation not really rearing [26:56] its head. [26:57] >> Josh and I talk a lot on the show about [26:58] the wealth effect and um what drives [27:01] people to spend more money. And I've [27:04] mostly rejected the idea that people [27:06] spend money based on how they how much [27:07] they think their house is worth because [27:09] you don't see it on the screen. Yeah. [27:10] No, it feels good. Sure, I suppose. But [27:13] your investment account, [27:16] it leads to overconfidence. I mean, it [27:19] just does. You can take money out of [27:20] your investment account, pay for a a [27:22] home renovation or a toy or whatever, [27:25] and in 10 days, the bucket is full again [27:27] >> because the semiconductors just keep [27:29] giving you free monopoly money. So, that [27:31] like I think there's no there's no doubt [27:32] about it that is driving a huge amount [27:35] of the spending. Today, [27:36] >> in 2022, the stock market did [27:40] go in reverse. It was a bare market and [27:42] it didn't last five years, but all of [27:45] the names that everybody loved got cut [27:47] in half for the most part. Like Amazon, [27:49] Facebook, uh, Nvidia fell two-thirds [27:52] like legitimately lost like a lot of a [27:54] lot of money and people didn't stop [27:56] spending at least as far as as far as I [27:57] know. [27:57] >> No, because we had income growth. [27:59] >> Okay. [27:59] >> Yeah. That's the key difference between [28:01] now and then. [28:02] >> So if so if we [28:03] >> and we had a lot of pandemic savings. [28:04] >> Yeah. I I know it's it's not apples to [28:06] apples at all, but so you're saying that [28:07] if what would what would it take for the [28:10] spending to slow down? Would it be the [28:13] stock market or the labor market or or [28:15] both or who knows? [28:16] >> Well, I think so there's a couple [28:17] things. I mean, I think right now we're [28:20] sort of for us, I think for economists, [28:22] it's like, is it linear or nonlinear? [28:24] >> Which part? [28:25] >> The con the slowdown in consumption. [28:27] Right now, you're basically in a linear [28:28] I think a linear slowdown in consumer [28:30] spending. Consumer spending is growing [28:32] about 2%. a little below, a little [28:35] above, it depends. But, you know, more [28:37] or less that's where we are. Um, [28:40] if you start to see layoffs [28:43] in any meaningful extent, not that we [28:45] have outside of technology, um, that'll [28:48] probably hurt consumer spending pretty [28:50] quickly. Um, [28:53] that's one option. The other way would [28:55] be you get a a market correction of some [28:57] kind. Um, if you just assume like [29:00] savings are stable, like consumption [29:02] will probably naturally slow a little [29:04] bit anyway because income growth is so [29:05] sluggish. [29:06] >> But doesn't matter. [29:06] >> That that doesn't mean it's going to [29:07] like fall off of a cliff. You know what [29:09] I mean? Like it's just okay, we're [29:10] instead of growing two, we'll grow like [29:13] one to one and a half. [29:14] >> In a world where these hyperscalers and [29:16] others joining the party are spending [29:18] $700 billion of capex, doesn't matter if [29:20] the consumer pulls back a little bit. [29:21] >> A little bit is not enough, right? [29:23] Consumer consumption is generally pretty [29:24] smooth anyways. So really when you think [29:25] about people say oh the US consumer is [29:27] the economy but investment is the [29:29] business cycle investment is the thing [29:31] that's volatile that moves with the [29:32] business cycle and right now it's risk [29:34] on right there's clearly a lot of [29:35] capital commitments that are tied to [29:37] this there's all sorts of planned [29:38] spending planned additions of energy of [29:41] data centers that haven't happened yet [29:43] and as long as there's the belief that [29:44] that this is going to keep continuing [29:46] which is the case right now that's going [29:48] to happen and that's what's going to [29:49] drive sort of where the market goes and [29:50] market sentiment and I think that's just [29:52] >> what's so unusual about this is that [29:54] it's like equipment investment and like [29:56] um non-residential business fixed [29:58] investment that's driving it and that's [30:00] what's interesting I mean if you're [30:02] talking about business cycle economics [30:04] right like typically the investment [30:05] piece that cracks is residential [30:08] investment right [30:09] >> right um so what's interesting about [30:11] this is that because typically if you [30:13] look at it I mean equipment like [30:15] non-residential business investment um [30:18] capex [30:19] that actually follows growth it doesn't [30:21] it's not a leading indicator [30:22] historically right I mean basically The [30:24] way it works is companies think growth [30:26] it's the accelerator effect is what we [30:27] call it, right? Like companies think [30:28] growth is going to pick up and so they [30:29] start investing more. Um what's unusual [30:32] about this cycle is that it's not like [30:34] growth expectations are really taking [30:36] off in any material extent. Um but you [30:39] have this sort of spectacular capital [30:41] spending boom nonetheless. So that to me [30:43] is like to your point like why are all [30:46] these like traditional things not [30:47] working? [30:50] >> We don't think of tech as like a [30:51] cyclical sector, right? we think of as [30:52] housing, manufacturing, maybe some like [30:55] segments of consumer spending, but like [30:57] tech is the cyclical thing. It is the [30:59] thing that matters for it's driving all [31:01] the vault in terms of GDP, in terms of [31:03] um why we're getting the outcomes we're [31:05] getting in stock market too, right? So [31:08] what we don't have like a leading [31:09] indicator for tech outside of like if [31:11] you're really locked in on some [31:12] particular names and maybe certain [31:13] orders for this tech supply chain, maybe [31:16] then you can have a read into the [31:18] leading indicators of this dynamic. But [31:21] um it's really just about like business [31:23] fix investment that's all tied to AI [31:24] boom. [31:25] >> So tech AI it's sucking everything in. [31:27] We're going to get to some of the work [31:28] that you've done on everything looking [31:29] like a semiconductor stock. The [31:31] framework that I'm that I'm working with [31:33] today and by the way things changed so [31:36] fast and people act like what's [31:38] happening today has been in place [31:39] forever. Like nobody wanted the mag [31:42] seven stocks in the fall. These things [31:43] were getting destroyed. Oracle got [31:45] Oracle fell by 60% because people are [31:47] like there's no way that Sam Alman is [31:48] gonna be able to pay that 5year 300 [31:50] billion dollar contract. There's no way [31:52] and all of a sudden it's hot again and [31:53] people are like up bubble. Um so the way [31:56] the way that I see the world today is um [32:00] that everything is being driven by the [32:01] investment boom by the by the buildout [32:04] and I don't know when the handoff [32:06] happens but how much of the world is [32:10] even using these tools the agentic um AI [32:15] stuff like what's what's interesting [32:17] about where the tenure was before the [32:19] war and today is that you got the [32:22] closure and the