AI Capex Replaces Housing as US Economic Engine, Kawa Says

Kawa：AI资本支出取代房地产成为美国经济主要引擎 创纪录的人工智能基础设施支出已取代房地产成为美国财富和盈利增长的主要驱动力；若这些由债务驱动的支出未能产生下游利润，股市将面临脆弱风险。 Luke Kawa 与主持人共同探讨了当今金融市场的极高周转速度与内部撕裂：尽管大盘指数逼近历史高位，底层市场广度却异常低迷。讨论深入分析了超大规模计算厂商的庞大资本支出如何取代传统房地产，成为宏观经济周期和消费财富效应的核心引擎。双方探讨了硬件端 AI 受益标的与惨遭抛售的非必需消费品股票之间的剧烈分化，以及在全球持久增长预期（而非失控通胀）推动下 10 年期美债收益率破 5% 的历史性债券抛售。Kawa 最终指出，虽然在算法交易与超短期散户投机冲击下市场已变得“快到难以逆势做空”，但历史经验表明，等待远期盈利预期下调回落，依然是研判最终市场见顶最严谨的信号。 市场广度恶化至历史极端水平：虽然 S&P 500 指数距离历史高点不足 0.5%，但其成分股中仅有 51.2% 位于 200 日移动平均线上方——这种背离程度上次出现还要追溯到 2000 年 3 月互联网泡沫见顶的次日。 (4:59) 超大规模厂商的资本支出贡献了今年全市场约 50% 的盈利增长，目前主导 Russell 1000 指数估值倍数的是未来三年的远期销售预期调整，而非近期业绩。 (10:00) 在持久的全球增长预期和稳定的通胀保值预期支撑下，基准 10 年期美国国债收益率冲破 5%，使固定收益沦为低优先级边缘资产，难以匹敌 AI 带来的巨额预期回报。 (17:15) 美国股市财富规模已正式超越住宅房地产财富，AI 资本支出与股市升值已实质上取代房地产，成为经济周期与财富效应的主要驱动力。 (45:00)

Financial markets have entered an era defined by extreme velocity, where massive artificial intelligence capital expenditure has replaced housing as the core driver of economic momentum. While headline equity indices hover near record highs, market strategist Luke Kawa argues that underlying participation is historically narrow, masked by rapid algorithmic trading and sustained by legacy tech revenues. Narrow Breadth and Capex-Driven Valuations (4:59) Headline equity benchmarks belie severe internal fragility. Kawa notes that heading into late September, with the S&P 500 within 0.5% of record highs, only 51.2% of its component stocks traded above their 200-day moving averages—a technical divergence matched only at the March 2000 dot-com peak, though similar weakness in 1998 persisted for two years before equities rolled over. Meanwhile, rate-sensitive consumer equities have suffered deep drawdowns, epitomized by Nike tumbling from $180 in 2021 toward $35. Market performance is overwhelmingly anchored by infrastructure investments rather than downstream software productivity. A Goldman Sachs study cited by Kawa found that hyperscaler capital expenditures alone accounted for roughly half of the S&P 500's earnings growth this year. Furthermore, Russell 1000 valuation multiples show almost no statistical correlation with near-term sales revisions, trading instead on three-year forward sales forecasts. However, Stanford economist Hanno Lustig calculates that tech platforms must eventually capture revenue equal to roughly 9% of U.S. GDP simply to earn an adequate return on cumulative capex projections exceeding $1.2 trillion. Surging Real Yields and the Allocation Squeeze (17:15) Benchmark 10-year U.S. Treasury yields breaking above 5% reflects a synchronized global growth impulse rather than unanchored inflation fears, as market-implied inflation breakevens remain steady. Capital-intensive investments in AI power and data infrastructure are spilling into heavy industry, while Macquarie strategist Thierry Wizman notes that long-term rearmament, defense reshoring, and supply-chain reconstruction represent an additional structural, capital-draining inflationary force. Because nominal growth and inflation remain elevated, fixed income has become what Kawa calls a 'residual' asset. Multi-asset managers find 2% to 3% real bond yields unattractive relative to expected AI equity returns, forcing real bond yields higher to compete. Kawa observes that if AI supremacy is treated in Washington as an urgent national security imperative, funding it may eventually require forms of financial repression—such as wartime-style 'AI savings bonds' designed to curtail consumer spending and direct scarce materials toward data centers. Compressing Reaction Windows and Trading Velocity (25:00) Kawa characterizes the prevailing environment as 'too fast to fade,' defined by the rapid normalization of geopolitical shocks and unprecedented intraday sector rotations. This velocity is propelled by a structural confluence of multi-manager pod shops operating with tight risk limits, programmatic dispersion strategies, and expanding retail participation. Retail flows on Robinhood returned to roughly 89% of their late-June peak by September, with speculation heavily focused on options expiring in four days or fewer. Microstructure changes have also eliminated historical market reaction lags. In earlier eras of machine-readable news, algorithmic parsing triggered temporary one-second liquidity vacuums as quote providers stepped back; today, text ingestion and execution are near-instantaneous. While traders are experimenting with autonomous agentic trading bots to scan academic research and mispricings, Kawa warns that widespread algorithmic order execution risks thinning resting order-book depth and exacerbating single-day volatility. AI as the New Business Cycle Engine (36:30) AI infrastructure has largely supplanted residential real estate as the primary transmission mechanism of the business cycle. With aggregate American household wealth in equities having officially surpassed wealth held in real estate, market appreciation and massive data center construction now drive the wealth effect and corporate investment even while the housing market remains frozen under high borrowing costs. This cycle functions through a self-reinforcing loop: corporate capex lifts equity valuations, stimulating consumer spending and sustaining earnings. Yet Kawa emphasizes that the current buildout is largely financed not by AI profits, but by the lucrative legacy franchises of tech monopolies alongside rising corporate borrowing. Emerging balance-sheet strain—visible in widening credit default swaps for Oracle and CoreWeave—points to a transition from hyper-growth to growth-at-a-reasonable-price (GARP). To identify an ultimate market peak, Kawa recommends tracking 52-week forward earnings estimates rather than breadth, noting that estimate rollover reliably signaled six of the last seven bear markets. The market's self-reinforcing momentum hinges on whether downstream applications can generate genuine productivity and enterprise demand before hyperscaler balance sheets become overburdened. If forward earnings estimates peak or credit spreads on leveraged tech providers widen materially, the infrastructure-driven equity expansion will face severe fundamental pressure.

## Background 本期节目由彭博社（Bloomberg）旗下的《Odd Lots》播客于 2026 年 10 月 1 日在 RSS 及各大网络平台发布。根据页面简介，节目特邀前彭博记者、现任 Sherwood News 市场主管 Luke Kawa 做客，探讨指数狂飙与个股低迷、利率上升与经济韧性并存的怪异现象，重点剖析 AI 投资浪潮、地缘冲突、持续通胀以及当下市场演变与新闻节奏极度飞快的宏观特征。 ## TL;DR 节目深入剖析了美股表面强劲但底层极度撕裂的怪象：标普 500 指数逼近历史高点的同时，仅约半数股票站在 200 日线上，市场广度创下类似互联网泡沫时期的恶劣水平。当前宏观经济驱动力已从传统房地产彻底转向超大规模云厂商的 AI 资本支出（CapEx），但该模式严重依赖三年后的远期销售预期，并在信贷端积累出明斯基时刻般的脆弱性。与此同时，巨额财政赤字、昂贵的地缘冲突重整与 AI 耗能推高了通胀底线，而 0DTE 短期期权、量化 Pod Shop 和 AI Agent 的普及更让市场陷入“太快而无法做空（Too Fast to Fade）”的高频波动之中。 ## Atmosphere & style 整场对话在专业硬核与幽默风趣之间无缝切换，前同事之间的重聚让对谈显得松弛而默契。讨论既有对市场广度恶化、信贷异动与地缘通胀的严肃推演，也穿插着关于英超曼城制裁、散户炒作和末日心理学的机智调侃。听感上犹如置身华尔街交易台与学术沙龙的交界处，在密集的宏观逻辑与技术反思中，透出一种面对瞬息万变狂暴市场时的无奈、清醒与刺骨警觉。 ## Bulletin - `05:00` 美股市场广度极度恶化，标普 500 指数逼近历史高点之际，仅有 51.2% 的个股运行在 200 日均线上，类似背离在历史上仅出现于 1998 年与 2000 年互联网泡沫见顶次日。 - `10:00` 高盛数据显示，标普成分股约一半的盈利增长预期几乎完全由超大规模云厂商的资本开支拉动，罗素 1000 成分股的高估值主要取决于三年期远期销售预期，而非近期订单。 - `25:00` 2026 年市场节奏被定性为“太快而无法做空”，受 7x24 小时 AI Agent 交易、多管理人平台紧止损机制以及 4 天内到期的超短期虚值看涨期权投机潮驱动。 - `45:00` AI 已正式取代房地产成为核心宏观商业周期载体，美国家庭股票财富占比已超越房产财富占比，推升了由股市浮盈支撑消费的循环闭环。 ## Key takeaways - `00:00` 宏观信息严重过载且被过度钝化，投资者对伊朗战争局势及 AI 突破等重磅事件展现出麻木感，整体股市呈现单极化推进。 - `05:00` 当前市场并非对高利率脱敏，等权重指数与罗素 2000 深陷泥潭，大盘指数完全由对 AI 资本开支扩张的单向豪赌所托举。 - `08:00` 市场呈现出极端的“二选一”分化格局，资金不仅在半导体与软件板块间剧烈轮动，且对传统美国消费类股票形成单边抽血效应。 - `10:00` 消费龙头此前被视为软件 ARR 的避风港，但当前科技硬件的估值依赖于三年期远期销售预期，若要证明此轮基建合理，巨头未来收入需达到 GDP 的 9% 左右。 - `15:00` 全球长债收益率破 5% 的核心驱动力是全球经济与实体工业增长脉冲，但 TIPS 盈亏平衡通胀率反常稳定，折射出市场对央行抗通胀信誉的盲目假设与实际负回报风险。 - `20:00` 高增长与高通胀环境下债券沦为资产配置的被动残差项，2% 至 3% 的真实收益率无法匹敌 AI 预期回报，迫使真实收益率被动抬升。 - `22:15` AI 霸权被上升至国家安全层面的无休止军备竞赛，但政客无法公开通过金融压抑抑制民用消费，宏观政策面临尖锐的物理资源争夺悖论。 - `25:00` 市场节奏进入“太快而无法做空”的状态，Robinhood 散户单股净买入额强劲反弹，大量投机资金押注 4 天内到期的超短期限期权。 - `30:00` 自主 AI 交易智能体（Agentic Trading）的普及与算法同质化可能削减真实挂单与止损结构，加剧单日极端波动并改写微观流动性。 - `35:00` 市场广度并不足以充当趋势反转的信号，中长期唯一核心是指标层面的企业盈利松动；依赖政府赤字和企业债的 CapEx 繁荣孕育着经典的明斯基时刻风险。 - `37:30` AI 资本开支的二阶导数已经发生逆转，估值体系被迫从超高增长向 GARP 切换，且底层天量开支依然高度依赖传统互联网业务现金流输血。 - `40:00` 除了 AI 资本开支和结构性福利支出，军备竞赛、供应链重组与战争风险的极度昂贵性构成了被市场严重低估的第三大底线通胀推手。 - `45:00` AI 资本支出接替了过去房地产在财富效应和投资建工上的周期晴雨表角色，若下游无法实现切实商业化回报，整体消费与 GDP 闭环将面临断裂。 - `50:00` 高频算法与现代自然语言处理使突发新闻交易近乎无延迟完成，彻底改变了早期机器新闻阅读时代的秒级撤单与流动性骤降现象。 ## Detailed walkthrough ### `00:00` 喧嚣之下的市场麻木与矛盾开局 播客开篇，主持人 Joe Weisenthal 与 Tracy Allaway 聊起了眼下极其诡异的市场氛围。当时正值 9 月 29 日录音，Tracy 刚刚休假一周归来，笑称自己每次休假市场都会“爆雷”，而这次出事的正是债券市场。然而更深层次的荒诞感在于：眼下的金融市场仿佛处于一种“发生了太多事，却又像是什么都没发生”的状态。 从地缘政治到科技跃进，重大头条接踵而至——与伊朗的战争局势瞬息万变，投资者看到相关谈判进展的新闻后转头就翻篇；AI 公司不断推出革命性新模型，但由于技术突破的频率太高，这类重磅事件几乎被降格为背景噪音；华尔街人事变动如 David Zervos 离开 Jefferies 等消息虽然引发关注，但整体股市似乎对除了“赚钱”之外的所有宏观剧变都变得麻木或不知所措。市场虽然在标普 500 指数上录得超过 12% 的年内涨幅，但表象之下的撕裂与怪异感前所未有。为了梳理这场纷繁复杂的宏观大戏，他们邀请到了彭博前同事、现任 Sherwood News 市场主管的 Luke Kawa 展开深度漫谈。 ### `05:00` 极其糟糕的“市场广度”与单极驱动的 AI 豪赌 讨论首先切入当下技术派交易员最关注的核心痛点——市场广度（Market Breadth）。Luke Kawa 直言，当前市场的广度“烂得不可思议（god awful）”。标普 500 指数距离历史最高点仅差约 1.5%（本周初甚至仅相差 0.5%），但全市场竟然只有 51.2% 的个股站在各自的 200 日均线之上。这一数据在历史上具有极其刺眼的对照意义：上一次指数如此逼近历史高点、而站在 200 日线之上的股票比例却如此稀少的时刻，正是互联网泡沫见顶破裂的次日。虽然 1998 年也曾出现过类似恶劣的广度且随后市场又狂欢了两年，但这一统计依然引发了诸多投资者的恐慌与反弹。 这种极致的背离直接戳破了一个核心幻象：市场并没有对高利率“脱敏”。标普等权重指数（S&P 500 Equal Weight）和代表中小盘的罗素 2000 指数（Russell 2000）都在高利率环境下深陷泥潭。眼下支撑整体指数的，本质上是一场对 AI 催化力的单向豪赌——市场正在押注 AI 浪潮将像滚雪球一样吞噬一切增长空间，直到信贷收紧、借贷瓶颈或监管干预强行打断这一进程。市场甚至荒谬到了“对欧洲工业股的信心超过对处于强劲增长期的美国消费者信心”的地步，这种现象在历史上除 2017 年欧洲短暂景气之外几乎前所未见。 ```mermaid flowchart TD accTitle: 利率冲击下的市场极化分流图 accDescr: 展示美联储高利率下，强抗风险的AI巨头与受重创的传统周期与消费股的分歧走势。 