The AI Bubble, Like the Housing Bubble, Is a Big Problem—and It’s Not Complicated

Less than 20 years ago, the collapse of a nationwide housing bubble triggered the Great Recession. Millions of homeowners faced foreclosure, unemployment soared, and the subsequent decline in construction contributed to another surge in house prices during the pandemic. The current boom in artificial intelligence (AI) is laying the groundwork for a potentially similar economic downturn.

The Simplicity Beneath the Complexity

Often, experts complicate economic problems unnecessarily. Richard Bookstaber, a hedge fund manager who predicted the financial crisis following the housing bubble’s collapse, acknowledges the AI bubble but also points to risks in the private credit market and geopolitical concerns—such as potential disruptions to the supply of semiconductors from Taiwan or oil from the Strait of Hormuz—as significant threats. Whereas these warnings are valid, the core issue remains straightforward.

Echoes of the Housing Bubble

Bookstaber highlights that these potential problems are interconnected within a “complex and tightly coupled system,” where the specific trigger matters less than the speed at which stress spreads. This mirrors the situation preceding the 2000s housing bubble. From 1996 to 2006, real house prices nationwide increased by 70 percent [1], a dramatic departure from the previous century’s average of keeping pace with overall inflation.

This run-up in house prices occurred despite a relatively high vacancy rate and without a corresponding increase in rents, which largely tracked inflation. The boom in residential construction peaked at 6.7% of GDP in the fourth quarter of 2005, then plummeted to 2.4% of GDP by the third quarter of 2010 after prices began to fall.

The Real Impact: Demand Destruction

The Great Recession wasn’t primarily caused by the financial crisis itself, but by the collapse of the housing bubble. The loss of demand resulting from the construction slowdown amounted to 4.3 percentage points of GDP, equivalent to approximately $1.3 trillion in today’s economy. The loss of trillions in homeowner wealth further reduced demand by an estimated 1-2 percentage points of GDP ($320-$640 billion today).

While the financial crisis provided a spectacle—with politicians reluctant to let Wall Street bankers suffer the consequences of their actions—it was a secondary effect of the underlying bubble.

The AI Bubble: A Similar Pattern

The current situation mirrors the housing bubble: a grossly inflated stock market fueled by the AI boom. A freeze-up in private credit would be less concerning if not for its role in driving the AI bubble. Just as the subprime mortgage market’s collapse extinguished the fuel for the housing bubble’s expansion, a disruption in credit for AI could have severe consequences.

The Rise of Chinese AI

Adding to the potential risks is the rapid growth of Chinese AI companies, which are focusing on user-friendliness and affordability. Some reports suggest they captured 30 percent of the global market by December 2018 [2], and their market share has likely increased since then. While U.S. Companies prioritize massive computing power, Chinese firms are developing practical, low-cost applications. This approach could prove more successful in the long run, potentially undermining the inflated profits stock investors are anticipating.

Geopolitical Considerations

Geopolitical factors, such as former President Trump’s policies toward Iran, could also deter international users from relying on American AI technology, fearing potential disruptions in access.

The Inevitable Burst

The exact cause of the AI bubble’s burst is unpredictable, but its existence as a significant driver of the economy is a real problem. The complex financial structures supporting the housing bubble were intricate, but the bubble itself was simple. The same holds true for the AI bubble. The focus should be on the bubble itself, not the specific trigger that causes it to deflate.