inflation picking up [32:25] concurrently with the crazy boom in [32:29] anthropics revenue going from like 9 to [32:31] 45 and it happened at the same time. And [32:36] the point is nobody's even using these [32:38] things and all we're hearing about is a [32:40] shortage of compute to so to suggest or [32:42] to think or to use the the framework [32:44] that like it's late or getting long in [32:46] the tooth. I kind of think it's just [32:48] starting. It's really hard to know in [32:51] real time and it does seem like the best [32:53] approximation is what we can see in the [32:54] present. What I can see in the present [32:55] is risk on right. I can see just that [32:57] there is a lot of appetite. The capacity [32:59] on compute is clearly very much tight as [33:01] far as as far as anthropics concerned [33:03] and at the same time they are seeing [33:05] revenue growth. So that's all reasons to [33:08] keep investing right reasons to keep [33:09] being optimistic. At some point there [33:12] are I I I'm sure there is a point where [33:14] exponential curves become s curves but [33:16] that point is not right now. I love that [33:18] you said that because all we could [33:19] observe is is [33:21] >> is that the line's going up today? Who [33:23] knows? [33:23] >> So, I guess I guess one thing I was [33:25] thinking about and I I really enjoyed [33:27] your uh your conversation with Denise [33:29] last week. Um you know, she's great. The [33:32] uh this whole notion of like it's early, [33:35] it's it's not a bubble because the [33:37] earnings are are are so strong. And but [33:40] I also would say at some level like the [33:42] earnings are tied to some kind of [33:44] temporary phenomenon with this at some [33:45] point the data center buildout will stop [33:48] and the earnings won't be there. So what [33:50] are we really talking I mean just [33:51] because it's not just because the PE [33:53] multiple isn't like ridiculously high [33:55] doesn't necessarily mean [33:57] >> I agree that tells you nothing. [33:58] >> Yeah. So I I think that there's a little [34:00] bit of that uh going on on the street. [34:02] Um the other thing I would say um is [34:05] this is like a really I mean people talk [34:07] about product I mean I go to client [34:08] meetings it's like oh productivity boom [34:10] like you know this is a very unusual [34:12] productivity boom like what is the point [34:14] of investment ultimately the point of [34:16] investment is to raise household living [34:18] standards can we say that that's what's [34:20] happening with AI like if anything it's [34:22] a really weird productivity boom when [34:23] the prices for information technology [34:25] commodities like Scandanda was talking [34:26] about software laptops it's actually [34:29] going up like you you pull up those [34:31] charts in the 90s or 2000s it was [34:34] deflating [34:36] month after month after month sometimes [34:38] at accelerating rates. Um you go back to [34:41] that period too again very strong [34:43] productivity growth we you know we also [34:44] had very strong growth in real income [34:47] right like there should be some [34:49] relationship between stronger [34:51] productivity [34:52] and stronger real compensation. [34:55] We just had negative real compensation [34:57] growth this year. So right now all the [34:59] growth is flowing to you know margins I [35:02] guess. Um and I don't I mean that to me [35:04] is like this is like this is why I say [35:05] like is it really a productivity boom [35:07] like it's not yet raising household [35:09] living standards and I don't know how [35:12] long that can continue. [35:13] >> I honestly think for productivity a lot [35:15] of what people are talking about are [35:16] things that happened in the previous [35:18] call it one to three years like you look [35:21] at Q1 data on productivity not great. If [35:23] we think about the supply shocks that [35:24] are hitting because of hey prices of [35:26] price of gasoline is much higher than it [35:28] was in Q1 price of uh all sorts of [35:30] energy airfares. You have a lot of other [35:32] shortages and other shocks that are [35:34] materializing too. [35:35] >> Cattle prices. [35:36] >> Yeah. Right. Cattle and beef all that [35:37] stuff is going to feed through as well. [35:39] If that's happening that's probably [35:41] going to also weigh on productivity too. [35:43] And so productivity growth is not [35:44] actually as rosy now as it probably was [35:47] in 23 24. Um, [35:49] >> but doesn't doesn't the the whose [35:52] productivity? [35:53] >> That's right. The macro data stats that [35:55] the Fed might look at that their [35:57] productivity. I do think it's like even [35:59] in terms of diffusion, you're right that [36:01] we actually haven't seen mass adoption [36:04] of AI in terms of in the real economy. [36:07] >> Nobody's using it for the I mean, I know [36:08] our listeners probably are like, you [36:10] know, you guys are, but in the real [36:11] world, [36:12] >> yeah, there's there's there's limits [36:13] even in terms of large scale large um [36:15] corporations. [36:16] >> Of course you do. If if you're a smaller [36:18] company or a smaller organization, you [36:20] probably can adopt these tools. You [36:21] don't have to worry as much about [36:22] security risks that stuff. If you're a [36:23] larger company, there's all sorts of [36:25] walls. You can only use co-pilot. You [36:26] can only use these types of tools. You [36:28] can't use those because they might um [36:30] ultimately present security risks. And [36:31] so actually adoption may not I think [36:33] it's [36:34] >> there's also a measurement there's also [36:35] a measurement thing with this. Like I [36:37] mean like everyone's like looking at the [36:38] ramp index and like how how like I mean [36:41] index [36:42] >> like what is that even telling me, [36:43] right? was like, "Oh, look, OpenAI is [36:45] going down and everyone's using Claude [36:46] now." Like, okay. I mean, I don't even [36:48] know what to do with that. So, I would [36:50] just say that if it's a genuine [36:52] productivity boom. I mean, people, it [36:54] feels like everyone assumes that like [36:57] everyone's margins will expand because [36:58] of this, [36:59] >> but they are [37:01] margins are at an all are going up, [37:04] >> right? Ultimately, costs need to come [37:06] down to households. [37:07] >> All right. So, we'll talk about in a [37:08] second. I just want to end with end this [37:09] the topic with this. So Neil, you said [37:11] um we ran a one-year daily return [37:14] correlation between every S&P 500 name [37:17] and the semiconductor ETF SMH. 15 non- [37:21] techch S&P 500 companies collectively [37:24] worth $2 trillion in market cap now move [37:26] with semis at correlations of 0.5 or [37:29] higher. 