Always Think Symmetrically About Risk

25 September 2025

One of my mentors used to say always think symmetrically about risk. Put another way, it is important to consider upside and downside risk. Although most market participants tend to benefit from rising asset prices, risk is often framed in terms of left tail or downside risk. To be fair, with US equity valuation in the top 5% of all observations, risk-reward probably is skewed to the downside on an extended time frame.

The current speculative episode in Artificial Intelligence (AI) has many parallels with the first Dot.com bubble in 2000. Indeed, on some measures the divergence between price and fundamentals is even greater. Moreover, the MSCI US Information Technology sector now trades at 10 times sales which is greater than at the peak in 2000. In that episode, the market correction was concentrated in that sector. However, there was meaningful out performance by “value” equities, from 2001-2003. While we would expect value to outperform again when there is a correction in the AI bubble, there are some important similarities and differences to note from the 2000 episode.

A key similarity to the first episode is the capital intense nature. Early in the bull market, the mega-cap tech companies famously had a “capital lite” model. However, the second leg of the bubble (AI) by the hyper-scalers has become capital intensive. As Gerard Minack noted recently, capital intensity has created a positive feedback loop between rising investment spending and rising profits: the firm selling the capital good immediately reports its profits in full, while the firm buying the capital good depreciates its cost over time. Consensus forecasts expect the Mag 7 investment spending to rise to over $450 billion over the year to the June quarter 2026.

A related point to the reflexive feedback loop between investment spending and profits is the recent “vendor financing” announced by NVIDIA. That was also feature of the first episode when the telecom equipment makers (Cisco, Lucent and Nortel and others) extended loans, equity investments or credit guarantees to their customers who used the cash/credit to buy back the equipment. The reverse feedback loop was ugly in the 2001-2002 bust.

Another similarity between the current episode and the late-90s cycle is that investors are probably overestimating the likely return on this investment. According to MIT 95% of organisations are getting zero return on the capital expenditure. However, another lesson of the first bubble is that equity prices are likely to correct before a decline is reported by firms.

An important lesson from the first episode is that “value” stocks outperformed when the speculative “growth and momentum” equities corrected. As a factor, MSCI US Value is trading at a 50% discount to MSCI growth. However, US value is trading at 20 times earnings which is one standard deviation above its long-term average (16.9 times earnings). That might be heroic if the business cycle deteriorates. Another US sector trading at a material discount to the technology sector (leaders) is health care. The relative discount is almost 60%. Although the outright and relative valuation of the health care sector declined even further in 2009 and 2002 at the trough. On the positive side, positioning and consensus beliefs are already very pessimistic (bullish for the non-consensus investors).

Chart 1

Source: Bloomberg

20250925_Always_Think_Symmetrically About Risk.pdf

Download PDF • 483KB