AI Investment Dwarfs Dot-Com Era, Raising Fears Of A Crisis Four Times Larger

Global financial markets are facing risks that could dwarf previous crashes, with AI investment alone running roughly 17 times higher than the 2000 dot-com boom.

The dot-com collapse and the 2008 subprime housing crisis caused estimated losses of $5 to $6 trillion and $15 to $20 trillion respectively, setting a grim benchmark for what another collapse could cost.

By contrast, the scale of current AI-driven investment runs approximately four times the size of the subprime housing crisis, suggesting the potential for catastrophic losses if sentiment reverses.

Property, the largest concentration of global wealth, is overvalued by up to 50% using price-to-affordability metrics, though significant differences exist depending on location and market.

U.S. property is overvalued by roughly 10% on average, but individual cities show high bubble risks, as do parts of Australasia, Europe and Asia, especially major financial hubs.

Global stock prices remain elevated despite recent corrections, with the Shiller CAPE ratio sitting at a significant premium to its 25-year average.

McKinsey research across 10 countries accounting for 60% of global GDP found that $500 trillion in real assets now supports more than $1,000 trillion in financial assets, a staggering imbalance.

For every $1 in net new investment in real assets over the past 20 years, liabilities have grown by almost $4, of which about $2 is debt, reflecting how credit has supercharged the global economy.

Debt levels are stretched across both the real and financial economies, with a layer cake of leverage existing at the investor, fund, instrument, and underlying investment levels simultaneously.

For capital importers like the U.S., where foreign investors hold exposures of around $30 trillion, falling asset prices would be amplified by currency weakness as investors exit, creating adverse feedback loops.

Crowded trades and herding behavior, including exchange-traded funds and quantitative trend-following models, would accelerate any downward spiral once selling pressure builds.

Retail investors driven by FOMO, FEMO, and TINA dynamics remain deeply exposed, with eventual panic described as NOGO, meaning no option but getting out, taking hold as risk aversion rises sharply.

Signs of stress are already visible, including overpriced equity and debt issuance by SpaceX, gating of funds invested in unlisted equity and debt, and rising defaults in highly leveraged parts of the economy.

High valuations and volatility are tolerable only when funded by long-term equity capital, not borrowed money that must be continually serviced and repaid under pressure.

All financial bacchanals end when an event, often minor but only obvious in hindsight, sets off an uncontrollable chain reaction, and markets repeatedly forget just how costly that lesson turns out to be.