12 of these 15 are industrials. [37:32] names like Verdive, Eden, Caterpillar, [37:35] Cumins, Hubel, Hubel, Hubel, Hubble, um, [37:39] Comfort Systems. I don't even know. I [37:40] don't know these a lot of these [37:41] companies. Um, whatever you go on. Uh, [37:43] you say these are not tech stocks. They [37:46] trade like semis because their order [37:48] books have become AI capex order books. [37:50] Caterpillar sells backup generator sets [37:53] and engines into hyperscaler data [37:54] centers. Verda sells cooling and power [37:57] management. Ensal components. And Geova [38:00] sells gas turbines for data center [38:01] power. This is kind of interesting. The [38:03] gig sector classification has not caught [38:05] up with the economic exposure. [38:08] >> That sounds so good when you sound good [38:09] when you read it back to me. Yeah, [38:11] >> it's good stuff. [38:12] >> Buy that guy's research. [38:13] >> But [laughter] inside but inside the [38:15] stock market, I was looking I was [38:16] talking to Charid and Sean today. [38:18] >> Um, so healthcare is breaking out. It's [38:20] been stuck in the mud for a while. [38:22] Industrials look awesome. That's 20% in [38:24] the S&P. So for as much like people talk [38:26] about, oh, it's just a MAG 7. Daniel [38:28] chart uh throw chart 12 on this [38:31] surprised me um and I look at the market [38:33] pretty damn closely and I don't think I [38:34] knew this probably because they diverged [38:36] like very recently but the Mag 7 are up [38:40] um two uh 7.3% year to date. 7.3% year [38:44] to date for the Mag 7. Not bad. But the [38:46] 493 [38:48] are up 12.6%. [38:50] Huh? Did you guys know that? [38:52] >> Well, I knew about it. A big spread. [38:54] >> You mentioned it I think uh last week, [38:56] right? [38:56] >> I forgot about it. [laughter] I have [38:58] other things in my mind. [38:59] >> I I do listen. Um but I think what that [39:01] analysis is showing is that [39:04] >> this is going to support a lot of the [39:06] equal weighted indexes because it's such [39:08] a profound effect on the broader [39:09] economy, right? Like Caterpillar is [39:11] probably part of that S&P 493, right? [39:13] Like I think I was reading that [39:15] Generrack uh is doing better now because [39:17] of the data center buildout to build [39:19] backup uh you know um power uh for these [39:22] facilities. So um [39:25] that's why I say like just because I [39:27] mean is is the equal weight like [39:29] historically we look at that as a sign [39:31] of like market broadening but if the AI [39:34] tech capex boom is touching lots of [39:38] industrial names uh as an example um and [39:42] lots of freight right like the stuff [39:44] needs to move around the country that's [39:46] you know that's also probably helped by [39:48] all this so is it really like a sign of [39:51] breath or is it just It's a sign that [39:52] things are [39:54] really concentrated into one area of the [39:56] economy. Like what is like how much is [39:58] like traditional non-residential [40:00] structures like doing? [40:01] >> So when you say what area of the [40:02] economy, you mean corporate profits? [40:04] >> No, I mean like the tech sector, the [40:05] tech boom. [40:06] >> Damn it. I was trying to do a segue. [40:08] >> I make corporate profits. [40:10] >> So go ahead. [40:11] >> I mean it just seems like the the [40:12] breadth and the boom go together too, [40:13] right? It's just that there is you you [40:15] see companies like what Ford is kind of [40:17] trying to sell batteries now. [40:18] >> Ford's an AI company. [40:19] >> Yeah. Ford's trying to provide batteries [40:20] to support this sort of power boom, [40:22] right? So, we have it's it's just become [40:24] it it sort of touches so many other [40:26] sectors and even beyond whatever [40:29] rational logical um connection between [40:32] like oh this affects power which affects [40:34] like transportation. It's more like it's [40:37] a risk on environment and so this is one [40:39] in which transactions happen and [40:41] correlations do go to one in both [40:43] directions. So you actually if if it's a [40:44] an environment where investment appetite [40:46] is particularly solid that's one in [40:49] which other other companies other [40:51] sectors have a chance to participate as [40:52] well. [40:52] >> The other thing I was thinking I'd love [40:53] to hear your thoughts on this Michael is [40:56] that you know when people when clients [40:58] ask me like do you think the market's [41:00] pricing in a slowdown? [41:02] >> See like you're looking at it like but I [41:06] don't know I mean discretionary stocks [41:08] don't look great right? I mean like [41:12] that's what my that's what Jeff has been [41:14] telling me like uh discretionaries [41:16] underperforming consumer staples. [41:18] >> Which discretionary stocks? [41:19] >> Equal weight like he's been pointing [41:22] that out. I mean restaurants don't look [41:23] good. I mean you can say that's all [41:24] because of like people taking the fat [41:26] shot. [41:26] >> I I will say that unironically. [41:29] >> Um I think that it's I mean it's not [41:31] like their same store sales are doing [41:32] particularly well. Um or uh you know you [41:36] mentioned healthcare. I mean that's not [41:38] necessarily a cyclical sector but or I [41:41] mean what about what's going on with [41:44] like financials right I mean so when [41:47] people's like I don't I mean maybe the [41:48] market is like it's you know is is the [41:51] market pricing in some some slowed no so [41:53] tell me why [41:53] >> no okay so I'm looking at the equate [41:56] discretionary over staples and it's it's [41:58] still in an uptrend but flattening it [41:59] looks fine whatever it's sort of neither [42:00] here nor there [42:01] >> but I think some of the story is it's [42:03] really hard to separate I don't think [42:05] consumer discretionary stocks are [42:06] necessarily [42:07] always a reflection of the consumer and [42:10] restaurants are a great example of this. [42:12] A lot of these names, Cava um is like [42:15] the poster child of this. A lot of these [42:16] names came public sort of recently were [42:19] in vogue. The valuations were stupid. [42:21] The slowdown happened in uh consumer [42:24] preference changed. Shake Shack. Nobody [42:26] wants to eat a burger for lunch anymore. [42:28] The stock the stock is getting [42:29] destroyed. Is that a reflection of a [42:30] slowing macro environment? I don't know. [42:33] I don't even know what the dollar stores [42:34] tell you anymore about the state of the [42:36] like [42:37] >> how many needles we're buying? [42:38] >> Well, yeah, but seriously, you look at [42:40] Dollar Tree and Dollar General when [42:41] they're going up. Is that because that [42:43] particular consumer is doing better? Is [42:45] it because the middle income consumer is [42:48] trading down? I have no idea. Now, [42:50] obviously, you can listen to the [42:51] conference calls and they'll give you a [42:52] little bit more information, but I think [42:54] it's really tricky. So, I don't think [42:55] the market is pricing in a slowdown. I [42:57] think it's sort of hard to hard to say [42:58] that. Maybe the the PE is coming in a [43:01] little bit. Um, which is probably [43:02] healthy. Like there's a little bit of a [43:04] governor on the stock market, but I [43:05] don't I don't know. What do you think? [43:07] >> I mean, I think the consumer is spending [43:09] in total at a reasonable pace, but I [43:12] think one thing that's going to be [43:13] changing is they're facing more [43:15] inflationary pressures and it's again [43:17] there's obviously energy, but it's also [43:18] like food prices. If you think about a a [43:20] lot of what is being guided on the [43:22] staple side, right, suggests they're [43:24] going to pass through a lot of costs. So [43:25] you think about all the pet chem prices [43:26] that are in all your household cleaning [43:28] products and your household paper [43:29] products. It's basically just there is [43:31] assuming okay we're going to be able to [43:32] preserve our margin and that means we're [43:34] going to pass it through to the retailer [43:36] to the consumer. So someone's going to [43:37] hit the squeeze. It might be the [43:38] consumer in that case and that's really [43:40] like a real consumption squeeze. So this [43:42] the money can keep flowing even though [43:44] we may not necessarily be getting [43:45] richer. [43:45] >> Let me ask you guys this. I think a lot [43:46] of this is investor preference. For [43:48] example, Clorox, I don't know anything [43:51] about the business. I don't follow the [43:52] stock but guess what? The stock looks [43:54] terrible. [43:55] >> Is Clorox undergoing some stress? Is [43:57] there a competi like competitive [43:58] landscape that I don't Yeah, I'm sure. [43:59] I'm sure there is. But if you look at a [44:01] stock like Hyatt, which is at an [44:02] all-time high today, [44:04] >> and Delta, which is at an all-time high [44:05] today, I don't care what Domino's Pizza [44:08] stock is doing. That's a completely [44:09] idiosyncratic story that has nothing to [44:11] do with the broader consumer. Now, if [44:12] I'm Domino's Pizza CEO and Chipotle CEO [44:15] and I'm telling uh the analyst that is a [44:17] consumer macro pressure story, of course [44:18] you're going to say that. But I can show [44:20] you a million other examples that [44:21] completely refute that story. If [44:23] anything, what you're what you're [44:23] pointing out is actually very real in [44:25] the data about goods to services. So, we [44:27] see in goods, obviously, consumer [44:28] spending is kind of not as great. It's [44:31] not we're not seeing the spending in [44:32] terms of food and staples and even what [44:34] you kind of whatever else you get in the [44:36] grocery store. But if you see it in [44:37] terms of like services in terms of air [44:39] air travel, like air travel, even though [44:40] airfares are are going up quite [44:42] considerably, the actual volumes of air [44:44] travel are quite robust, right? This is [44:47] not something [44:47] >> restaurant sales haven't been that [44:48] great. [44:49] >> Restaurants have not been as great. I [44:50] think restaurants is also facing a big [44:51] like food squeeze. So everything your [44:53] costs on the food side have gone up [44:55] especially for Shake Shack. So [44:56] >> and I really think I completely [44:58] underestimated the GLP1 story [45:00] >> and it is having a material impact on [45:03] areas of the economy. It really is. [45:04] >> So in that sense you also get the [45:05] rotation there too. If you're not going [45:06] to spend as much at the grocery store, [45:08] you're not going to spend as much in uh [45:09] certain types of like food services [45:11] experiences, but you might spend it on [45:12] other kinds of recreation services. You [45:14] might spend it on accommodations, [45:15] resorts. Um and I do think we see this [45:17] we see that story play out as well in [45:18] the day. [45:19] >> All right. Uh shifting gears a little [45:20] bit, Greg Ip wrote uh a story, the [45:24] record divide between corporate profits [45:26] and worker pay. And Denise was on this [45:30] uh Denise Chisum was on last week [45:32] talking about um falling unit labor [45:34] costs being a tailwind for the stock [45:36] market. Great for the stock market, [45:37] pretty shitty for society. Um so Greg [45:40] said, "To understand why people are so [45:42] miserable about the economy, look no [45:43] further than Thursday's report on gross [45:45] domestic product, not how much GDP grew, [45:47] but how it was divvied up. Worker [45:49] compensation, wages, and benefits grew [45:50] 0.8% in the first quarter. Um, while [45:53] domestic corporates profits jumped 2.7%. [45:56] Daniel, charts on, please. As a result, [45:58] labor share of gross domestic income [46:00] sank to 51%, the lowest since records [46:04] began in 1947, and profit share climbed [46:07] to 12.1%, the highest since 1950. So, [46:10] just go back and forth between these [46:12] charts. So, this is corporate profits [46:13] basically hitting all-time highs. And [46:15] the flip side of this is wages just not [46:18] getting not getting their fair share and [46:20] this is ripping the country apart. Is [46:23] this not being accounted for properly or [46:24] do you think like this really is the [46:25] story? [46:26] >> I think it's probably two things I can [46:28] think of. One is the labor market has [46:29] been underperforming. Um so when we say [46:31] like [snorts] hey given like what we're [46:34] seeing in the labor market being so [46:35] sluggish consumer looks pretty good and [46:37] given these facts like oh okay it's [46:39] happening despite the labor market not [46:40] being as good this is quite impressive [46:42] but the labor market is sluggish in a [46:44] lot of ways like benchmark to what we [46:46] saw in the 2010s even this is slower [46:48] than that and yet we're also seeing the [46:51] structural trend a lot of result by the [46:53] tax system right the tax the tax code [46:55] has sort of created a lot of biases in a [46:58] way that makes it just um you're going [47:00] to be more inclined to to to try and [47:02] stay ste at the margin steer away from [47:04] W2 income towards either trying to own [47:06] your own um business. Uh there's [47:09] obviously a lot of stuff with capital [47:10] versus labor. I'm not going to get too [47:11] much into that, but these are all ways [47:13] in which um there's going to create some [47:15] bias and some of that might be real and [47:17] some of that just might be [47:18] classification. So it's it's not just [47:20] like some big political thing. It's [47:22] superstar firms is probably another [47:24] argument, right? like just like firms [47:26] that have really high margins and we [47:28] have more of them in our economy. [47:29] Therefore, naturally, as a result, [47:31] profit share goes up. [47:32] >> But this is this is very much impacting [47:34] the political landscape and how people [47:36] vote. And this is like a huge part of [47:38] the story. [47:40] >> Yeah. I mean, I think it's the if you [47:42] aren't owning your own business, right, [47:43] and you're basically relying on a W2, [47:46] then what exactly is like what is that [47:48] what does that labor market look like? [47:50] I'd actually argue right now we're [47:51] seeing a pickup in white collar [47:52] employment. we're seeing a pick pick up [47:54] there which is kind of defying all the [47:56] odds and prognostications on the AI um [47:59] replacing white collar and but for blue [48:01] collar it's not necessarily all great [48:03] right it's actually uh if you look at a [48:05] lot of different types of uh employment [48:07] there job growth has not been as [48:09] fantastic um there's a lot of [48:12] construction jobs tied to building data [48:14] centers outside of that not so great um [48:16] and even manufacturing if you're maybe [48:19] working for Boeing obviously jobs are [48:21] being created because they're ramping [48:23] production after all their snafoos. [48:25] Outside of that, not as impressive, I'd [48:26] say. [48:26] >> Neil, do you think that those lines um [48:29] forget about converging, do you think [48:31] they stop diverging? Is