The "AI Bubble"

Key Concepts

• Market Concentration: The extent to which a small number of stocks constitute a large percentage of a market index (e.g., S&P 500). Currently at historically high levels in the US (36% in top 7 stocks).
• Market Valuations: A measure of how expensive stocks are relative to their expected future earnings, often assessed using metrics like the Cyclically Adjusted Price-to-Earnings (CAPE) ratio. US valuations are nearing 1999 peaks.
• CAPE Ratio (Cyclically Adjusted Price-to-Earnings Ratio): A valuation measure that uses average inflation-adjusted earnings from the previous 10 years. A high CAPE ratio suggests higher expected future returns will be lower.
• AI Bubble: The potential for inflated stock prices driven by speculation surrounding artificial intelligence technologies. Its existence is currently uncertain.
• Value Stocks: Stocks trading at a lower price relative to their fundamentals (e.g., book value, earnings). Historically, they have shown resilience during market downturns.
• Small-Cap Value Stocks: Value stocks with relatively small market capitalizations. Often outperform during periods of market stress.
• Diversification: Spreading investments across different asset classes, geographies, and sectors to reduce risk.

US Market Concentration, Valuations, and Historical Bubbles

The US stock market currently exhibits two concerning features: high market valuations and significant market concentration. 36% of the S&P 500 is comprised of just seven stocks, and 32% of the total US market is held in these same stocks – the most extreme concentration in US market history dating back to 1927. Furthermore, US stock market valuations are approaching levels seen during the 1999 dot-com bubble, which was followed by a decade of flat or negative returns.

Historical Parallels: Canada and the Dot-Com Bubble

Ben Felix draws parallels to the Canadian market experience during the dot-com bubble. In July 2000, Nortel Networks comprised 36% of the entire Canadian market index (TSSE 300). The company subsequently collapsed, becoming worthless and dragging the Canadian market down with it. Nortel peaked at a Schiller P/E ratio of 60.62, exceeding the US market’s peak during the same period. However, despite this extreme concentration, the Canadian market recovered by July 2005 and delivered strong returns, outperforming the US market which struggled for over a decade. This highlights that extreme concentration doesn’t automatically guarantee prolonged underperformance.

The US Dot-Com Bubble and Subsequent “Lost Decade”

The US market in 1999, while not as concentrated as Canada’s during the Nortel bubble, also experienced high valuations. While some companies like Microsoft and Amazon proved transformative, the majority failed. This led to the dot-com bubble burst and a “lost decade” for US stocks, with the market remaining flat or below flat (in Canadian dollar terms) until July 2013.

Value and Small-Cap Value as Protective Strategies

A recurring theme across both the Canadian and US examples is the outperformance of value and small-cap value stocks during and after market downturns. While the overall Canadian market was impacted by the Nortel crash, Canadian value stocks did not crash and delivered stronger returns on the recovery. Similarly, during the US “lost decade,” US value and small-cap value stocks generated positive returns while the broader market remained flat.

The AI Bubble – Uncertainty and Recent Trends

The video acknowledges the uncertainty surrounding the current “AI bubble.” While it’s impossible to know if a bubble exists in hindsight, significant investment is flowing into AI-related companies. Since the launch of ChatGPT in November 2022, AI-related stocks have accounted for 75% of S&P 500 returns, 80% of earnings growth, and 90% of capital spending growth (as of a September 2023 JP Morgan report). This rapid growth mirrors the patterns observed during previous technological revolutions (railroads, the internet).

Global Market Concentration and Valuations

The video points out that US market concentration, while high, isn’t unique. Many other global markets are even more concentrated. Examining the 10 largest stock markets (excluding the US) in November 2015, the average weight of the top seven stocks was 40.94%, with Switzerland reaching 60.11%. Despite this, these markets delivered an average return of 8.44% over the subsequent 10 years, demonstrating that high concentration doesn’t necessarily preclude positive returns.

Valuations as a Stronger Indicator

While market concentration has a weak and statistically insignificant relationship with future returns, market valuations appear to be a stronger indicator. Analysis of the 10 largest developed stock markets from 1982 to 2024 reveals a clear monotonic relationship: higher starting CAPE ratios are associated with lower future 10-year returns.

The Japanese Example: A Cautionary Tale

The Japanese stock market in the late 1980s serves as a cautionary tale. It experienced a massive bubble, surpassing the US market in capitalization, but crashed at the end of 1989 and has not recovered in real terms as of almost 2026. Diversification across markets and within the Japanese market (into value and small-cap value stocks) would have protected investors.

Key Takeaways and Recommendations

The video concludes that diversification and discipline are crucial. A diversified investor should be comfortable holding both winners and losers, understanding that long-term success relies on the winners outperforming the losers. While the US market’s current valuations are concerning, it’s not necessarily a signal to exit the market. Instead, investors should moderate their return expectations and maintain a disciplined, diversified approach.

Notable Quote:

“Diversification is the only free lunch in investing for good reason.” – Ben Felix.

Technical Terms Explained:

• Schiller P/E Ratio (CAPE Ratio): A valuation metric using 10-year average inflation-adjusted earnings.
• Market Capitalization: The total value of a company’s outstanding shares.
• Monotonic Relationship: A relationship where one variable consistently increases or decreases as the other variable changes.
• Statistically Significant: A result that is unlikely to have occurred by chance.

The video emphasizes that while historical patterns don’t guarantee future outcomes, understanding past bubbles and their aftermath can inform investment strategies and manage expectations. The core message is to prioritize diversification and maintain a long-term perspective, rather than attempting to time the market based on current conditions.