So are we in an AI bubble? Here are clues to look for. - Planet Money Recap

Podcast: Planet Money

Published: 2026-01-10

Duration: 25 minutes

Guests: Robin Greenwood, Gotti Barlevi

Summary

The episode dives into whether the AI industry is experiencing a bubble, exploring the indicators economists use to detect bubbles and the potential economic impact if such a bubble bursts.

What Happened

The episode begins with an exploration of the current stock market surge, particularly driven by a few AI giants like Microsoft, Amazon, and NVIDIA, leading to speculation about an AI bubble. Economists like Robin Greenwood from Harvard Business School have been studying bubbles and have identified certain clues that might indicate a bubble, such as high valuations and volatility. Greenwood's research also reveals that bubbles often form in new, exciting industries, making it difficult to assess whether AI is truly in a bubble phase.

The episode discusses the challenge of detecting bubbles, referencing economist Eugene Fama's skepticism about their existence due to market efficiency theories. Greenwood and his colleagues attempted to prove Fama wrong by analyzing historical data, identifying patterns in stock market behavior to predict bubbles, albeit with limited accuracy.

Four main clues were identified as potential indicators of bubbles: high valuations, volatility, stock issuance, and acceleration. The current AI landscape showcases some of these indicators, particularly high valuations and volatility, but lacks significant new stock issuance, suggesting an early bubble stage.

Economist Gotti Barlevi discusses the implications of potential bubbles and the debate over whether governments should intervene to prevent them. He highlights that not all bubbles have severe economic consequences, with the impact depending on factors such as borrowing levels and industry connections.

The episode also considers the potential benefits of bubbles, suggesting that they might spur investment in underfunded areas like research and development, thus having positive long-term effects despite short-term irrational exuberance.

Using the dot-com bubble as a historical example, the episode illustrates how the excess investment in fiber optics later benefited society by enabling the broadband era, suggesting that bubbles can sometimes lead to valuable infrastructure or technological advancements.

Finally, the episode mentions the AI boom's potential risks and benefits, indicating that while a burst could lead to significant financial losses, the technology and infrastructure developed might still benefit society in the long run.

Key Insights

• High valuations and volatility are common indicators of financial bubbles, and the current AI sector exhibits both, suggesting it may be in the early stages of a bubble.
• Historical analysis by economists has identified four main indicators of bubbles: high valuations, volatility, stock issuance, and acceleration, with the AI sector currently lacking significant new stock issuance.
• Bubbles can have positive long-term effects by encouraging investment in underfunded areas like research and development, as seen with the dot-com bubble's contribution to the broadband era.
• The potential risks of an AI bubble include significant financial losses, but the technology and infrastructure developed during the boom could still provide societal benefits.