there a breaking [48:33] point? [48:35] >> So, I I think I sent you this chart, but [48:37] basically, if you pull up a chart of [48:38] like nominal GDP [48:39] >> chart time, Daniel [48:41] >> um and uh like sort of nominal [48:44] compensation growth, like I think it [48:47] kind of speaks for itself, right? Like [48:48] nominal GDP is growing about 6% [48:53] uh and nominal compensation like wages [48:56] and salaries mostly is is running below [48:59] 4%. [49:00] >> Has that ever has this ever happened? [49:01] >> I mean I'm sure it's like it's rare like [49:04] like the disconnect is quite unusual. [49:07] Now there's a I mean and I I I've said [49:09] this to our clients and I'll say it to [49:11] your audience. How you feel about how [49:13] this thing reconciles will should [49:15] dictate how you feel about the [49:16] trajectory of like policy going forward. [49:19] >> Go on. Well, I mean, if you think um if [49:22] you think that all this spending is [49:25] going to lead to a meaningful inflection [49:27] higher in like labor income and tight [49:28] job market and you know, people seeing [49:30] stronger wage growth, [snorts] then you [49:32] should be very hawkish. Like I mean, you [49:34] should you should expect, right? Like, [49:36] so if the gray line converges towards [49:37] the red line, then [49:39] it's not like I think Scanda has a [49:41] forecast for like one hike at some [49:43] point. [49:43] >> Yeah. [49:44] >> Um in the next year, forget that. I [49:46] mean, they're going to go 75 to 100. [49:48] They'll they'll take away all the [49:49] insurance cuts from last year. Okay, so [49:52] that's at least 75. The Fed the Fed [49:54] never I don't think really just goes [49:55] once. I mean, um [49:58] >> but if you think that we'll see that [50:01] that slowing in nominal wages and [50:03] salaries will p pull down to some extent [50:06] consumer spending from like maybe 2% to [50:08] one one and a half. Uh you continue to [50:11] see this sluggish growth in residential [50:14] investment. You continue to see sluggish [50:16] growth in structures investment. So if [50:17] that red line converges onto the gray [50:19] line, well then maybe the Fed can wait [50:22] it out. Maybe the Fed can wait it out. [50:24] And so I think that to me is the kind of [50:26] conversation like so I think Scandal is [50:29] a little bit more on the hawkish side of [50:30] things right now than I am. Um it's [50:32] unusual for us to be [laughter] that [50:33] that that unaligned, but that's kind of [50:35] where where the debate is right now. So, [50:38] you know, for me, like [50:41] in my career, I've always put more [50:43] weight on labor and housing. And so, [50:47] that's why I probably sit more on the [50:49] dovish side of things at the moment. Um, [50:53] you know, how can you, you know, as an [50:54] example, like how can like apparel [50:56] prices have been rising very, very [50:57] rapidly over the last year, but what's [51:00] happening to the real volume of clothing [51:02] that's being sold? It's actually [51:03] contracting. Like, how can firms make [51:05] that stick, right? It's very difficult [51:07] to make the price increases stick if [51:09] labor income isn't there. [51:10] >> Where else is so so areas? All right. [51:13] Kevin Walsh. [51:14] >> Oh god. [51:15] >> Well, not your favorite. Um I think last [51:18] time you were on here, you were you were [51:20] not so happy about the prospects. [51:22] >> I mean, someone has to do the the the [51:24] dirty work. I mean, you can't get half [51:26] the people on the street to actually say [51:27] what they really think about them. So, I [51:28] guess if I if one person has to do it, I [51:30] guess I will. You know, that's sort of [51:31] how I think about it. I mean, that's one [51:33] of the benefits of working at a smaller [51:35] place. You don't have to be like a [51:37] diplomat. But yeah, I don't I don't [51:39] know. I mean, what's the upside? [51:41] >> What's the upside of what? [51:42] >> Him. [51:43] >> Oh, well, he's in charge now. He's the [51:45] boss. [51:45] >> Well, I mean, I think it's as likely [51:47] that the Fed captures him than he [51:49] captures the Fed, right? He's going in [51:50] there talking about regime change. I [51:52] think [snorts] uh if I were if I were a [51:55] betting man, I would say that the Fed is [51:57] more likely to influence him than the [51:59] other way around. [52:00] >> I think actually it's a bit of like [52:01] Chinese finger trap, right? the harder [52:03] you try to pull away, try to the harder [52:04] you try to push for a doubbish case, the [52:07] hard the more you're likely to kind of [52:09] stoke a reaction from the rest of the [52:12] his colleagues. [52:14] >> Well, yeah, if you if you make bad [52:15] arguments, [laughter] [52:17] >> well, if you make bad arguments, don't [52:19] be surprised when like the staff or your [52:20] colleagues just smack you down, right? [52:21] And I think that's the that's the big [52:23] issue with like he's kind of trying to [52:24] throw in trimmed me and I'm actually a [52:26] data guy now for like [52:28] >> Yeah. Look, look at his career, Michael, [52:30] if you've ever seen him, you know, when [52:32] when he first got when he was governor, [52:35] he never gave actually speeches on the [52:37] economic outlook. Like, you know, it's [52:38] like, oh, here's Governor War or Lorie [52:40] Logan like talking about the economic [52:42] and and and policy outlook. He never [52:44] gave those speeches really. Um, if you [52:46] go through his like record, like most of [52:48] his speeches are like sort of like very [52:51] kind of like highlevel like [52:54] philosophical. he's like freaking going [52:56] around talking about Emanuel Kant and [52:58] like these like people that you've like [53:00] heard in like literature like class like [53:02] in your freshman year college or [53:03] something. I mean it's it's very bizarre [53:05] to be talking in those terms for like a [53:07] central banker. [53:08] >> Can't people change? [53:10] >> Um I think the change is very curious [53:12] considering it happened during a period [53:14] where he was actively campaigning for a [53:16] job. And so to me that's like to me [53:18] that's like the knock on him. I [53:20] >> I saw somebody [53:21] >> but for me it's just not going to work. [53:23] Do you know what I [53:24] like going in there and being like, "Oh, [53:26] look at this trim mean measure. It's [53:27] lower than everything else." Like if his [53:29] job is to go in there and try to [53:31] convince the people around that table [53:33] that we're about we're on the precipice [53:35] of a golden age. I mean, you hear them [53:37] talk about this all the time like pull a [53:38] Greenspan as if Greenspan didn't hike [53:40] aggressively in the late 1990s. Um [53:43] anyway, um so he's going in there to [53:46] sell this golden age thesis. Basically, [53:48] what that means that we're on the front [53:49] edge of a productivity boom. That means [53:50] we have a lot of spare capacity. [53:52] >> So don't do anything. Yeah, we have a [53:53] lot of spare capacity in our an economy. [53:55] That means we actually have more room to [53:57] cut rates. That's what he's going in [53:59] there. That's the pretense under which [54:01] he was brought in. [54:02] >> So, Scanda, [54:04] >> yeah. [54:04] >> Oh, were you done? [54:05] >> I can keep [laughter] I can keep going. [54:08] Uh, Scanda knows. [54:09] >> So, Scanda, my question to you is um how [54:12] how much power does he have to influence [54:16] policy? Can he just say we're doing [54:18] we're cutting, we're raising? I think I [54:21] think he he he can't just snap his [54:22] fingers. He's got some bully pulpit [54:25] power. He's got some ability to set the [54:26] agenda at a meeting. Like those are like [54:28] for like a really shrewd operator. There [54:31] are ways to leverage that. Well, but [54:33] it's certainly not something so unitary. [54:36] It's a you vote at a committee. You vote [54:38] among a set of people that's uh each got [54:40] one one person, one vote. Um and there's [54:43] a set of members. Some of them are part [54:45] of the board of governors and some of [54:46] them are regional Fed presidents. Um, [54:48] and you already saw what, four descents [54:50] at the last uh, Fed meeting. One of them [54:52] was on the more doubish side, but three [54:53] were on the hawkish side. That's a lot [54:54] of descents. And they're probably going [54:56] to descent again because they're going [54:58] to say, "Ah, yeah, we should get drop [54:59] this easing bias, but we really need to [55:01] be moving towards a tightening bias. Um, [55:03] we already got some battle. [55:05] >> That'll be the next battle." And we're [55:06] headed in that direction. I'm I would [55:08] separate what I think the Fed should do [55:10] from what the Fed will do. [55:11] >> So, what do you think they should do? [55:12] What do you think they will do? [55:13] >> I think they should steal themselves. [55:15] the fact there are a lot of supply [55:16] shocks in the economy and there there [55:17] are a lot of supply shocks that are [55:19] really hard to look through. The first [55:20] thing you do is don't underestimate [55:22] their scale and duration. Um, which is I [55:24] think a big part of what got them in [55:25] trouble in 2021 and 22 which is that [55:27] it's just uh these things can last a lot [55:29] longer than they than you think because [55:31] they take a long time to travel through [55:32] the value chain. Different people are [55:34] move different companies are moving [55:35] their margins. Um the scale of this [55:37] stuff can be a lot larger than you think [55:38] in real time. Um so don't go in uh [55:42] basically underbaking these forecasts. [55:44] So dumb question. How much how much do [55:46] you think the overnight rate impacts? [55:47] Like what does that have to do with [55:48] supply shocks? [55:49] >> It doesn't really. But the biggest issue [55:51] is that the Fed has been underbaking [55:52] where inflation is supposed to be by [55:54] this point. And when you keep making the [55:56] same error, you're liable to say, "Oh, I [55:58] must be missing something really big. I [56:00] must be must need to raise interest [56:01] rates to keep inflation expectations [56:03] anchored." And I think we're headed down [56:04] this path. I mean, I'm basically being [56:06] bearish about the Fed because of the [56:07] reason Neil uh pointed out, which is [56:09] that they are going to confuse this [56:10] stuff. They're going to look at the [56:11] inflation data. They're going to say, [56:13] "Oh, it's it's hot. It's hot for so [56:14] long. It's in too many different [56:16] components because it's so broad because [56:18] it's lasted so long. It therefore must [56:20] be something we have to do something [56:22] about." [56:22] >> So, you think they should look through [56:23] it, but they won't. [56:24] >> So, he thinks they should do what my [56:25] call is. His call is that they won't. [56:27] That's why he's negative. So, I think [56:29] that's it's an interesting framing. I [56:30] mean, we'll see. I mean, [56:31] >> but the market is pricing at a rate [56:32] hike. Show chart 13, please. So, this is [56:34] the implied Fed funds rate. Um, we're [56:36] looking at through the end of September [56:37] 2027. What is this line telling us? [56:40] we've it's kind of the cowardly [56:42] compromise of saying that there is [56:43] either going to be a set of hikes, [56:45] right? If they're Fed's going to hike [56:46] once, they're going to hike multiple [56:47] times, or they're just going to stand [56:49] pat. I think it's going to be hard for [56:51] them to steal themselves through what [56:52] might be some more inflation and a a [56:54] broader pickup. And at a time when [56:57] financial conditions, if at a first cut, [57:00] pretty supportive. The interest rate [57:02] currently is roughly around neutral, [57:05] modestly restrictive, and at the same [57:07] the labor market's not showing the same [57:09] downside risk. We're moving away from [57:10] the left tail. Inflation has got some [57:12] right tail properties. It I do think [57:14] it's largely supply, but that's a hard [57:18] when you have all these things moving in [57:19] the same direction of the hawkish dovish [57:21] debate. [57:23] >> They're not going to go once. They're [57:24] going to go 75 to 100. [57:26] >> If they start if if they start to go, [57:28] don't be surprised if it's three three [57:29] hikes in a So Neil, you're a market [57:31] economist. If you knew that they were [57:33] going to go what, three times 25? [57:36] >> If they're doing it for the reasons he's [57:38] talking about, it would be very bad for [57:39] the economy in the capital markets [57:41] because because it's basically you're [57:43] hiking rates, not because demand is like [57:46] we're trained to think rate hikes are [57:48] okay because it means that the economy [57:49] is strong because demand is there and [57:51] that means earnings are there. [57:52] >> But the economy is not that strong. [57:54] >> Well, that's Yeah, the economy is not [57:57] horrible. It's it's it's okay. It's [57:59] okay. [57:59] >> It's not overheating to the point that [58:01] >> it's very uneven. I mean, I thought the [58:02] Beige book this week was very [58:04] interesting. It's a I haven't seen a [58:05] beige book like that in a while. It's a [58:06] very uneven economy, I think, is is a is [58:08] a fair kind of characterization of it. [58:10] Um, [58:11] >> but if the Fed's hiking and it's because [58:15] like we have this sort of like broad [58:17] like grow growth and labor markets are [58:19] fine, like I think the markets are going [58:20] to be fine with that. But if they're [58:22] hiking for the reasons that Scandanda is [58:24] talking about, like that is not a good [58:25] outcome. like hiking. It's like, [58:28] remember what? Remember what Josh was [58:29] talking about last week? [58:30] >> I have no idea. [58:31] >> Well, he was talking about how like it [58:32] makes no sense to hike because of oil [58:34] prices going up. Now it's like, okay, [58:35] now you're going to hike because cattle [58:37] prices are up. Now you're going to hike [58:38] because uh you know, I mean, can the Fed [58:41] control the flow of oil through the [58:42] straight of hormones? Absolutely not. [58:44] Can it does it have any power over the [58:46] El Nino? Like no, it does not. The one [58:48] thing I wanted to come back to on Worsh [58:50] was I did find it fascinating [58:53] like in the last week that Bernani, what [58:55] did he say about like his views on the [58:57] balance sheet? He called it a [58:58] meaningless statement. I thought that [59:00] was really fascinating