subgraph Macro["美联储政策与信贷环境"] Rate["高利率与紧缩工具 (钝器打击)"] end subgraph AI_Ecosystem["AI 与超级计算巨头"] Hyperscalers["超大规模云厂商 / AI龙头"] TechSpread["信用利差保持低位<br/>自发投资 / 发债需求阶段性回落"] TechBull["驱动整体指数走高<br/>吸收全球增长押注"] Hyperscalers --> TechSpread --> TechBull end subgraph Traditional_Economy["传统经济与周期性资产"] RateSens["利率敏感型与周期股<br/>(建筑商 / 零售 / 消费)"] ConsumerWeak["股价触及52周低点<br/>估值遭受长期折价"] RateSens --> ConsumerWeak end Rate -->|"冲击较小 / 现金流充裕"| Hyperscalers Rate -->|"重创借贷与终端需求"| RateSens ``` - 标普 500 指数本周初距离历史极值仅差 0.5%，但仅有 51.2% 的成分股运行在 200 日移动平均线上。 - 相同广度水平在历史上只对应极少数极端时期，如 1998 年与 2000 年互联网泡沫顶峰次日。 - 利率作为美联储的“钝器”，对房地产建筑商和普通耐用品消费的打击远大于对 AI 超大规模企业（Hyperscalers）的冲击。 - 超大规模云服务商（Hyperscalers）在 9 月的美国投资级（USIG）债券市场上发行量甚至为零，主因是二季度财报后资本支出（CapEx）激增一度吓到股票市场，迫使其放缓了公开发债扩张的步伐。 ### `08:00` “二选一”市场、扩散效应与消费股崩塌 Luke 进一步将这种分化定义为极端分化（Dispersion）。技术分析师 Helene Meisler 将其称为“二选一市场（either-or market）”，Luke 则打趣借用《宋飞正传》的角色称其为“Elaine-George 市场”——在任何一个时刻，只能有一个人混得好。早前市场在半导体和软件股之间互相抽血，而当前的分化逻辑则演变成了“AI 扩散受益股”对阵“传统惯性股”。从多模态、Agentic Commerce（智能体商业）到底层硬件和垂直软件层，资金不断在 AI 链条内部寻找多层受益者，而链条之外的企业则惨遭抛弃。 这种资金抽血的极端受害者是广谱的美国消费类股票。从 3 月初（与中东局势升级大致同一时间节点）开始，非必需消费品以及诸多标杆零售股呈现出单边断崖式下跌。以耐克（Nike）为例，其股价从 2021 年的 180 美元天花板崩跌至当下的 35 美元左右。不论是传统运动品牌还是零售巨头沃尔玛、Costco，大量曾经稳健的消费白马都在高通胀、高利率与 AI 资产的极致虹吸下陷入低迷，展现出冰火两重天的微观图景。 ### `10:00` 软件股惩罚、三年前瞻预期与超大规模资本支出驱动 沃尔玛（Walmart）和好市多（Costco）等消费龙头在过去一段时间之所以享有极高估值，本质上是因为名义增长加速、债券收益率走高与股市展现韧性的背景下，市场迫切需要寻找标的。当软件股面临被人工智能（AI）颠覆的剧烈恐慌时，消费股凭借消费者每月稳定的消费习惯，展现出极像 SaaS 软件中 ARR（年度经常性收入）的营收特征，因而成了软件板块的“避风港”或替代品。然而，随着软件板块出现回暖反弹，这些消费股的估值倍数随即出现缩水。市场似乎遵循着一种逻辑：“在 AI 表现亮眼时，你总得找个对象来惩罚”。 关于 AI 是否真正兑现了生产力跃升、削减劳动力成本或催生新产品的盈利预期，无论是理论还是实际数据，目前支撑美股盈利增长的核心引擎依然是纯粹的“AI 硬件脉冲”。高盛（Goldman Sachs）近期报告指出： - **超大规模资本支出驱动**：今年标普成分股约一半的盈利增长预期，几乎完全由超大规模云厂商（Hyperscalers）的资本支出（CapEx）直接带动。 - **三年期远期销售预测**：高盛对罗素 1000 指数成分股估值倍数的分析显示，当期或近期的销售预期修订几乎不具备解释力，因为近期的积压订单早已被充分消化；真正决定当前高估值的核心指标，是针对“三年后”的远期销售预期。例如 Bloom Energy 以及大量光模块/光学（Optics）板块股票，凭借高企的三年期销售预期获得了极高的估值溢价。 ```mermaid flowchart TD accTitle: AI交易逻辑与传导路径 accDescr: 展示超大规模资本支出如何推高硬件盈利与三年预期，并最终需要向生产力落地交接的过程。 subgraph Hardware["第一阶段：硬件与资本支出脉冲"] A["超大规模云厂商 CapEx"] --> B["硬件/光学/能源板块收益"] B --> C["推高美股整体盈利预期"] C --> D["三年前瞻销售预期获高估值倍数"] end subgraph Transition["交接与长期兑现难题"] D --> E["英伟达财报重组：主权 AI/机器人/物理 AI"] E --> F{"能否实现广泛生产力？"} F -- 否 --> G["难以支撑天量资本回报 (如占GDP 9%)"] F -- 是 --> H["全行业劳动力节约与经济繁荣"] end ``` 这种完全依赖上调资本支出的交易模式引发了深度讨论。卖方分析师对 2023 年及未来三年的资本支出预测呈现出类似“美杜莎头发图（Hair charts）”般层层上修的轨迹，整个 AI 交易中有相当庞大的一部分只是在押注“资本开支预期会不断被调高”。英伟达（NVIDIA）在前几个季度的财报结构调整中，已开始强调主权 AI、机器人技术以及本地部署（on-prem）的物理 AI，试图将硬件需求从超大规模云厂商和新型云服务商（Neoclouds）向外拓展。经济学家 Hanno Lustig 曾提出估算：为了证明当前预期的这一轮人类历史上最大规模的基础设施繁荣具备合理性，少数几家巨头未来创造的收入甚至需要达到 GDP 的 9% 左右，这一要求堪称极其激进。 --- ### `15:00` 全球与美债收益率飙升背后的经济脉冲与通胀预期之谜 美国 10 年期国债收益率重新突破 5%，30 年期长债收益率同步攀升。这一轮美债乃至全球主权债券收益率的暴涨，如果用主成分分析法（PCA）去提炼最核心的单一驱动因素，结论非常纯粹——那就是“全球经济增长脉冲”。无论是在美国还是全球范围内，工业制造与实体活动表现极其强劲，且这种强劲预期还在持续发酵。正如市场流行的一句话所言，“凡是能被重新改造用于 AI 的东西，都将被全力改造”；庞大的资本投入与经济脉冲正在惠及整个产业链，推升名义经济增速与通胀压力，进而迫使全球央行维持强硬紧缩立场。 令人困惑的反常现象在于，即便名义收益率急速走高，美国通胀保值债券（TIPS）所隐含的盈亏平衡通胀率（Breakevens）却表现得异常稳定： - **央行信誉假设**：主流观点通常认为，盈亏平衡通胀率保持稳定反映了市场信任美联储捍卫双重使命的“央行信誉”，押注其最终能够有效抑制长期通胀。 - **实际回报率的警示**：如果从历史持有体验来看，投资者在 2011 年至 2021 年期间的任何一个时间点（除页岩油破产危机前夕的极短暂窗口外）买入并持有到期 5 年期美国国债，其扣除通胀后的实际收益率（Real returns）全为负值。因此，市场在名义利率剧震之际依然笃定长期通胀预期纹丝不动，构成了当前宏观交易中最耐人寻味的一大矛盾。 ### `20:00` 真实收益率上升与跨资产配置的“AI 债券”难题 当被问及在美联储多年未达成通胀目标后，市场为何依然没有对美联储维持通胀稳定的能力失去信心、且通胀预期未在图表上失控飙升时，Luke Kawa 坦言自己对此“确实感到由衷意外”。回顾 2022 年至 2023 年在高通胀环境下的资产配置讨论，债券在很大程度上沦为了资产负债表上的“残差（residual）”：只有当投资者担忧经济增长且预期通胀不高时，才会大规模买入债券；一旦经济增长强劲且通胀高企，资金就会流向现金、短期工具、大宗商品和股票。 ```mermaid flowchart TD accTitle: 高通胀与 AI 繁荣下的跨资产配置决策流程 accDescr: 展示在增长强劲且 AI 收益可观的背景下，债券如何沦为资产配置的残差，以及真实收益率被迫走高的逻辑机制 subgraph Macro["宏观环境评估"] A["经济增长稳固 + 通胀高企"] end subgraph Assets["资金优先配置路径"] B["股票 & 大宗商品"] C["现金 / 超短期资产"] D["AI 基础设施赛道"] end subgraph BondDilemma["债券资产定位"] E["债券沦为『残差』项"] F["2%~3% 真实回报难敌 AI 吸引力"] G["名义收益率走高 / 真实收益率被迫抬升"] end A --> B A --> C A --> D B & C & D --> E E --> F F --> G ``` 当前的收益率上行主要是由真实收益率（real yields）驱动的。跨资产投资者普遍认为股票收益代表的是一种真实的资产流，如果要让债券相对于股票更具吸引力，就必须抬高其真实收益率。在 AI 浪潮提供惊人预期回报的背景下，投资者根本无法说服自己去买入仅有 2% 甚至 3% 真实回报率的债券。面对这种资本抽离，另一个极端的选择是走向“金融压抑（financial repression）”，即效仿战时国债（War Bonds）发起全国性号召，要求民众削减日常开支、将资金投入“AI 债券”，以确保算力建设。 - **残差资产**：在经济稳固与通胀并存时，债券沦为最终才做决策的被动配置标的。 - **真实收益率比较**：投资者往往将股票盈利收益率（Earnings Yield）与真实收益率（而非单纯的名义收益率）对比，以此权衡是否持有固收资产。 - **极端路径猜想**：通过推行限制消费的金融压抑政策，强制动员资本购买 AI 债券以支撑高昂的基建与数据中心电力需求。 --- ### `22:15` AI 霸权、国家安全与宏观政策的“零和博弈” 主持人和嘉宾就 AI 在国家安全层面的地位展开了更深层的政治经济学探讨。当前华盛顿广泛认为“AI 霸权”事关生死存亡的国家安全。如果这一逻辑被贯彻到底，政策制定者理应主动抑制非 AI 领域的经济活动（例如对私人游艇等一切可被拆解为数据中心零部件的非必要消费品加征重税），以腾出真实的物理资源支持基建。然而，现实中没有政客敢公开对选民说“为了 AI 建设，你暂时不能买房”，这种资源挤占必然会引发严重的民间不满。 与冷战时期的核竞赛不同，核力量在拥有原子弹后便进入了相对稳固的“核大国俱乐部”状态；但 AI 算力与模型能力的竞赛是永无休止的，没有任何一个节点能让人安然宣称“我们已经全面具备能力”。Luke Kawa 感叹，回看全球金融危机（GFC）之后的整个周期，政策制定者完全没有意识到那时的日子有多轻松——当时的宏观经济几乎不存在严重的硬性权衡取舍（trade-offs）。曾经被学界奉为圭臬的“央行只要宣称通胀目标是 2%，通胀就会锚定在 2%”的前瞻指引理论，在当下的硬现实面前已被证明彻底沦为空谈。 - **资源争夺悖论**：AI 建设需要大规模的电力、土地和计算资源，与普通民众的生活资源（如住房、电费、生活环境安宁）形成直接竞争。 - **无止境的军备竞赛**：AI 技术不存在稳定的均衡状态，资本支出扩张周期被无限拉长。 - **前瞻指引神话破灭**：过去依赖口头预期的央行信誉在通胀持续冲击下遭受质疑，市场不再盲目听信利率指引。 --- ### `25:00` “太快而无法做空”：2026 年的速度压缩与零售投机潮 在为当下的市场定调时，对比上一轮零利率时期的“TINA（别无选择，There Is No Alternative）”，Luke Kawa 给出的核心词是“太快而无法做空（Too Fast to Fade）”——资产价格与盈利预测的变化节奏已经快到让逆势交易者无法从容反向操作。2026 年被这种极致的“速度”与“时间压缩”彻底定义：半导体与软件板块之间的月度相对轮动经历了一年内最为剧烈的四次回摆，白银等大宗商品也出现了罕见的暴力拉升，这种剧烈波动的背后是能够全天候自主运转的“AI Agent 代理交易”以及技术的颠覆性破坏力。 ```mermaid flowchart LR accTitle: 2026年市场极速运转的动因构成 accDescr: 展示技术压缩时间、期权工具、散户流向以及微观市场结构如何共同导致资产价格快速变动 subgraph TechDrivers["技术与感知层面"] T1["头条事件爆发频率提升"] T2["AI Agent 7x24小时全天候运转"] end subgraph RetailFlows["散户资金与交易品种"] R1["单股净买入额高位反弹"] R2["0DTE/超短期限虚值看涨期权"] end subgraph MarketStructure["量化微观结构"] M1["Pod Shop 紧止损模型"] M2["多空离散度交易 Dispersion Trade"] end TechDrivers --> FastMarket["『太快而无法做空』的极端市场"] RetailFlows --> FastMarket MarketStructure --> FastMarket ``` 从 Robinhood 的散户交易数据来看，单股滚动 21 天净买入额在 6 月底（美光财报前后）见顶，此前的一个峰值则出现在 2025 年 10 月至 11 月初，该时期对应着投机性资产（如比特币与 Palantir）的联合筑顶。尽管 9 月传统上是散户淡季，但当前单股购买量已强劲反弹至 6 月峰值的 89% 左右。散户资金不仅扎堆在美光（Micron）、SpaceX 等知名标的上，甚至前 25 大热门交易名单中频繁出现连资深市场人士都未曾见过的生僻代码。投资者疯狂涌向距离到期仅剩 4 天以内的超短期看涨期权，通过极致的杠杆和博取暴利的心态推动市场以极高频率反转。 - **核心定性**：Too fast to fade（速度太快，无法做空/反向对冲）。 - **投机资产历史周期**：2025 年 10 月比特币与 Palantir 同步见顶，2026 年 6 月探底后再度与期权投机潮形成回升共振。 - **散户交易数据**：9 月 Robinhood 散户单股净购买量恢复至 6 月历史高点的 89%，4 天内到期的短期期权占据主导地位。 - **市场结构催化剂**：多管理人平台（Pod Shops）的扩大、极其严格的日内止损机制、针对特定个股的离散度交易（Dispersion Trade）以及动量倾向（Momo）鲜明的自主 Agent 算法，共同放大了价格与新闻事件共振后的剧烈摆动。 ### `30:00` 单只股票波动与 AI 交易智能体对市场结构的潜在冲击 当前市场呈现出一种指数层面波澜不惊、但个股层面剧烈波动的分化状态。就像一只浮在水面上的鸭子，水面上看起来纹丝不动，水面下的双脚却在拼命划水（"duck with its feet underwater going crazy"）。哪怕整个大盘在数月内完全没有明确方向，市场里永远不缺惊心动魄的话题。 主持人谈到了自己虽然不做实盘交易，但一直想动手尝试的一个项目：利用 AI 技术构建一个代理式自主交易机器人（Agentic Trading Bot），在类似 Kalshi 等预测或交易平台上进行模拟纸面交易（Paper Trade）。其思路是让 AI 阅读大量学术论文，挖掘人们猜测可能存在错误定价（Mispricing）的领域，然后搭建一套无懈可击的系统化交易模型去自动执行。 Luke Kawa 表示，目前确实看到许多人正全身心投入或至少在尝试搭建此类系统。他认为这在未来对市场结构会带来非常耐人寻味的影响： - 如果越来越多的交易决策由智能代理按既定规则接管，市场上的真实挂单和止损单（Stop Orders）会不会减少？ - 这类机制对市场的透明度和深度（Market Visibility）究竟意味着什么？ - 智能体算法的同质化和自动化执行，极有可能进一步加剧单日极端波动（Single-day Volatility）。近年来市场结构的演进方向，似乎都在不知不觉中推高了单日的瞬时波动烈度。 ```mermaid flowchart LR accTitle: Agentic Trading and Market Volatility accDescr: 流程图展示了从学术论文挖掘定价偏差到AI自动化交易执行，进而影响市场挂单结构并加剧单日波动的连锁反应。 subgraph Strategy ["策略生成与执行"] A["研读学术研究<br/>寻找定价偏差"] --> B["搭建 Agent 智能体<br/>系统化交易规则"] B --> C["自动化接管交易<br/>算法快速响应"] end subgraph Impact ["市场结构与波动"] C --> D["传统挂单与止损结构改变<br/>市场可见度变化"] D --> E["单日极端波动加剧<br/>流动性瞬间抽干"] end ``` --- ### `35:00` 泡沫警报、真正的转向指标与明斯基时刻的阴影 面对顶着利率高企、地缘政治冲突（例如与伊朗的战争风险）依然不断上攻的美股，许多人担忧市场已步入泡沫化阶段。被问及什么样的警示信号才意味着股市会出现“持久性的趋势逆转”时，Luke Kawa 直言，市场广度（Breadth）并不是一个充分或必要的反转指标，因为广度往往会给出假信号，就像 1998 年发生过的那样。 作为一名秉持“基本面至上”的无聊投资者，Luke 强调驱动股市中长期的唯一核心永远是企业盈利（Earnings）： - 尽管分析师的预期调整向来严重滞后，投资人若等到标普 500 远期每股收益（Forward EPS）自 52 周高点开始下修时才离场，在过去 6 到 7 次熊市中，除了 2020 年因封城导致暴跌速度极快的新冠危机让人猝不及防之外，其余每次即使晚到离场，只要此前持有两年，最终斩获的回报依然相当可观。 - 因此，投资者完全可以耐心等待指数层面的企业盈利真正出现松动迹象，再做出撤退决策。 更值得警惕的断裂点在于信贷市场（Credit Market）的金融工程把戏。Luke 指出，此前市场曾有一个坚定共识：超大规模云厂商（Hyperscalers）的高昂资本开支（CapEx）如果无法收回，要么是 AI 热潮降温，要么就必须诉诸更加激进、更具创意的债务融资模式。如今，“融资方式确实变得越来越有创意了”。从企业利润来源看，当前支撑美股繁荣的一大支柱是超大规模云厂商的资本开支（且越来越多通过发行债务融资），另一大支柱则是巨额的美国政府财政赤字： > “你能想到比超大规模云厂商的企业债和政府财政赤字更安全的利润来源吗？这简直安全得不能再安全了——直到它突然暴雷为止，这是最经典的明斯基时刻（Minskyan formula）。” --- ### `37:30` AI 资本开支的二阶导拐点与硬件供应链分化 对于那些处于超大规模云厂商资本开支接收端的供应商来说，这笔持续流入的巨资一度被市场视作犹如“债券票息”般稳定可靠的结构性收入。但信用市场的焦虑感已经在部分 AI 深度参与者中蔓延开来（例如针对 CoreWeave 的信用违约互换 CDS 的定价讨论）： - 市场的担忧往往直接映射在杠杆率和信用评级上。例如在主流超大规模厂商中，甲骨文（Oracle）因其相对同业显著偏高的债务负担与略逊一筹的信用资质，在信贷市场中格外扎眼。 - 整个 AI 赛道已经告别了“凡沾 AI 必涨”的无差别狂欢期。盈利预期上调的速度（即二阶导数，Second Derivative）已经发生逆转，市场不再无脑计入史上最快的盈利上调预期。 - 股票估值体系正被迫经历从“超高增长（Hyper-growth）”向“合理价格增长（GARP, Growth at a Reasonable Price）”的艰难换挡，在经历三季度动能股的剧烈重构后，许多股票仍在苦苦挣扎。 - 半导体板块内部的分化已经极其惨烈：在标普 500 指数中，整个第三季度创下 52 周新高的芯片公司仅有两家——AMD 和思佳讯（Skyworks Solutions）。 - 正如《华尔街日报》专栏作家 Greg Ip 所言，本轮 AI 热潮的一个极其怪异之处在于：支持这场天量开销的资金，并不是由 AI 自身的商业化果实来支撑的，而是由传统互联网业务的“超级印钞机”在输血——例如 Meta 极其暴利的广告业务和谷歌无可匹敌的搜索业务。没有这些原有现金流的垫付，单靠 AI 目前微薄的回报根本无法维持如今的烧钱速度。 ### `40:00` 巨头资本支出、足球八卦与被低估的“战争通胀”冲动 讨论由甲骨文（Oracle）在超大规模云计算（Hyperscaler）竞赛中的地位展开。主持人指出，尽管 Oracle 是一家非常成功且优秀的企业，但它并不具备其他超大规模云厂商那样纯粹的“资金水龙头”（money spigot）。在当下的信贷市场上，人们持续观察到 Oracle 的信用违约互换（CDS）利差出现飙升。 紧接着，金融终端上跳出了一条突发新闻：曼城足球俱乐部因涉嫌虚增收入及削减 9 亿英镑成本而面临英超联赛的重罚与制裁。话题随即展开了一段轻松的体育插曲： - 作为曼联球迷的 Luke Kawa 直呼这是“因果报应”（FAFO，即 *Fuck Around and Find Out*），戏称只要曼联能顺带递补拿到冠军，他甚至愿意容忍阿森纳追溯获得奖杯。 - 主持人随后聊起 Joe 曾撰文将英超联赛视为“欧洲牛市的做多理由”，并提到朋友 Guan 送了一件深圳新鹏城足球俱乐部（Shenzhen Peng City FC）的浅蓝球衣——这支俱乐部与曼城同属城市足球集团（City Football Group）网络。 - 话题还延展到美国观众对足球的“三年热度”现象，主持人笑称自己是周末早晨带娃看球的“小众文青球迷（hipster fan）”，并调侃如果阿联酋资本未来不能继续在英国肆意挥霍，中国足球是否会迎来巨资注入。Luke 则幽默表示自己原本期待将话题转向冰球。 回到严肃的市场宏观层面，麦格理（Macquarie）分析师 Terry Wiseman 的一份研报引发了深度共鸣。人们在讨论当下的通胀与财政冲动时，注意力大多集中在两个巨头因素上：**AI 资本开支**与**政府结构性福利支出**。然而，Wiseman 提醒市场严重低估了第三大驱动力——**战争与地缘重组的极度昂贵性**。 当下市场在交易中东局势及伊朗战争风险时，视角极为狭隘，往往仅关注油价波动、霍尔木兹海峡何时封锁或柴油裂解价差。但现实是全球范围内的国防军费连年攀升、战备重整成本极其高昂，再加上半导体供应链的本土化回流重构，无一不需要天文数字的资金投入。当今时代的一个核心特征，便是社会在何种程度上愿意容忍那些“必然推高底线通胀（bottom-line inflationary）”政策所带来的副作用。 此外，Luke 还指出当前市场另一处令人费解的定价扭曲（borderline psycho）：在美股名义 GDP 增速高达 6% 左右、单季个人消费预计拉动 2.8 个百分点增长的环境下，标普 500 等权重指数距离历史高点的距离居然落后于欧洲等权重指数，美国非必需消费品（Consumer Discretionary）板块甚至成了市场的“出气筒”。 ```mermaid flowchart TD accTitle: 驱动宏观通胀冲动的三大支柱 accDescr: 结构性政府福利支出、AI基础设施资本开支，以及昂贵的地缘冲突与重整开支，共同推高通胀底线。 subgraph MacroImpulse["宏观资本消耗与通胀冲动"] A["AI 基础设施 CapEx 竞赛"] B["政府结构性福利刚性支出"] C["地缘冲突与全球再武装"] end C --> C1["伊朗局势/能源供应链摩擦"] C --> C2["国防军费常态化增加"] C --> C3["半导体产业回流本土建厂"] A --> D["极度高昂的底线通胀压力"] B --> D C1 --> D C2 --> D C3 --> D ``` --- ### `45:00` 经济周期的迁移：从房地产到 AI，以及“世界末日”心理 讨论进一步深入到美国宏观经济驱动力的本质转换。在以往的经济周期中，房地产市场始终是经济的核心晴雨表——房产增值带来房屋净值信贷（HELOC）和财富效应，继而刺激耐用品消费。然而如今，即便住房市场深陷冰冷惨淡的泥淖，整体经济却并未崩溃。 其核心原因在于：**AI 已经取代房地产，成为全新的经济商业周期载体**。 今年早些时候，美国家庭源自股票的财富占比正式超越了源自房产的财富占比，这是一个具有里程碑意义的交叉点： - **财富效应与建设冲动的完全转移**：过去由房地产带动的建工投资和耐用品采购，如今被数额更大、超越以往任何时期的 AI 数据中心与基建 CapEx 所接管；中产与高净值人群则通过股市浮盈来维持消费。 - **永动机模式与循环风险**：AI 资本支出推高相关企业股价，财富效应支撑强劲消费，消费进而维系整体 GDP 增长并进一步推动股票上涨。但这种近乎“永动机（perpetual motion machine）”的闭环存在致命漏洞——企业最终需要盈利，消费者最终需要有实际需求和购买力。如果链条下游无法带来切实的商业回报与消费者变现，整个庞大的 CapEx 繁荣将失去支撑。 围绕这种极端预期，Luke 与主持人探讨了一种正在蔓延的社会心理：AI 究竟会带来奇点般的乌托邦（实际财富激增、物品极度廉价）、颠覆性的劳动力大洗牌，还是引发生存浩劫（如模型失控诱发核灾难）？ 这种非生即死的技术末日论甚至开始反向侵蚀储蓄观： - 传统逻辑是“担忧核战争就更该买股票，因为世界毁灭了钱也没意义（if it's in the price）”； - 如今的反向思维则是“如果十年内世界注定走向毁灭，为什么现在还要存钱？” - Luke 戏谑地引用自己曾发过的推文总结这一悖论：“如果 AI 发射核武器毁灭了全人类，我们的投资组合将付之一炬（on fire），你还会后悔当初没去吃顿大餐；而如果 AI 没有毁灭我们，我们的投资组合同样会火热爆表（on fire）。” 节目在轻松的告别与复盘中收尾。主持人与 Luke 达成共识，当前市场的本质特征不仅是底层结构正在剧变，更是**叙事与新闻流切换速度达到了不可思议的极致**。交易者即使不相信周二跳出来的某条正面头条，也深知市场算法会对其作出剧烈反应，所有人只能在这密集的红绿头条信息流中不断来回拉扯。 ### `50:00` 机器新闻阅读的演进与市场流动性瞬时变化 在对新闻与市场反应节奏的深入探讨接近尾声时，主持人和嘉宾关注到了重大新闻突发瞬间的微观结构变化。讨论指出，去研究突发新闻发生频率与反应节奏的历史趋势将会非常有趣。在早期的“机器新闻阅读交易”（early machine reading trading——尽管如今已很少有人再使用这个称呼）时代，每当市场上出现突发重磅头条时，许多交易员和算法实际上会选择瞬间退避并撤离报价。其原因在于，当时的机器算法通常需要大约一秒钟的时间来抓取、阅读新闻并判断该如何做出反应，这导致头条新闻公布的瞬间市场往往会出现短暂的流动性骤降。 如今，随着技术和自然语言处理能力的极大跃升，算法几乎不再需要这种迟滞时间，但大家依然十分好奇现代市场在突发重大头条瞬间的流动性画像究竟变成了什么样，以及是否有人专门针对这种毫秒级的流动性演变做过学术或实证研究。谈及对高频信息冲击和市场波动的担忧时，大家打趣地自嘲“我担心的事情……我担心的事情可多着呢”，随后以轻松的幽默感顺理成章地结束了对宏观速度与新闻节奏的深度探讨。 ```mermaid flowchart TD accTitle: 机器阅读新闻交易模式的历史演变 accDescr: 对比早期机器阅读交易与现代算法在突发新闻下的流动性与延迟表现 subgraph Early["早期机器阅读交易模式"] E1["重磅头条新闻发布"] --> E2["算法耗费秒级时间解析文本"] E2 --> E3["做市商/机构短暂撤单回避"] E3 --> E4["盘口出现短暂的流动性骤降"] end subgraph Modern["现代高频与算法环境"] M1["重磅头条新闻发布"] --> M2["高频自然语言处理近乎即时定价"] M2 --> M3["极速成交与快速价差重构"] end ``` - **早期机器阅读的延迟特征**：早期文本抓取与情感分析算法响应需要大约 1 秒，做市机构在此窗口内通常主动撤离，造成微观结构上的流动性凹陷。 - **现代市场的未知领域**：当代量化系统处理非结构化新闻的速度已呈指数级提升，重大头条出现瞬间的流动性深度与价差变化仍缺乏公开系统性的实证研究。 ### `50:20` 节目收尾、社交账号与社群信息 随着讨论告一段落，主持人 Tracy Alloway 与 Joe Weisenthal 宣布本期《Odd Lots》正式收官，并向听众分享了制作团队与嘉宾的联系方式。他们呼吁大家在各大社交平台上保持关注，追踪关于快节奏宏观环境与市场叙事的最新动态，并提醒听众可以前往彭博官网获取往期播客音频、专属特稿与更新。 节目同时向听众推荐了活跃的 Discord 交流社区，欢迎所有市场爱好者在全天候的社群中讨论每期节目的核心主题。主持人呼吁喜爱此类宏观市场专访的受众在各大播客平台留下五星好评，并特意提及彭博订阅用户的专属权益——只要在 Apple Podcasts 上找到 Bloomberg 官方频道并完成认证绑定，即可享受无广告打扰的纯净收听体验。 - **主持人社交账号**：Tracy Alloway（`@TracyAlloway`）、Joe Weisenthal（`@TheStalwart`）。 - **制作团队社交账号**：制作人 Carmen Rodriguez（`@CarmenArmon`）、Dash Bennett（`@Dashbot`）、Kale Brooks（`@kalebrooks`）、Kevin Lozano（`@kevinllozanno`）。 - **特邀嘉宾社交账号**：Sherwood News 市场主管 Luke Kawa（`@LJKawa`）。 - **社群与订阅渠道**：彭博官方主页（`Bloomberg.com/oddlots`）、24/7 Discord 交流社群（`discord.gg/oddlots`），以及 Apple Podcasts 彭博无广告订阅频道。 ### `50:50` 商业赞助与服务推广 节目尾声播放了多家商业伙伴的品牌赞助信息，涵盖企业保险、通信网络、商业信用卡与定制礼品服务等领域： - **辛辛那提保险公司（Cincinnati Insurance）**：强调在瞬息万变的金融与商业世界中，面对无法预测的“黑天鹅”突发事件比追逐下一个风口更为重要。该公司与独立代理人合作超过 75 年，致力于为企业运营、房产及贵重资产提供防范风险的坚实后盾（官方平台 `CINFIN.com`）。 - **康卡斯特商业（Comcast Business）**：展示其为全球 100 多个国家提供的 AI 驱动安全网络服务。该业务不仅覆盖了超过 90% 的《财富》500 强企业和数百万小微商家，还配备了强大的全天候专业团队，即便在凌晨两点也能提供如日间一般的高水准即时支持。 - **达美航空天合优享 Reserve 商业运通卡（Delta SkyMiles Reserve Business American Express Card）**：鼓励企业主将日常商业开支转化为飞行里程，助力商务出差或休假出行。限时优惠期间新申领用户最高可获得 20 万点奖励里程（该卡年度最高开卡奖励，申请截止日期为 11 月 4 日，适用最低消费额度与相关条款；申请入口 `www.delta.com/podcast`）。 - **4imprint 品牌定制礼品**：主打不仅是“完成送礼清单”，更是传递品牌价值与真诚谢意的高品质定制解决方案。提供包括高级服饰、饮水具、数码配件等成千上万种客制化产品，并提供专业的售前支持与 360 度满意保证（服务网址 `4imprint.com`）。 ## Notable quotes > `05:00` 烂得不可思议（god awful） — Luke Kawa > `08:00` 二选一市场（either-or market） — Helene Meisler > `10:00` 在 AI 表现亮眼时，你总得找个对象来惩罚 > `15:00` 凡是能被重新改造用于 AI 的东西，都将被全力改造 > `25:00` 太快而无法做空（Too Fast to Fade） — Luke Kawa > `30:00` duck with its feet underwater going crazy > `35:00` 你能想到比超大规模云厂商的企业债和政府财政赤字更安全的利润来源吗？这简直安全得不能再安全了——直到它突然暴雷为止，这是最经典的明斯基时刻（Minskyan formula）。 — Luke Kawa > `45:00` 如果 AI 发射核武器毁灭了全人类，我们的投资组合将付之一炬（on fire），你还会后悔当初没去吃顿大餐；而如果 AI 没有毁灭我们，我们的投资组合同样会火热爆表（on fire）。 — Luke Kawa > `50:00` 我担心的事情……我担心的事情可多着呢 ## Mentioned Joe Weisenthal, Tracy Allaway, Luke Kawa, David Zervos, Helene Meisler, Hanno Lustig, Greg Ip, Terry Wiseman, Carmen Rodriguez, Dash Bennett, Kale Brooks, Kevin Lozano, Jefferies, Sherwood News, Bloomberg, Goldman Sachs, Macquarie, Robinhood, NVIDIA, Meta, Google, Oracle, AMD, Skyworks Solutions, Bloom Energy, Micron, SpaceX, Palantir, CoreWeave, Nike, Walmart, Costco, Manchester City FC, Shenzhen Peng City FC, City Football Group, Kalshi, Cincinnati Insurance, Comcast Business, Delta Air Lines, American Express, 4imprint ## Video description <p>The market is acting in ways that make it very hard to get a handle on what exactly is going on right now. The overall indices are surging, while many individual stocks are doing badly. Rates are rising, but the economy is still robust. There's capex spending and the AI trade. There's the war in Iran. And of course there's persistent inflation and the Federal Reserve's ongoing tightening cycle. On this episode, we speak with our former Bloomberg News colleague Luke Kawa, who is now head of markets at Sherwood News. We go over all the big market themes, and he explains why the one consistent theme of this moment is the sheer speed of the moves and the pace of the news itself.<br><br><a href="https://events.bloombergevents.com/event/OddLotsLiveChicago/summary">See Odd Lots Live in Chicago!</a></p><p>See <a href="https://omnystudio.com/listener">omnystudio.com/listener</a> for privacy information.</p> [Source](https://podtrac.com/pts/redirect.mp3/tracking.swap.fm/track/UVBrz8bN8aM2Xe47PEPu/traffic.omny.fm/d/clips/e73c998e-6e60-432f-8610-ae210140c5b1/8a94442e-5a74-4fa2-8b8d-ae27003a8d6b/cf3994af-6383-4b79-a266-b4d50143f150/audio.mp3?in_playlist=982f5071-765c-403d-969d-ae27003a8d83)

[5:00] Luke Kawa: 烂得不可思议（god awful） [8:00] Helene Meisler: 二选一市场（either-or market） [10:00] 在 AI 表现亮眼时，你总得找个对象来惩罚 [15:00] 凡是能被重新改造用于 AI 的东西，都将被全力改造 [25:00] Luke Kawa: 太快而无法做空（Too Fast to Fade） [30:00] duck with its feet underwater going crazy [35:00] Luke Kawa: 你能想到比超大规模云厂商的企业债和政府财政赤字更安全的利润来源吗？这简直安全得不能再安全了——直到它突然暴雷为止，这是最经典的明斯基时刻（Minskyan formula）。 [45:00] Luke Kawa: 如果 AI 发射核武器毁灭了全人类，我们的投资组合将付之一炬（on fire），你还会后悔当初没去吃顿大餐；而如果 AI 没有毁灭我们，我们的投资组合同样会火热爆表（on fire）。 [50:00] 我担心的事情……我担心的事情可多着呢

[1:45] Joe Weisenthal: so much is happening and nothing is happening. [1:45] Joe Weisenthal: trying to make money. [4:59] Luke Kawa: absolutely god-awful. [6:20] Luke Kawa: AI impulse will continue to eat and swallow the world. [6:20] Luke Kawa: It's this idea that only one can do well at the right time... At first, it was semis versus software. More recently, it's been inertia stocks versus the general AI beneficiaries. [12:35] Luke Kawa: ambitious, ambitious, I'll put it that way. [25:00] Luke Kawa: This is... too fast to fade, both prices and earnings just moving too fast to fade. And speed is something that I think has completely defined 2026. [30:00] Luke Kawa: the classic duck with its feet underwater going crazy, [30:00] Luke Kawa: exacerbate single day volatility. [36:30] Luke Kawa: second derivative has turned [45:00] Luke Kawa: If AI kills us all by setting off a nuclear weapon, our portfolios will be on fire. And you'll feel stupid for not having gone out to eat. If AI doesn't kill us all... then our portfolios will also be on fire. [50:00] I worry about plenty of things

[0:00] There are some market stories where you want every detail. Joe and I have made quite a few podcasts on that basis, but sometimes you've only got 10 minutes and just want to know what's moving markets fast. That's the Barclays Brief podcast. Every week, experts from Barclays Markets and Research get you up to speed on what's happening and what to watch next. Concise, focused, and brief. Search Barclays Brief wherever you get your podcasts. Some people treat ChatGPT like some kind Thank you. projects. Get started at ChachiBT.com by selecting work mode available on plus and pro plans. Never bet against American grit or American energy. Through innovation, Venture Global is not only building some of the largest energy facilities in the world right here in the United States, but delivering American energy at a fraction of the cost and a fraction of the time. That's venture global. That's unstoppable energy. Hey, Odd Lots listeners, the Odd Lots tour continues and our next stop is in Chicago. That's right. Joe and I will be at the City Winery Chicago on October 15th for a live Odd Lots recording. Tickets are on sale now at Bloomberg.com forward slash Odd Lots. And of course, a special thank you to Barclays for supporting Odd Lots Live. So that's October 15th at City Winery in Chicago. Get your tickets now. Bloomberg Audio Studios. Podcasts. Radio. News. Hello and welcome to another episode of the Odd Laws podcast. I'm Joe Weisenthal. And I'm Tracy Allaway. - Tracy, it dawned on me the other day... The market's kind of crazy right now. No, no, it's like, seriously, it's like we're sort of used to all this stuff and everything. And then one day, one moment... I realized markets are really crazy these days. I would say weird. I find them very weird at the moment with like the biggest tension being what's happening in the bond market and the stock market. By the way, I have a huge caveat for this episode, which is we're recording this on September 29th. I've been away for the past week on vacation. I probably should have put up my normal going on vacation warning so something is about to blow up. Turns out it