U.S. equity markets have shown a repeated pattern: speculative investment themes surge and then collapse, often violently, yet the broader market absorbs the shock without entering a prolonged downturn. In the most recent instance, a sharp memory-chip rally inflated and then reversed within roughly four months, wiping out trillions of dollars in market value and playing a role in the failure of a hedge fund.
Despite those concentrated losses, the headline S&P 500 remains only 1.6% shy of its all-time high, and an equal-weighted version of the index recorded a fresh peak last week. Declines among AI-related names have been largely counterbalanced by gains in other areas of the market, limiting headline damage.
These boom-and-bust cycles are not new. Investors have seen similar episodes in 3D printing, Chinese equities, low-volatility products, special-purpose acquisition companies, clean energy, cannabis, space ventures and stocks tied to cryptocurrencies. Many of those themes fell by more than 80% at their lows without producing lasting damage to the broader financial system.
Several conditions have helped these episodes remain relatively contained. Extended periods of easy monetary policy and elevated government spending increased the pool of capital that could chase speculative ideas. The advent of zero-commission trading lowered the barrier to entry for individual investors and expanded participation in highly speculative trades.
More recently, margin borrowing and leveraged exchange-traded funds have amplified momentum in crowded trades, while successive waves of new technologies and financial products have continually offered fresh themes that promise rapid returns. Together, these forces have fed recurring cycles of fevered buying and steep reversals.
Crucially, past collapses were often financed principally with investor equity rather than broad-based debt. That structure concentrated losses on shareholders and other investors, while banks and many financial institutions retained capital buffers sufficient to continue lending. In effect, the transmission of losses was muted because the more systemically important lenders were not the primary source of financing for many of these bets.
That dynamic may be different in the case of artificial intelligence. An estimated $7 trillion in spending on data centres could be deployed over the next four years, creating the potential for significant capital misallocation if anticipated productivity gains from AI fail to materialize. A growing portion of that investment is being financed through borrowing, rather than solely through equity capital.
If the broader AI trade proves to be a bubble, losses could move beyond technology equity holders and reach lenders and, ultimately, the wider economy. The market’s historical ability to absorb speculative collapses offers no guarantee that a larger, debt-funded downturn tied to AI infrastructure would remain similarly contained.
In short, while prior speculative busts have been damaging for shareholders and niche investors, they have not historically derailed the broader market. The scale and financing mix behind planned AI-related investment introduce a plausible pathway for greater contagion, meaning the same pattern of containment may not hold if those investments reverse sharply.