because obviously [59:01] Worsh worked for for uh for Bernani and [59:04] the balance sheet for whatever reason is [59:06] like Kevin Worsh's hobby horse and [59:08] Bernani just took a big mass of crap all [59:11] over that. So I think it's kind of like [59:13] emblematic of how he actually feels [59:15] about him without saying it. But [59:17] >> I want to get your guys' take on this. [59:19] Um, an a listener emailed us about the [59:23] labor market. He said, "Many of the [59:26] distractors of this bull market, I think [59:28] it meant detractors, often say the lack [59:30] of jobs and the lack of good, [59:32] high-paying jobs and knowledge services [59:33] are reasons for doubt and concern. The [59:36] latest report from the Bureau of Labor [59:38] shared that the number of job openings [59:40] is now at a 2-year high with 7.6 million [59:43] open jobs. More interesting is that the [59:46] largest jump in openings was in [59:47] professional and business services [59:50] adding 668,000 positions combined with [59:53] the past two months of higher than [59:54] expected non-farm payroll data. Do we [59:57] think sentiment is ready to turn and [59:59] finally embrace that AI is not going to [60:02] eat all the jobs and is in fact helping [60:04] to grow and create opportunities for [60:07] workers? [60:09] >> Well, first I I mean I sort of reject [60:11] the premise that AI will actually eat [60:13] all the jobs. I mean, you know, that's [60:15] kind of like the lump of labor fallacy. [60:17] I never really bought into it. It's [60:18] really hard to talk about like AI taking [60:20] away all the employment with, you know, [60:23] the unemployment rate at 4.3%. [60:26] I mean, or on a prime age employment is [60:28] still fairly good. I mean, so I don't I [60:30] don't really buy into that. I mean, I [60:32] will say that you don't want to get too [60:33] much into the indicator macro. Like, I [60:35] don't really believe that professional [60:37] and business services openings went up [60:39] by 700,000 uh in in a month. fake [60:41] postings are [60:42] >> I mean I think the data are real. The [60:44] news is fake. I mean you know to borrow [60:45] from from Trump um I you know I mean [60:48] it's like anything else you could find a [60:49] different indicator like the Indeed job [60:51] postings numbers are weaker uh on net [60:53] over the last few weeks. So [60:55] >> I don't think the signal from that [60:56] report really comes um from openings [61:00] frankly. I mean it's like openings are [61:02] like the fakest thing. [61:03] >> All right. Well here's here's a more [61:04] tangible um [61:05] >> it's hires and quits that matter a lot [61:06] more. [61:06] >> Here's a more tangible reading. initial [61:08] jobless games was it last summer when [61:10] they started to go up and people were [61:12] getting I'm sure I know Josh said this [61:15] these don't usually slow down like this [61:17] is the ultimate object in motion stays [61:18] in motion type thing and they did slow [61:20] down they peaked and then they [61:21] normalized um [61:23] >> are you guys surprised at the low level [61:25] of initial claims just how like okay the [61:28] the labor market has been [61:30] >> I am surprised that we saw such a big [61:32] slowdown in job growth in 2025 and it [61:35] kind of stabilized on its own right so [61:37] that's something that is one of those [61:38] like object motion states in motion. I [61:40] think there's a lot of good reasons to [61:42] take that seriously like there's a lot [61:43] of information in the present is just [61:45] respect. Um but it did start to show [61:47] stabilization and so we started to see [61:49] that probably around what was that jobs [61:51] report in December you started to see [61:52] signs of like okay things post shutdown [61:55] were starting to to like um just just [61:58] show the kind of stability that job [61:59] growth was not going to keep falling off [62:00] a cliff. Um and I think it's actually [62:03] most interesting on when you have a [62:04] white collar. So you think about there's [62:06] a subset of prof professional and [62:07] business services that really matters [62:08] because it also includes things like [62:09] temp help jobs and a lot of other things [62:11] that are not exactly high wage there you [62:14] are seeing a pickup you are seeing a [62:15] pickup in in job growth and so it's [62:17] hiring is picking up there and if you're [62:19] really trying to put it together is this [62:21] because um like what is the the macro [62:24] fact that's driving this is not like [62:26] completely obvious but I think if I take [62:27] one step back it's look it's been a [62:29] soggy labor market for th that segment [62:32] for about three years basically you had [62:34] the sort tech session of 2022, you had [62:37] basically uh okay, the Fed is hiking, so [62:39] therefore we need to be on watch for [62:41] recession. We shouldn't we overhired [62:42] during the pandemic, we need to slow [62:44] down. Then it became, well, this AI [62:46] stuff is so risky. Should we really be [62:48] increasing headcount? But like the [62:50] actual cost of capital signal has long [62:52] been signaling you it's okay to spend. [62:54] >> Well, also, don't you think it's it's [62:55] just so it would be so deeply unpopular [62:58] for JP Morgan, for example, to start [63:01] doing massive layoffs? like do you think [63:03] that companies and I know JP Morgan's [63:05] you know an extreme example um but do [63:07] you think companies in the aggregate [63:09] have any sort of hesitation to do that [63:12] for fear of some sort of retaliation by [63:14] either their own workforce or larger [63:16] forces at play political forces or or [63:18] the such [63:19] >> I think there's actually a lot of risk [63:20] that you lose a lot of knowledge to be [63:22] able to orchestrate future solutions [63:24] right so you end up like lo let letting [63:27] go of someone who is responsible for a [63:28] key system you're letting go of someone [63:30] who's able to like build the set of [63:33] things that work kind of actually [63:34] leverage the technology. I mean, it's [63:37] not obvious to me that like who's most [63:40] exposed to AI. We don't really know, [63:42] right? Like we don't know which kind of [63:43] jobs are actually most exposed. Oh, aha, [63:46] these jobs are going to be the ones that [63:47] are easily automated. It may very well [63:48] be the case that's the white collar [63:50] people who need to be actually executing [63:52] a lot of the um sort of how how to take [63:54] advantage of the technology itself. I [63:56] mean, you do see specific areas in the [63:57] job market that are booming right now [63:58] because of AI, but I would just say that [64:00] layoffs are usually like the last thing [64:02] to go. I mean, [64:03] >> by the time layoffs are are showing you [64:05] something in terms of a signal, like [64:07] it's over, buddy. [64:08] >> So, let's let's leave the audience with [64:10] this. Um, and you guys are great. Uh, [64:12] Scandanda, how could people find your [64:13] work? [64:14] >> Um, so employeeamerica.org is a great [64:16] place to go. Uh, you can access our [64:17] research. We have some public research [64:19] if you want to subscribe to. We have a [64:20] few different distributions there. Um, [64:22] if you want to follow me on Twitter, [64:23] right now I'm probably