Sold At “Irrational Exuberance”. Still Lost Money | Sam Ro in the AI Bubble Paradox

Key Concepts

• Market Valuations & Potential Bubbles: Current valuations are elevated, but justifying them requires considering historically high profit margins and the potential impact of AI. Timing market corrections remains difficult.
• AI as a Transformative Force: AI presents significant opportunities for productivity gains and economic growth, but also carries risks of overinvestment and diminishing returns.
• Shifting Business Models of the Magnificent 7: These companies are transitioning from asset-light, high-margin businesses to asset-heavy, capital-intensive models focused on AI infrastructure.
• Evolving Valuation Metrics: Traditional valuation methods may be insufficient to assess companies undergoing structural changes, necessitating a more nuanced approach.
• Increased Competition in AI Infrastructure: The AI infrastructure space is becoming more competitive, diminishing the competitive moats of the largest tech companies.

Market Valuations and the AI Revolution

The discussion begins with an assessment of current market valuations, acknowledging the possibility of a bubble – with Sam Row stating “100% certainty that we’re either in a bubble or this will eventually be a bubble” – but emphasizing the historical difficulty of timing such corrections, referencing Alan Greenspan’s “irrational exuberance” warning in the late 1990s. Valuations do matter over the long term, but are unreliable for short-term market timing. Elevated valuations are partially justified by historically high profit margins, as demonstrated by a chart from Truist, though the sustainability of these margins is a key concern.

The S&P 500 Forward P/E Ratio currently sits at 22x earnings, exceeding the 30-year average of 17.1x, placing it over one standard deviation above its historical average.

The Impact of Artificial Intelligence

A central theme is the potential impact of AI on productivity and earnings. AI is viewed as potentially as transformative as the internet or the automobile, but the costs associated with implementation could offset some gains. Revenue per worker has been flat for 15 years but is showing recent signs of growth, potentially driven by AI advancements, as illustrated by a BFA chart. However, the speakers anticipate significant overbuilding in AI infrastructure, drawing parallels to the railroad boom of the 1860s and the telecom bust of the early 2000s, predicting future write-downs and stranded assets. Sam Row is “convinced that all these players will overbuild and there will be write downs.”

The Magnificent 7 and Shifting Dynamics

The dominance of the “Magnificent 7” (Apple, Microsoft, Alphabet, Amazon, Nvidia, Tesla, Meta) is being challenged. While these companies drove market gains in 2023, many underperformed the S&P 500, suggesting a potential breakdown of the narrative that the market is solely reliant on them. This dispersion is viewed as a healthy sign.

Crucially, the Magnificent 7 are undergoing a structural shift from asset-light to asset-heavy models, investing heavily in AI data centers. Kai Woo’s report, “Surviving the AI Capex Boom,” highlights this transition. This shift necessitates a re-evaluation of traditional valuation methods, as running a capital-intensive business differs fundamentally from one driven by code and intellectual property.

Competition and the Utility-Like Nature of AI Infrastructure

The discussion emphasizes that building AI data centers isn’t solely the domain of the Magnificent 7. Companies like Oracle, Bitcoin miners (Coreweave), and “Neoclouds” are entering the space, diminishing their competitive moat. AI infrastructure may ultimately function as a utility – essential but not necessarily generating supernormal returns on capital due to constant asset replacement and obsolescence, similar to the historical examples of railroads and telecommunications. This could lead to multiple compression, even with strong earnings growth.

Market Narratives and Contrarian Investing

The segment challenges the assumption that underperformance of leading companies automatically translates to a broader market downturn. Investor sentiment and prevailing narratives don’t always accurately reflect underlying economic fundamentals. The current widespread concern about an AI bubble, highlighted by the Bank of America Fund Manager Survey (where it’s identified as the biggest tail risk by 38% of respondents), might actually be a contrarian indicator, potentially signaling a buying opportunity. Historical data demonstrates that the top 10 companies by decade consistently change, suggesting that current market leaders are unlikely to maintain their dominance indefinitely, as exemplified by the fate of AT&T.

Conclusion

The discussion paints a complex picture of the current market landscape. While acknowledging the potential for a bubble and the risks associated with AI investment, the speakers emphasize the importance of understanding the structural changes occurring within the tech sector. The shift towards capital-intensive investments by the Magnificent 7 necessitates a re-evaluation of traditional valuation metrics and a more nuanced approach to assessing future growth potential. Ultimately, the conversation suggests that while caution is warranted, dismissing the transformative potential of AI and the evolving dynamics of the market would be a mistake.