Has AI reached its Wile E. Coyote moment? Since the start of the huge boom in AI-related capital spending, particularly in data centers, many observers have warned that the boom appears unsustainable. AI is a remarkable technology, which can generate significant economic benefits (even if these benefits are not yet clear in the data). But it’s doubtful that the companies making these huge investments will be able to capture enough of those spillovers to compensate them for the enormous sums they’re spending.
If these skeptics are right, we should expect there to be a Wile E. Coyote moment: Investors and lenders backing AI companies look down, realize they’ve fallen off the edge of a cliff, and the boom collapses. Such sudden reckonings have occurred several times in the past, most notably with the bursting of the dotcom bubble in 2000 and the bursting of the housing bubble a few years later.
And in recent weeks, we’ve seen increasing signs of nervousness about AI’s financial prospects. Price of hyperscaler stocks are down 27 percent from their peak; the cost of insurance against non-payment on AI company bonds has soared. We don’t know for sure that this is a true turning point for the industry, but the mood has definitely changed.
Today’s introduction will not attempt to settle the hotly (and endlessly) debated question of whether AI really is a bubble, much less whether it is the moment of truth. Instead, I will focus on a conditional question: If AI is a bubble and this bubble is about to burst or is likely to burst soon, what are the likely economic consequences? This is, as I will explain, a question we can resolve with ample historical evidence. Because even though AI as a technology is completely new, bursting bubbles have been a recurring economic theme for over a century, and we know a lot about the macroeconomic consequences of bubbles.
To preview my conclusions: a bursting of the AI bubble will almost surely cause a recession. The extent of this recession is very uncertain, although I favor the idea that it will be worse than the recession at the end of the 1990s tech boom. But even if the recession is short and shallow, history and theory suggest that recovery from this recession will be extremely slow and unemployment will remain high for several years.
Beyond the paywall, I will address the following points:
1. A Brief History of Modern Recessions
2. Why recovery after bubble burst is slow
3. Comparing a possible AI collapse to the Internet industry recession
4. Why the policy response to an AI collapse will likely be inadequate to prevent a painful recession
Gn bussni