posting mostly [64:24] about the Knicks, but uh but but I [64:27] promise after the the series ends and [64:29] hopefully well, um, we back back at it [64:32] and um, we're usually breaking down a [64:34] lot of data related to jobs and CPI and [64:36] PCE and everything the Fed's tracking. [64:38] >> All right. Uh, Neil, we've got a QR [64:40] code. If people are watching this, they [64:41] could they could [64:42] >> Yeah, you can also find me on the New [64:43] York Knicks Instagram account. [64:45] >> Hell yeah. Um, [laughter] [64:46] where else are you posting? [64:48] >> I post on LinkedIn. I like the longer [64:50] form content. We obviously post charts [64:52] on Twitter as well. X, excuse me. Um, [64:55] speaking about an area where uh they cut [64:57] to the bone and you know, I'm actually [64:59] glad that more tech people didn't follow [65:00] his lead. Like there was a lot of fear [65:02] about that. Like they cut everybody at [65:03] Twitter and it still works. And I know [65:04] to the degree that it works, but it but [65:06] it did, you know, it's still still doing [65:07] its thing. All right. I want to leave [65:09] people with this. Um, [65:12] everybody wants to know when it's going [65:14] to end. I get it, right? It's just it's [65:16] it's just how we're wired for better or [65:18] for worse. Um and maybe maybe it's [65:20] tomorrow, maybe it's in 2032 like and [65:23] when I say it ends, I mean like the [65:25] cycle turns. Okay. So what is like one [65:28] or two things pieces of actual evidence [65:31] or actual data that you guys would look [65:34] to for this particular cycle because we [65:36] know this is a unique one. So what are [65:38] you going to be keeping an eye on? [65:40] >> You want me to go first? [65:41] >> Sure. [65:42] Well, I would say at the end of the [65:44] year, I think, you know, into next year, [65:47] maybe the conditions are not as [65:49] favorable as they are at the moment, [65:50] right? Like, so we probably don't get as [65:53] much of a fiscal push, right? Like the [65:54] one big beautiful bill tailwind is going [65:57] away by the end of the year, that that [65:59] transitions to more of a headwind. Um, [66:01] if the Fed is on hold, even as nominal [66:04] compensation growth remains weak. I [66:05] mean, everyone's talking about a passive [66:07] easing of policy because of inflation, [66:09] but for the labor market, that's a a [66:11] passive tightening of policy. Um, so [66:14] those two things are kind of on my mind. [66:16] Um, [66:16] >> that was way too smart. I don't know [66:17] what you just said. [66:20] >> SC. [laughter] [66:22] >> Yeah, business cycles are risk cycles. [66:24] And so the risk and the willingness to [66:26] actually spend is what we should all be [66:29] caring about if we're trying to think [66:30] about the big macro risk, the big draw [66:32] down you want to avoid or the big um [66:34] risk to your job. [66:35] >> It sounds like you're you're saying [66:37] credit spreads [66:38] >> maybe credit spreads. [66:40] >> That would be my answer. [66:40] >> Yeah. So credit spreads are a very good [66:42] proxy say especially episodic though by [66:44] the time it turns out [66:46] >> but but I feel like it's like the e [66:47] here. Okay, here's the thing. Um the [66:50] during I it I think it was liberation [66:52] day. The the thing that kept me like [66:54] moderately comfortable with the [66:57] direction of the economy is the credit [66:59] market really wasn't freaking out. [67:01] >> It like it just it just wasn't. And so [67:03] you're right by the time like it really [67:04] shows it'll probably be too late. At [67:06] least that's kept me on like the right [67:08] side of like the positive trend. [67:09] >> Yeah. To me the thing when you think [67:11] about like where's the balance sheet [67:13] constraint that's going to cause this, [67:14] right? And so it could show up in [67:15] corporate credit spreads. may be [67:16] something that's the things to me that [67:18] are most likely are something that where [67:21] it's the willingness to spend with [67:22] probably the non- tech companies on AI [67:24] expenditure. That's really what we [67:26] should actually be focused on because [67:28] for them it's an expense and they need [67:29] to see payoff on that expense at some [67:31] point. Whereas if it's just spending by [67:34] the hyperscalers they're it's all part [67:36] of their moat. It's part of uh um how [67:38] they're trying to make sure that they [67:39] stay competitive. But for the non- tech [67:41] companies like are they seeing a return [67:43] on their investment? I'm sure they are [67:44] seeing something but like you're [67:46] obviously hearing about like token [67:47] expenditures and all that stuff. So when [67:48] that gets stretched, which is kind of [67:50] this the scenario we think about 2000 [67:52] where it was a lot of frontloading of [67:54] expenditure around Y2K u at some point [67:56] just got tired of it, right? It's like I [67:58] I don't need to keep spending more and [67:59] more and more. I'll just spend the same [68:00] amount and that's important. [68:02] >> What's nice is that Nvidia is now [68:04] breaking out their revenue by [68:06] hyperscaler and everybody else to give [68:08] us maybe a better sense of like no, [68:09] we're actually we're not just relying on [68:12] these, you know, yeah, [68:13] >> these behemoths. [68:14] >> Um all right, boys. How are we feeling [68:15] for the rest of the series? better than [68:17] I did uh 24 hours ago. I'd say we took [68:20] their heart out yesterday. You know, [68:22] >> I think they're a resilient bunch. [68:23] They're not like they're not the Cavs. [68:25] Like um Castle's a dog. He's not going [68:27] to give up and so is one. [68:29] >> No, it's not going to come easy. But I [68:30] also I mean I felt like the Knicks I [68:32] still do. The Knicks are the more mature [68:34] team. They'll handle it. And um I think [68:36] they needed that the Spurs needed that [68:38] game more than we did. [68:40] >> I think they're pretty exhausted from [68:41] the previous series. And [68:42] >> goodbye ran out of steam. [68:43] >> Yeah, he ran out of steam the last few [68:45] minutes game. I I it still doesn't feel [68:48] I still can't believe this is happening. [68:50] It's been [68:51] >> same. [68:51] >> It's been so long. [laughter] [68:53] >> I mean, he It's funny you mentioned PJ [68:55] Brown. I mean, I remember I was an [68:57] 11year-old kid when when [68:58] >> Hated that guy [68:59] >> when uh when when the Knicks uh you [69:01] know, in '94. I I still remember that [69:04] series. A series we should have won. [69:05] >> That one's a little bit foggy. I was [69:07] like I was I was nine. [69:09] >> So, I don't remember. [69:10] >> We were We were up 3-2. [69:12] >> Yeah. [69:12] >> In the Garden. And we should have won [69:15] that game. [69:16] To me, I'm just more cynical after two [69:19] decades of Dolan and it's like how how [69:21] is this coming together? [69:21] >> There's no time for citizen. [69:22] >> Sorry, no time for citizen. [69:23] >> We're done with that. All right. Um, [69:25] thank you everybody for listening, for [69:26] reaching out. You guys are awesome. This [69:28] is so much fun. Scandal would love to [69:30] have you back. This is really great. [69:31] Thanks, guys. We'll see you next time. [69:33] Appreciate you watching. [69:40] >> [music]