was the bond market. Yeah. Actually, it's funny you say that because the way I think about the market right now is so much is happening and nothing is happening. In the specific sense that it's like, okay, we're in the middle of a war. with Iran. And so many, we're just like, oh, it's just this headline. Trump says negotiations are going good. And then it's like people move on. There's some crazy model leap, et cetera, by one of the AI companies. Huge news, but also so regularly cadence now that even that almost becomes noise. Obviously, the movement's in the bond market, I guess, like, That is probably the one thing that does feel... novel about maybe just the last few weeks but even there a lot of that angst has been going around a long time so it's like There's so much of everything. There's nothing. There's no like one thing that just, you know, I was thinking like yesterday, September 28th, it was announced that David Zervos was going from Jeffrey. I saw that. I was in Connecticut with no electricity. And even I saw that. related story of the day. It feels like stock markets are sort of overwhelmed and or inured to everything happening except trying to make money. No, like that is the one impulse that remains. That never gets old. No. And that's arguably the point of the whole thing. Anyway, we need like a sort of big rambling state of the markets episode. And we really do have the perfect guest, someone who used to be our colleague here at Bloomberg. We are, of course, speaking to Luke Kawa, head of Luke, thank you for coming back on Oblots. - Ah, great to be here. Thanks for having me, El. - Let me start on something niche, actually, 'cause there's so much going on. Suddenly, lately, markets having a great year, S&P up, [5:00] or over 12% or about that as the time I'm talking. I see a lot of talk about breath. This is something everyone's talking about market breadth these days. What's that all about? Yeah, I think why they're talking about breadth is because it's absolutely god awful. There's been a lot of different stats to show it. You can say the average stat is XYZ below its high, while the S&P 500 is, you know, as of writing about, you know, 1.5% off of its high. The one I came up with that bothered a lot of people the other day was, coming into this week, the S&P 500 was 0.5% off record high. And just 51.2% of stocks were above their 200-day. The last time we had that proximity to a high with that few stocks above the 200-day was literally the day after the dot-com bubble peak. So that gets a lot of people mad. But then you have to point out, okay, wait. Breath was also that bad in 1998. The market ripped for another two years after that. So it really is a case of, and I think to unify a lot of what you were talking about in the intro, is the market inured to higher rates now? No! A lot of the market isn't. The S&P 500 equal weight certainly isn't. The Russell 2000 certainly isn't. However, basically, to me, there seems to be just this big bet that the AI impulse will continue to eat and swallow the world until who knows what puts a stop to it, whether that's kind of credit conditions, the capacity to lend, whether that's regulation. But right now, the pricing to me just looks like more and more betting on a singular factor to be able to drive growth. to be more confident in European industrials than the U.S. consumer during a period of strong growth. I cannot ever remember that happening outside of maybe like Euro boom, what was it, 2017 for a couple months. I mean, it is true that in higher rates, so... Rates are a blunt tool for the Fed, right? And they're going to hurt something like a home builder a lot more than an AI company that is issuing like mega deals into the bond market, I would imagine. Is it the case that, you know, maybe the bet on AI makes sense in that environment? Maybe the AI companies are the ones that are less rate sensitive than a lot of other businesses out there. explosion higher in yields. And we've talked about like, oh, it's the AI impulse this, that. It's also coming to a time where hyperscaler credit spreads, none of them are at their wides of the year at the long end. Yeah. It comes after September of you can tell me how much USIG... and USIG issuance there was by hyperscalers. I believe that's a zero for September. So it's, yeah, so very, very interesting. - They cut back a lot because they got a little bit spooked with the CapEx response from the market. - So that to me is something that like, hey, a lot of this in the bond market, both a response to stronger growth and the potential for issuance, a lot of this is like the classic Mark Dow belated overreaction to something we've been thinking and talking about for a long while. And then I think the question is, How long can the cyclicals that are doing well, you know, operationally at least, how long can you maybe plumbing 52-week lows, a lot of these? And is it just because AI is doing so well? And I think that's the kind of important part to be making, the important point to be making, that so much of this market has been dispersion. Helene Meisler calls it the either-or market. I've called it, you know, the Elaine George market, Jerry's friends, only one can do well at the right time. of 2026. At first, it was semi-software. I think more recently, it's been like this idea of inertia stocks versus the general AI beneficiaries. And that to me is a bit of an important twist. It's this idea that like, It's not just one layer of AI that benefits. It's multiple layers through software, through the hardware providers that get to benefit from an impulse like Muse, like Agentic Commerce, and then others that get hit. So that to me is at least a bit of a nuance in how the market treats AI and the impulse. But for me, it's just really a sign of the mechanics that have continued to dominate and just continuing to play out. on March 3rd, and then it's been straight down. So it's like basically when the Iran war started, Wait, was that when the CEO ate? Oh, I forgot about that. But it's just like any consumer name you can think of right now. The stock is awful. Nike is the classic example. That was crazy. That was $180 stock in 2021. It's $35 now. I mean, granted, it had a huge run-up in 2021, too. It's crazy. And when you look at Nike, one of the things that stands out, and I think you could apply this more recently to a Walmart, possibly to a Costco, [10:00] And I'll tie this back to software, oddly enough, but these were relatively expensive stocks that didn't have exceptional top line growth in a world where nominal growth has been accelerated. That's been the that's been what the bond market move is about. That's what stock market resilience is about. Why do you want to own those stocks? Why do you want to own the stocks that aren't growing that fast, that are expensive, when everything else is growing very fast? So I think for what happened for a while is that a lot of these consumer names, and Walmart and Costco in particular, they were stand-ins for the software stocks that were bound to be disrupted by AI. Because if you think about it, if you're going to spend a certain amount at Walmart every month, that looks to me like an ARR kind of revenue profile that you've got there. lot of ways. So I think that was a stand in for disruption fears. And as software has come back, you've seen kind of multiples there deflate. So I think it is we got to find someone to punish. Maybe it's just because everyone's very invested. But when AI is doing well, you got to find someone to punish. What about just in terms of pure earnings expectations, what we're seeing here, because this is the other thing that's in theory should be happening, which is you hear about productivity gains from AI, you should be seeing people get excited about the possibility of cutting labor costs, I don't know, coming up with new products with AI or whatever. And in theory, you should be seeing that go into the earnings estimates for future years. Are we seeing any of that? I would say in theory and in practice, what we continue to see more of is the AI hardware impulse as just the straight... Fuel for earnings, Goldman had a good note released a couple of days ago where they said about half of Earnings growth this year is expected to be just driven by hyperscaler CapEx. So pretty big impulse when we're talking about how strong earnings growth has been this year. And you can just link it and distill it just down to that one factor. And then also if you look at, okay, they also ran in a separate note. What do we think explains multiples across Russell 1000 stocks? And the answer they came up with was more so than usual, it's three years forward sales estimates. sales estimates and those revisions have almost no explanatory power. So sales growth right now and for the near future, it's priced in. We all know we see the backlogs, all this, all that. The real bet on what you're willing to pay for is the kind of three years out thing. And when I did a scan of stocks that were just both expensive and had so markets willing to reward them and fairly high sales estimates, your bloom energies, a lot of the optics stocks, those are the ones that pop out there. So that's been kind of the way to earn a higher valuation right now in this on the continuation of the hardware trade. Yeah, this is something interesting. Both Tracy and I love the hair charts, the Medusa charts of all sorts, where you see historical expectations of trajectory. And one of the ones you're starting to see more is just the CapEx one. And so you look at what CellSide had estimated hyperscaler CapEx would be in 2023 and over the next three years, And so I'm curious, like there