He Wrote the Book on Bubbles | Edward Chancellor On If AI is Different

Key Concepts

• Capital Cycle Theory: A framework suggesting that investment booms lead to overcapacity, competition, and eventually lower returns, often driven by the "Prisoner’s Dilemma" where firms overinvest to avoid losing market share.
• Intangible Economy: An economic shift where value is increasingly derived from R&D, software, and human capital rather than physical assets, though subject to similar boom-bust cycles.
• Anti-Bubbles: Market sectors that are depressed or undervalued due to being perceived as "old economy" or threatened by new technology, often offering superior long-term investment opportunities.
• Lindy Effect: The idea that the future life expectancy of a non-perishable thing (like a technology or a brand) is proportional to its current age.
• Hindenburg Moment: A catastrophic event that causes a sudden loss of public and investor confidence in a hyped technology.
• Hallucinations: Inherent errors in Large Language Models (LLMs) where the system generates incorrect or nonsensical information, limiting their reliability in critical applications.

---

1. The Capital Cycle and Technology Booms

Edward Chancellor argues that history—from the 1840s British railway mania to the 2000s TMT (Technology, Media, and Telecom) bubble—follows a recurring pattern:

• The Cycle: A new technology emerges, attracting massive capital. Because firms fear being left behind (the Prisoner’s Dilemma), they overinvest, leading to fragmented markets and duplicative infrastructure.
• Profitability Illusion: During the boom, aggregate profits often rise because companies capitalize expenses rather than depreciating them immediately. When the bubble bursts, the reality of misallocated capital forces a collapse in profitability.
• The AI Context: Current AI investment is characterized by extreme hype. Chancellor notes that while the "picks and shovels" (semiconductors, data centers) are currently booming, the market historically struggles to identify the long-term winners, often leading to massive losses for early investors.

2. Demand vs. Supply and the "Hindenburg Moment"

A critical risk in the current AI cycle is the disconnect between supply (massive capex) and actual demand.

• Data Miscalculation: During the dotcom era, false data regarding internet traffic growth fueled overcapacity. Chancellor warns that current AI demand forecasts may be similarly inflated.
• Reliability Constraints: Chancellor is skeptical of "Agentic AI" (AI that performs tasks autonomously). He cites the "hallucination" problem—where models make probabilistic errors—as a fundamental barrier to adoption in high-stakes industries.
• The Hindenburg Moment: He suggests that a single high-profile failure (e.g., an AI-driven system causing a major disaster) could trigger a "Hindenburg moment," causing a sudden, sharp pause in development and investment.

3. Investment Strategy: Finding "Anti-Bubbles"

Chancellor advises against shorting bubbles, as they can persist longer than expected. Instead, he suggests looking for anti-bubbles:

• Definition: Sectors that have been sold off because the market assumes they will be disrupted by new technology, even when their underlying business models remain robust.
• Examples:
  • Energy: During the 2020-2021 SPAC boom, traditional energy stocks were severely undervalued, creating a massive opportunity.
  • Legacy Software/Services: Companies providing essential services (e.g., car auctions, travel software) that are currently trading at depressed multiples due to fears of AI disruption may be mispriced.
• Caveat: Investors must perform rigorous analysis to ensure the company isn't truly a "zero" (like Blockbuster) and should be wary of companies with high stock-based compensation, which can mask poor profitability.

4. Intangible Capital Cycles

The discussion highlights that capital cycles apply to intangible assets just as they do to physical ones:

• Big Pharma: The 1990s saw a surge in R&D spending for "blockbuster" drugs that eventually failed to deliver adequate returns on equity.
• Human Capital: The "Harvard MBA indicator"—where the brightest graduates flock to a specific industry (e.g., Finance in 2007, Big Tech today)—often serves as a reliable contrarian signal for the peak of a cycle.
• Valuation: Intangible assets are often misvalued because accounting conventions expense R&D rather than treating it as a long-term investment.

5. Synthesis and Conclusion

Chancellor concludes that while technology booms can be beneficial for society in the very long run, they are often disastrous for investors who enter at the peak. He emphasizes that the "easy money" era of the last decade—fueled by post-COVID liquidity and low interest rates—is likely ending.

Key Takeaways:

• Avoid the Herd: When an entire sector is rushing to invest, the capital cycle investor should step back.
• Focus on Value: Look for companies with proven, durable business models that the market has unfairly discarded.
• Gold as a Hedge: Chancellor maintains a contrarian view on gold, viewing it as a necessary "asset without a liability" in a world of high debt and potentially rising long-term interest rates.
• Skepticism of Hype: The "fake it till you make it" culture of Silicon Valley, while sometimes self-fulfilling, significantly increases the risk of a painful market correction when the underlying technology fails to meet the inflated expectations.