Podcast Atlas

Bytes: Week in Review - Are we in an AI bubble?

source Episode summary Updated 2026-08-05 Tags: Podcast, Marketplace-Tech, Ai, Bubbles, Investing

Summary

This Marketplace Tech episode uses David Kirsch’s Bubbles and Crashes framework to ask whether the current AI boom is a financial bubble. [[MeganMcCartyCorino|Megan McCarty-Carino]] frames the question through capital spending, data centers, private credit, and public dependence on frontier AI services, while Kirsch argues that AI strongly matches most Tech Bubble Conditions. His conclusion is probabilistic rather than dismissive: AI may be real and socially important while still being priced and financed ahead of the time needed for durable business value.

Key Claims

  • David Kirsch is introduced as a historian and management professor at the University of Maryland who studies technological bubbles across more than 150 years of innovation.
  • Kirsch’s bubble framework focuses on uncertainty, novice investors, an investable path into the technology, and powerful surrounding narratives.
  • The source treats bubbles as a timing and expectation problem, not proof that the underlying technology is fake.
  • New technologies are especially bubble-prone because they destroy existing expertise and create uncertainty about use cases, business models, and value capture.
  • Infrastructure can act as a bubble timekeeper: railways and electrical distribution took decades to build, while AI infrastructure is being built quickly but still needs organizational and market absorption.
  • Kirsch gives AI a seven-out-of-eight bubble score because it has uncertainty, many novice investors, and strong narratives, but has fewer pure-play AI IPOs than a maximum-intensity public-market bubble.
  • The aviation analogy says early demonstrations do not settle the final business model; aviation needed airports, landing norms, weather systems, communications, military uses, airmail, and commercial operations before broad customer value became clear.
  • AGI Narrative is the most important divergence from the ordinary bubble model because the promise of general intelligence makes today’s spending look like an entry ticket to unforeseeable future opportunity.
  • The episode broadens “novice investor” beyond retail traders to professional investors in debt, data-center, private-credit, or hedge-fund markets who may still be novices with respect to AI’s technical and adoption risks.
  • Kirsch separates company failure from technology diffusion: ChatGPT, Google, and Anthropic could become operationally important even if some AI firms later fail financially.
  • The APM promo for another program at the end is not treated as part of the Marketplace Tech source.

Key Quotes

“seven out of eight” - Kirsch’s bubble score for AI.

“warning lights are flashing” - the high end of Kirsch’s zero-to-eight bubble scale.

Connections

Contradictions