is some number exists, right? So maybe I don't know what it is. They'll say like, okay, they're expected to spend 1.2 trillion over X period of time. And that's up. But like how much of the trade is essentially a trade that... then the estimates will keep getting revised higher for spending. I got to think a solid chunk right now. And there's at some point, I think there has to be this handoff from that to a kind of kind of more broader set. And when NVIDIA and it was either a quarter ago or one before that, they actually kind of rejiggered how they're organizing their earnings to account for the fact that we expect sovereign AI and robotics and physical AI on-prem stuff. than hyperscalers and potentially neoclouds, but they're doing all they can to support that going forward. So I think right now that is a huge, huge chunk of the trade, the bulk of the trade. But the promise of AI is that it has to be more applicable than that, right? To your point, it has to have the productivity booms. It has to have the labor saving. If it doesn't, then what are we here for is kind of, I think, the question at some point. Yeah. Did you see the chart? I think The CapEx expenditure on AI is now like the biggest infrastructure boom of all time. I saw the economist Hannah Lustig, I'm not sure if I'm pronouncing his name right, who's saying that to justify all the spending that they estimate, like a handful of companies will have to get revenue that's like 9% of GDP in order to justify it, which is just like unbelievable. It's ambitious, ambitious. I'll put it that way. Yeah. [15:00] *music* There are some market stories where you want every detail. Tracy and I have made quite a few podcasts on that basis, but sometimes you've only got 10 minutes and need to know what's moving markets and why. That's the Barclays Brief Podcast. Every week, experts from Barclays markets and research get you up to speed on what matters and what to watch next. About the time it takes to grab a coffee. So search Barclays Brief wherever you get your podcasts. Some people treat ChatGPT like some kind of smart search engine, and some use it to get work done. ChatGPT Work is a new way of working in ChatGPT that can take action across your apps and files, stay with a project for hours if needed, and turn a goal into finished work. starting point to a reviewable first version. So all the source materials, briefs, and scattered information that you have to grind through to turn into something useful can just become something useful. Put ChatGPT to work on your most ambitious ideas and projects. Get started at ChatGPT.com by selecting Work Mode, available on Plus and Pro plans. Running a business means you're always on the move. The Delta SkyMiles Reserve Business American is made for frequent travelers who want premium benefits that works as hard as they do. For a limited time, new Delta SkyMiles Reserve business card members can earn 200,000 bonus miles, the highest miles offered this year. And when you need to recharge or find a quiet place to work, the Delta Sky Club can turn a layover into precious productive time or a restful reset. You can enjoy 15 visits each medallion year to the Delta Sky Club when flying Delta and unlock unlimited Delta Sky Club access after spending $75,000 in purchases on your card during a calendar year. Plus, receive four one-time guest passes each medallion year so you can bring a colleague along. That's 15 Sky Club visits plus four guest passes for an annual value of up to $950. Value estimate based on $50 per visit rate to purchase additional visits or bring a guest to the Delta Sky Club. Apply today at www.delta.com slash podcast. Offer ends 11-426. Minimum spending requirements and terms apply. This is Robert Smith from Business History. If you're listening to this, there's a good chance you're a small business owner. And like every small business owner, you started with a dream. to do what you love and watch it grow. What you probably didn't dream about, Keeping up with cyber threats. That's where MasterCard can help with access to tools that help identify cyber threats to better protect your business. Building your dream business. Priceless. For cybersecurity in a changing world, there's MasterCard. Learn more at mastercard.com slash smallbusiness. So I promised myself I wasn't going to talk exclusively fixed income in this conversation, but we can't have fluke on without talking more about fixed income. Bond yields rising. So these are the headlines that I saw when I was on vacation last week. The 10-year going to above 5%, 30-year rising as well. What's your story for this? I guess I should divide it up, global bond yields rising and then U.S. Treasury yields rising specifically. Predominantly, I think it's the same story for both or the same. Like if you had to do a principal component analysis or something, I think what would stick out across the board is it's the global growth impulse, its activity. It's both in the U.S. and globally. I think we see the industrial side of every economy is doing incredibly well and expected to do so. And I think to this about that which can be repurposed for AI will be with reference to Ford. I believe that it seems as though the impulse is just getting more widely shared. You have an impulse this big, it's bound to lift a lot of boats. And I would say that's kind of what's happening throughout the global economy. That is far and away, I would say, the one thing to think of if you had to, if you just had an elevator pitch with someone, why are bond yields high? Because growth is really strong. That's putting inflation up and central banks are expected to respond to that. But we haven't had a huge move in like break evens in the U.S., for instance. Yeah, but I think that somewhat speaks to the idea that central banks are going to be quote unquote credible in keeping a lid on inflation over time. You can believe that or not, because if you've been a whole to maturity buyer of five year U.S. treasuries from at any point in 2011 to 2021, outside of like a little before the shale bust, if you got in there, you have negative real returns for the entire period. I'm actually very curious. I find the stability of break-evens to be very surprising because usually what people say is like, well, yeah, naturally people think the Fed is going to take its dual mandates. [20:00] seriously. And OK, we're going to they're going to have to raise rates to roughly think it's and therefore inflation will be a little over two percent. Are you surprised that after years and years of the Fed missing its inflation target, there hasn't been more like, no, we do not think we do not have the confidence that the Fed will be able to maintain stable inflation in the years ahead. Or why isn't it manifesting on that chart? I would say in a word, yes, I'm genuinely surprised. I think it is an odd dynamic. The best explanation I have for it is just thinking back to when I'd be in meetings and we're discussing asset allocation, especially during 2022 at UBS and 2023, high inflation, high inflation environment. basically bonds become the residual at a certain point like you buy bonds when you're worried about growth and you don't think there's going to be that much inflation if you think growth is fine and it's inflationary well you know it's cash or short term and a lot of stocks and or stocks commodities a little bit of cash so it's just like bonds are kind of the thing you make the final decision on so i would say the fact that a lot of this is driven by real yields might come back to a a thing of like how do we get bonds to be more attractive relative to stocks And a lot of people will do the your earnings yield less the plain vanilla yield on bonds should probably be doing it versus the real yield stocks or stock earnings more of a real stream of assets. So I think in order to get bonds to be a more competitive asset with stocks, you do need to see reals rising and that a lot of the move is effectively saying like, hey, I can't justify buying a 2% even real return, 3% real return when this is what I think is on offer in AI. So I think having it come through reels is part of that story and part of why, part of the decision-making process in terms of how a cross-asset investor would be looking at bonds. Or you go the financial repression route and just make everyone hold bonds, right? That's like that's an actual possibility or we need to have a big national campaign to get people to stop spending money and to put their money into ai bonds a la war bonds right i mean yeah i with people already putting some weight on p doom with not liking the noise from data centers in their environment i think what america needs right now is to be told to to spend less so that the ai boom can more fully run its course I not running as a third party candidate and I also probably not allowed but that would be definitely my platform I would classify that under financial repression but I sure you see it differently It is a form of financial repression. Yeah, yeah, totally. I just, yeah. Just make it patriotic that it's like, no, this is actually a worrisome dynamic to me in a sort of like thinking about the economy, which is people perceive that AI supremacy is going to be crucial for national security, right? I believe this is a widely held view in Washington, et cetera. If we actually then like took that idea seriously, we would be saying like, yeah, we want to depress the non-AI parts of the economy to free up real resources for the buildout, etc no one clearly no one can say this or like articulate this or say we're going to tax consumption of this we're going to tax consumption of boats anything that could