在AI面，投资叙事押注全经济生产力跃升，但企业高层多谈节省成本、少谈新增营收；而基础设施资本开支规模被预期到2030年可达6,600亿美元，迫使科技巨头从「现金充沛」转向大量借款（部分公司成为净负债）。不少AI新创仍亏损，且彼此交易、供应商融资等做法可能抬高估值。加密资产则被描述为「无收益、纯投机」：比特币自10月6日高点回落；杠杆还常出现同一批波动代币反复作为抵押、价格下跌引发连锁清算的自我强化机制。作者并点名近年「赌博民主化」：社群媒体推波助澜、杠杆ETF十年暴增、以及社群炒作后暴跌的拉抬出货（pump-and-dump）。
宏观层面上，金融条件仍偏宽松：市场预期美国利率续降、财政在川普主导下偏扩张，联邦赤字超过GDP的6%。然而通膨仍高于2%目标、公共债务接近GDP的100%且期限偏短，使联准会若升息反通膨将迅速推高政府融资成本；加上若把货币市场基金等「准货币」纳入，货币增速呈上行。作者担心放松银行资本比率带来顺周期的信用扩张，最终使投资人要求更高美债殖利率，造成债券价格下跌并外溢到股市。同时，私募领域杠杆攀升：大型银行对非存款机构放贷约1.3兆美元，另有「数兆」未动用承诺额度。2025年的趋势被概括为美元走弱、非美元主权资产配置上升、金价大涨（甚至被形容为6,000年的泡沫）；作者预期2026年仍有更多泡沫与Fomo，最终可能崩裂，但央行「安全网」使其不致重演大萧条，代价可能是更久的停滞性通膨与政治压力。
The piece argues today’s AI and crypto euphoria resembles past bubbles—from the South Sea Bubble (1720) to the Roaring Twenties—raising the specter of a 1929-style crash nearly 100 years later. It pushes back on the “you can only spot bubbles after the fact” stance popularized in the late-1990s dotcom era, noting that valuation tools like P/E ratios and discounted future revenues can flag prices that detach from fundamentals even before the break.
AI may lift productivity, but executives often cite cost savings more than new revenues, while Big Tech’s infrastructure build-out could reach about $660bn by 2030. As spending scales, firms that began cash-rich are borrowing heavily; several become net debtors, and many leading AI startups remain loss-making. Crypto is framed as a speculative asset with no income stream; bitcoin has fallen since an Oct. 6 peak, and leverage can be self-referential when the same volatile tokens repeatedly serve as collateral, accelerating liquidations on downturns.
Macro risk amplifiers include easing financial conditions, a US deficit above 6% of GDP, inflation still over a 2% target, and public debt near 100% of GDP with short maturities—making rate hikes fiscally painful. In private markets, banks have about $1.3tn of loans to non-depository financial institutions plus trillions in unused committed lines. The author also points to 2025 trends—dollar weakness and surging gold—and expects more froth in 2026 before stress or a shock reprices US risk; central banks may backstop markets, limiting a repeat Great Depression.