be stripped for its parts to be put into a data center etc but if it is a matter of like existential national security that would be the way we would be doing policy right almost certainly and then you have to go like with man if i being told i can still can buy a house because yeah of the i that that gonna bother the hell out of me so i think there's and it's not like and it's not like the nuclear the nuclear race where once a country has achieved a bomb it is then in the club of nuclear powers there's like it's never ending because it's like there's no like reason to presume it's like oh we got the capabilities here and when i when i hear that what i immediately think of is damn we don't know how easy we had it speaking as a policymaker back in the post gfc era oh yeah there were no like did trade-offs did not exist for an entire cycle we still probably didn't do it optimally but yeah pretty pretty darn close and now i'm i grew up hearing about how central banks just by virtue of saying inflation was going to be two percent that was one of the important factors that would help it be 2% over time. I think now we're learning that that's junk that, completely junk that. You got to prove it over time or else we'll be in a situation we are right now where people really apparently don't like to hold bonds. So Joe described this as a crazy market. I described it as a weird market. Can you choose your one word for describing the market? And then you're over at Robinhood now. I'm sure you see some interesting retail investing [25:00] trends and color and maybe a little bit of flow. Like what are you seeing at the moment? Yeah, so if I had to summarize and cross-asset how I see it, like – Last era was Tina. There is no alternative. Yes, there is no alternative. This is... too fast to fade, both prices and earnings just moving too fast to fade. And speed is something that I think has completely defined 2026. If you look at the monthly change between semis and software, those have had some of its most four violent moves in either direction this year alone. I remember when silver was going crazy, but that was this year. Absolutely nuts moves all over the And I think that was both because the power of the technology, the need to be able to look forward and see what has disrupted. And now because of agents who can work when we're asleep. So this idea of a compression of time and speed manifesting in the markets is really, really what sticks out to me. And in terms of the retail side, the trading I've seen. So I can describe net single stock market. purchases at Robinhood. So the absolute peak for that, the rolling 21-day peak, happened right about at the, in late June, let's call it. So, you know, around Micron's earnings, then that was the overall peak. There was one before that in early November, 2025, which coincided roughly with a really big spec asset peak in my view in October, 2025. That's still, you know, in some senses we haven't fully recovered. So I do a spec, a spec asset, like a lot of things And even call option volumes in total, those peaked for a while and rolled over. But that was a huge spec top that we're still kind of clawing our way back from in some senses. So there's that. And more recently, what we've seen is supposed to be traditionally a slow month September for retail. It's really rebounded quite, quite strongly. September last numbers I have are about 89% of the peak that we saw in June. We're back to. So it's been a really strong month for buying. I went on vacation and spent a lot of money. And it's a lot of it's the names you would suspect. It's your Microns, your SpaceX is those lead it. But every day when I look at this, there's a name in the top 25. I consider myself someone who knows a lot of tickers. There is a name that I do not recognize and have to look up very often. You know, it's fun. It's you know, it's a sign of how people look for volatility. People look to get rich quick, look at options trading. Stock market is a place And we do it with options with four days or less to expiry. Like that's that's the story of the world we live in now. The October 2025 speculative assets boom does not get talked about that much, but it's fun. It's a good chart. Like if you just pull up Bitcoin and Palantir, how close their peaks were together around Bitcoin. October. They both recently sort of hit a low in June. I'm sorry, I saw it was six. But I love like, actually, I love the framing of speed because it seems very relevant to even how I introduced this. And one of the things that like the AI singularity type people talk about, Kurzweil talked about, etc. is this notion of like, okay, if you measure time by how long it takes for the world, the Earth to revolve around the sun. That's like stable. If you measure time by how long does it take for there to be a headline worthy event, which is a different legitimately... Different way to measure time. This is like extraordinary because and it goes to the intro, which is just the pace of headline worthy events is just extraordinary. And then you see it manifest. Therefore, then in the speed of the movements of the whatever underlying asset you're trading. Yeah, spooze and blues should have been like all across my Twitter feed yesterday and it probably got drowned out within a couple of minutes. That's for the uninitiated. That's David Servo's old motto there back in the day. Wait, is it headlines, though, or is there also an element of market structure in here? Because the other thing that's happened is we have a lot more pod shops that are out there that seem to be affecting the market and causing sharp swings. We have a lot more products. We have a lot more zero day call options, things like that. reversals happening really, really fast now, much faster than they used to. I certainly think both the higher participation of retail, the increased AUM, if you want to call it that, in the pod shop models with tighter stops, even agentic trading, which from what I can see so far seems to be a little more Momo-driven than most in terms of the rules people will put in to do this. Yeah, I think all of that and overlay the dispersion trade [30:00] On top of that, I think all of that does contribute to an environment where you do get a lot more single stock volatility and a lot less at the index level. And that means even when markets are doing absolutely nothing, it's the classic duck with its feet underwater going crazy. We always have something to talk about, even when the stock market is completely aimless for months. This reminds me, you know, I don't trade, but one of the projects I sort of want to see how easy it is to build now is like an agentic trading bot and like paper trade on Kelsey or something. And like go do a bunch of academic research, find where people speculate that there might be mispricing and then like create a systematic thing. Make no mistakes and go. I haven't I've like started to fiddle around with that, but I haven't really done it. Do you see a lot of people like diving headlong into figuring this out? I see a lot of people like diving or tiptoeing at least into setting it up. It's honestly something I want to look at a lot more because I think it has some fun implications potentially for market structure. Like, hey, if you've got an agent setting up rules for trading, does that mean there's less orders, less stops out there? Does that mean what does that do to market visibility? That kind of thing So I interested in that and kind of how that manifests And I do need to I think talk to more people that have their fingers on the button so to speak But I really think it also has the potential to again exacerbate volatility It seems like basically everything we've done and been moving towards has the potential to exacerbate single day volatility. Thank you. that you have to grind through to turn into something useful can just become something useful. Put ChatGPT to work on your most ambitious ideas and projects. Get started at ChatGPT by selecting Work Mode available on Plus and Pro plans Running a business means you always on the move The Delta Sky Miles Reserve Business American Express card is made for frequent travelers who want premium benefits that works as hard as they do. For a limited time, new Delta Sky Miles Reserve Business card members can earn 200,000 bonus miles, the highest miles offered this year. And when you need to recharge or find a quiet place to work, the Delta Sky Club can turn a layover into precious productive time or a restful reset. You can enjoy 15 visits each medallion year to the Delta Sky Club when flying Delta and unlock unlimited Delta Sky Club access after spending $75,000 in purchases on your card during a calendar year. Plus, receive four one-time guest passes each medallion year so you can bring a colleague along. That's 15 Sky Club visits plus four guest passes for an annual value of up to $950. Value estimate based on $50 per visit rate to purchase additional visits or bring a guest to the Delta Sky Club. Apply today at www.delta.com slash podcast. Offer ends 11-4-26. Minimum spending requirements and terms apply. This is Robert Smith from Business History. If you're listening to this, there's a good chance you're a small business owner. And like every small business owner you started with a dream to do what you love and watch it grow What you probably didn dream about Keeping up with cyber threats That where MasterCard can help with access to tools that help identify cyber threats to better protect your business. Building your dream business? Priceless. For cybersecurity in a changing world, there's MasterCard. Learn more at MasterCard.com slash small business. If you listen to financial news, you know a lot of time is spent thinking about what's next. The next opportunity. The next investment. The next move. But sometimes what matters most is being ready for what you never saw coming. A storm damages a second home. A pipe bursts above a collection you spent decades building. An unexpected loss puts a growing business on pause. Those are the days no one sees coming. But for more than 75 years, Cincinnati Insurance has been ready when they do. With deep expertise, coverage for businesses, homes, valuables, and more, Cincinnati Insurance works with independent agents who take the time to understand what matters to you. Together, they help protect more than what you own. They help protect what you're building toward. And when a bad day does come, Cincinnati Insurance has real people there to help make things right. Because planning for the future isn't only about knowing what's next, it's about making sure you're ready for [35:00] but you can't predict. Let Cincinnati Insurance make your bad day better. Find an independent agent at CINFIN.com. Can I put you on the spot and ask you, I guess, the question that everyone is wondering? Because a lot of people will look at this market and say, like, it does in fact look a little bit bubbly. It's again weird that we've been able to shake off Interest rate pressures, a freaking war with Iran, all of that. What are the sort of warning signs that you're looking out for for a durable, durable, turnover, I guess, in stocks. You mentioned breadth earlier, but again, like breadth can give a false positive like we saw in 1998. Yeah, breadth is not something that I would use as like a sufficient or even necessary cause for this. I'm very boring and I like to stick for what drives stocks over the long term. Also kind of does help drive them over the medium term and shorter term. It's earnings. I took a look at it recently. So what would happen if you waited until percent, off their 52-week high? What if you, like... And we all know analysts are notoriously late. You've definitely missed the top. But, like, how bad would that be? How far off would that be? The only place where you're really, really getting killed... on that for the past six or seven bear markets is COVID, just because it happened so fast and absolutely nobody had time and we were already in the whole lockdown by the time it happened. But pretty much every other time, if you held for two years, backwards even after the time that estimates have been cut. So, you know, you're arriving late to the party, admittedly. Most of the time, you're still looking at pretty healthy gains. So my kind of thing is, hey, if you've been invested, you can probably afford to wait for earnings to start to crack at the index level and have some kind of warning signal you're getting out. Beyond that, I think it's folks like you who are keenly focused on the credit market and some of the peculiarities happening there that are going to be, I think, in the best position to be able to spot, okay, when has this kind of financial ingenuity gone too far? Because one of the things that was high conviction for me was when hyperscalers Now, either the AI boom slows or the financing starts to get more creative. The financing has gotten more creative. So, you know, at a certain point, it's interesting to think about because when you think about sources of profits right now, hyperscaler capex, obviously, and more of that debt financed, and we do have a big fiscal deficit. So can you think of more safe places to be sourcing your profits from than hyperscaler debt and fiscal debt? That's pretty darn safe until it isn't your classic kind of Minskyan formula there. Honestly, it has me surprised that multiples haven't even been more supported. Just the fact that this debt comes from kind of safe places. Right. Talk to us a little bit more about if you're on the recipient end of hyperscaler spending, it almost feels like a bond at this point, right? Like that's sort of not really, but that's kind of what you're saying, which is like these are big structural things. You can be fairly predictable. On the other hand, when you look at who some of the recipients of this money are going to be like. It's not literally the recipients of the CapEx money, but kind of like CoreWeave CDS. How do you think about credit angst among some of the AI players with the assumption that AI spending is just going to keep going through the moon? Yeah, I think you can... pretty closely map in a lot of cases just a straight up concern about the the level of debt and just look at your classic debt ratios and oracles there's a reason why oracle stands out versus all the other major hyperscalers on this and kind of use that to flag what's the reason that they stand out to just relatively more indebtedness oh yeah than peers or just like less credit worthiness to begin with that's why i think a big recent thing that's happened is the safest AI borrower possible. But yeah, no, I do think we live in a world where now a lot of companies both aren't necessarily treated as unimpeachable AI beneficiaries and also the... The second derivative's turned. We are no longer pricing in the fastest earnings or revisions across the board ever. At some point, you've got to make the transition from hyper growth to growth at a reasonable price. And a lot of stocks post the Q3 momentum road are still trying to do that. Only have two semiconductor companies that have made 52-week highs in Q3. Only two. AMD and Skyworks in the S&P 500. Interesting. Crazy. Speaking of like Oracle being different, Greg Ip at the Wall Street Journal had a really good tweet the other day. the fact that, okay, the one thing that makes the AI boom weird is that it's not being financed by the fruits of the AI boom, right? So you have all these internet, right? And so it's like the companies that are best positioned to spend have their own money printers. Facebook has gigantically profitable businesses. Google has search, etc. [40:00] Oracle is like, Find. No, I mean, nothing because Oracle is a good, successful company, but it does not have that like pure money spigot the way the other hyperscalers had. And actually, right now, today, or we continue to see. their CDS spike. Speaking of financial engineering and cash spigots, there's a headline that came across the terminal just as we started this conversation, Luke. Slight tangent, but I think it's still relevant. Man City inflated revenue cut costs by £900 million Premier League fines. They're going to be sanctioned. We don't know what the sanctions are just yet, but... Immediate reaction. Let's do sports. Immediate reaction couldn't happen to a better Manchester team. You're a Man U fan, right? Yes. So this is FAFO. I'll keep it to that. And I'm very much hoping that I will tolerate Arsenal retroactively getting trophies if it means we can also retroactively get one as well. And that takes a lot for me to say. Joe thinks the Premier League is the bowl case for Europe. I loved those competing articles. I will not say which one I enjoyed more. I love all my parents equally. You know what I, you know, our friend Guan? Yeah. He bought me, he got me a gift, the Shenzhen Pang City Football Club. He got me one of their jerseys. What colors? It's blue. They're owned by Man City. It's part of the same corporation or something like that. There's some sort of connection. Shenzhen Pang City FC is part of the City Football Group Network alongside Manchester City. And I think like Manchester City. What's the equivalent of the championship league over there? I don't know. I want to learn more. Every American. First of all, their jerseys are the same color as the Manchester City jerseys. But every American, you know, it's like, oh, we get excited about the World Cup and then we forget about soccer. Not me. Me and my son have been watching the UEFA tournament that's happening right now. Spain versus England was really good. Harry Kane missed a penalty kick. Love that. be a game. So I am not one of the Americans who just stops paying attention to soccer when it ends. I like that. You're the hipster American soccer fan. This is very good. No, hey, what else are you supposed to do? In the morning on the weekends I'd watch kids. There's things to do, but hey, for the rest of us what else is there to do? Maybe China football will become a huge thing if the UAE can't spend money in the UK anymore. It's probably going to be huge. We're just putting Luke on the spot. I was like, Luke, another football pleasure. If we're going to transition to a sport, I was really hoping we'd go to hockey next, but okay. To bring it back to markets for a second, there was an interesting note in my inbox, Terry Wiseman over at Macquarie. And he had pointed out that like, okay, so we talk about these huge spending impulses, etc., And the two big ones, obviously, AI and then structural government entitlements, et cetera, that keeps going up. And then he points out, though, that like wars are really expensive. And so markets, people look at the around war headlines. It seems like mostly through the lens of the oil price. Right. To the extent that people are trading. Their interest in the Iran war is basically like, OK, when is the straight war moves going to open? What are the prices of oil? Very diesel spreads, maybe. But he points out like wars are very expensive. And he's like, is this like this is a third driver of inflationary impulse that doesn't get much attention. But defense spending rising for years all around the world. Wars are expensive. Rearmament's expensive. Trying to reshore and build a domestic capacity of semiconductors is expensive. All of this is really, really expensive. It is the defining feature of this era, like how much... we are willing to kind of tolerate the side effects that are in our face every day of policies that are, quote, that are definitely bottom line inflationary. Anything else that really like stand out for you right now is being kind of eye opening in this market? I don't have much else. The only thing, like, I do think it is just borderline... Psycho that... game The idea that European equal weight is closer to an all-time high than U.S. equal weight when we're running like... Around 6% nominal growth right now in the U.S. Like, blows, absolutely boggles my mind that, like, U.S. consumer discretionary has become the new punching bag for the market in a strong growth environment. Like, in a quarter where, you know, it's estimated to contribute, what, 2.8 percentage points to growth. Nuts. Amazing. I guess one question I have in addition to that is, like... In a different era, there would have just been so much more attention to how, like, bad the house [45:00] market is. And like it used to be housing was the business cycle. Well, a housing was the business cycle for multiple reasons. People spent out of their home equity, et cetera. Now we have this like miserable punk housing market and it hardly seems to matter. But now AI is the business cycle, which worries me because you've got the same dynamic. AI is the business cycle and there's a cycle and this wealth effect. of like people spending out of their stock market gains. I'm just sort of like curious, like your take on, I don't know, either way you look at it, whether it's like the salience of the stock market and the wealth effect, or just the general market's lack of interest in housing these days. Well, I... Perfect timing, right? Wasn't it earlier this year when Americans share wealth from stocks versus from housing also crossed overs? Yeah, it does feel like a seminal moment in that front. And when you think about why housing was the business cycle, it's because you had these spillover effects to consumption via HELOCs or just via wealth effect from that. And also, you know, spillovers from housing markets doing better, you're buying more consumer durables. You can just take all of that and transfer it over to AI, right? Like you just think that's where the construction impulse comes from. That's where the wealth impulse comes from. And as we've talked about, it's rivaling or surpassing everything we've ever done in terms of the capex boom and size. So do I think it lasts forever? No. Do I think it's kind of more of an anomaly around the now? Certainly, because like at the end of the day, We're all either businesses trying to make a profit or consumers trying to buy something we like. Like all of this has to eventually be good for somebody downstream at the end of this or else there is no point to anything we're doing. So there is no point to the to the profit if there's no kind of consumer intent for it in the long run and a consumer income to buy it. Yeah, the circularity and the size is what. starts to worry you, I think. And when you hear that people are using their stocks, their stock market gains to buy stuff, which is helping the economy along, along with the AI CapEx build out, and that's leading to stock gains and relative economic growth. And then it just becomes very perpetual motion machining. Something that you've talked about for a long time that I've started to think about is this idea of something big is going to happen soon, like a major catastrophe, like caused by a bot or whatever, or sorry, an AI model, or there's something big could be like, utopia and everyone real weight real wealth goes up dramatically and things get a lot cheaper etc or there's something big could be a major reshuffling of people's place in the labor market in some way but i'm curious whether like okay some people thing that everyone just sort of has out there is going to happen whether that leads to like a dip in the savings rate because it's like well like this you know that's a real thing yet it's the exact opposite of the like right if you fear nuclear war then just buy stocks anyways because if it's in the price yeah i think now like you could reasonably argue the other side like hey if if the world's doomed in 10 years why am i saving it all i i i find that fun interesting someone with children, you could possibly countenance the, hey, I'm going to sacrifice their future welfare for that. So I don't think it's something that could gain broad traction. But in terms of infecting people's minds, I think we're kind of definitely already there. If AI kills us all by setting off a nuclear weapon, our portfolios will be on fire. And you'll feel stupid for not having gone on to eat. If AI doesn't kill us all... then our portfolios will also be on fire. Oh, that's a good one. Thank you. I've tweeted that before, so I'm recycling my own jokes. That's fine. That's fine. Luke, that was a lot of fun. I love your framing. Great stuff. Appreciate you coming back on Outbox. Always a pleasure. Keep it up, guys. Love following and reading what you guys do. There's a lot of fun. I could just we could talk to Luke for hours. Yeah. And I do think, you know, I went into this episode. It's like, oh, there's a crazy market. I love this. The the speed characterization. Yeah. I think that's perfect. No, I think that makes a lot of sense. And again, in my mind, there's two things happening here, which is you have a change in the actual structure of the market. And you also have just the headline. Yeah. Flying thick and fast. It's so hard to keep up. a positive headline on a Tuesday. Even if you don't believe the positive headlines, you know the markets will react to them. And so you're just sort of ping-ponging back and forth between the news flow. I want to build something where we just look. The history of red headlines would be an interesting... chart, right? Yeah. [50:00] And it would be really curious to look over trends about the pacing of them. There's got to be some way to do that. One thing I'd be interested in, it used to be the case with early machine reading trading, as we used to call it. No one calls it that anymore, that people would actually step away when there was a big headline. So there was a little bit of a dip in liquidity because it took like a second for a lot of the machines to read the news and figure out how to react. I doubt that's the case anymore. looks like around headlines but i don't know if anyone's done a study I worry about... I worry about plenty of things. Okay. Shall we leave it there? Let's leave it there. All right. This has been another episode of the Odd Thoughts Podcast. I'm Tracy Allaway. You can follow me at Tracy Allaway. And I'm Joe Weisenthal. You can follow me at The Stalwart. Follow our producers, Carmen Rodriguez at Carmen Armand, Dash O'Bennett at Dashbot, Kale Brooks at Kale Brooks, and Kevin Lozano at Kevin Lloyd Lozano. And follow our guest, Luke Kawa at LJ Kawa. For more Odd Lots content, go to Bloomberg.com slash Odd Lots. episodes and you can chat about all of these topics 24 7 in our discord discord.gg slash oddlots and if you enjoy oddlots if you like it when we do these markets episodes then please leave us a positive review on your favorite podcast platform And remember, if you are a Bloomberg subscriber, you can listen to all of our episodes absolutely ad-free. All you need to do is find the Bloomberg channel on Apple Podcasts and follow the instructions there. Thanks for listening. Thank you. If you listen to financial news, you know a lot of time is spent thinking about what's next. the next opportunity, the next investment, The next move. But sometimes what matters most is being ready for what you never saw coming. For more than 75 years, Cincinnati Insurance has worked with independent agents to help protect businesses, homes, valuables, and more. Because planning for the future isn't only about knowing what's next. It's about making sure you're ready for what you can't predict. Let Cincinnati Insurance make your bad day better. Find an independent agent at CINFIN.com. The world of business is constantly evolving, and Comcast Business keeps you totally in step with secure AI-backed networking in more than 100 countries. They're powering over 90% of the Fortune 500 and millions of small businesses. That's a lot of muscle. And behind it all, thousands of experts answering your call at 2 a.m. like it's 2 p.m. One partner powering how business gets done for companies around the globe. it all up. No one does business like Comcast Business. If you're a business owner, you need a card that can keep up. With the Delta SkyMiles Reserve Business American Express card, everyday purchases transform into miles you can use towards future travel. So business as usual can lead to a well-deserved getaway. And for a limited time, card members can earn 200,000 bonus miles, the highest miles offered this year. Whether you're traveling for business or planning your next adventure, go further. Apply today at www.delta.com slash podcast. Offer ends November 4th. Minimum spending requirements and terms apply. A business gift should do more than check a box. It should reflect your brand and show someone they're appreciated, recognized, and truly seen. 4imprint offers thousands of high-quality, customizable products. like premium apparel, drinkware, tech, and more, making it easy to create a gift that feels meaningful and on-brand. And with 4imprint's expert support and their 360-degree guarantee, You can before imprint certain your order will arrive exactly as intended. Explore gifting with confidence at 4imprint.com. For imprint. For certain.