Posted By 
By Noah Solomon
Special to Financial Independence Hub
Should I stay, or should I go?
If I go, there will be trouble
And if I stay, it will be double
So come on and let me know
Should I stay or should I go?
— Should I Stay or Should I Go, by The Clash
Bubble or No Bubble?
During the latter part of 2025, one of the most common topics in both the media and in conversations with clients was whether markets are in a bubble, particularly with respect to AI-related companies. Given the spectacular multi-year ascent of many tech stocks and their sky-high valuations, it is unsurprising that many investors harbour serious concerns regarding a potential comeuppance.
Nobody can know for certain whether such a bubble exists, let alone how and when the proverbial story will end. However, analyzing the current environment from a risk-reward standpoint can provide investors with a useful framework to consider their current portfolio allocations and to determine whether changes are warranted.
Between a Rock & a Hard Place: Loss vs. Opportunity Cost
All bubbles, both perceived and actual, harbour the two key risks of loss and opportunity cost (i.e., missing the boat).
All bubbles eventually burst and leave a wake of losses when they do. However, in instances of false alarm where suspicions of a bubble prove unfounded, those who run for the hills suffer the opportunity cost of missing the proverbial party and leaving significant sums of money on the table.
Even in cases where the bubble moniker has proven to be accurate, things have not been straightforward. Given that irrationality and emotions are the root cause of excesses, bubbles tend to grow much bigger and to persist far longer than what may seem possible (one need look no further than Japanese stocks in the 1980s or tech stocks in the late 1990s to validate this statement). Unlike the false alarm scenario, significant losses do eventually materialize. However, this does not change the fact that investors who step aside before the comeuppance can nonetheless suffer significant opportunity costs as prices continue their often-parabolic extent long after alarm bells begin ringing.
That being said, opportunity cost is not merely comprised of the returns on the investments you forsake, but rather how those missed returns compare to those of the assets for which you forsake them. Even in instances where shunned assets deliver positive returns, if their returns are no greater or less than those of the holdings which replaced them, then the net opportunity cost is zero. Alternately stated, it’s not just about the returns you’re missing, but rather about the returns of what you’re missing vs. the returns you’re getting in their stead.
Not all Bubbles are Created Equal: A Trip down Bad Memory Lane
Historically, different bubbles have been accompanied by very different investment environments, which in turn have presented investors with vastly different prospective risks and returns. This fact is clearly evident across the three debacles of the new millennium, which include the dotcom bubble of the late 1990s/early 2000s, the global financial crisis of 2008-9, and the fixed-income duration bubble of 2021-22.
Whereas there is no such thing as a good bubble (by definition, they all eventually burst), some bubbles occur in environments that are far more ominous than others. In the worst cases, the prospective opportunity costs of avoiding them are acute. Conversely, there are bubbles during which the prospective opportunity costs of avoidance are far less pronounced. In essence, the greater the potential returns are for non-bubble assets, the lower the associated opportunity costs of avoiding the bubble.
The Dotcom Bubble (2000-3): Attractive Alternatives aplenty
The aftermath of the dotcom bubble resulted in significant losses for many investors. In early 2000, U.S. large-cap stocks stood at their highest valuations in modern history. Given the historically inverse relationship between valuations and future returns, it should have been no surprise that disappointment ensued. Between the summer of 2000 and the spring of 2003, the S&P 500 Index declined by 45% in inflation-adjusted terms, while the tech-oriented Nasdaq Composite Index fell 79%.
However, these losses were largely avoidable while simultaneously achieving reasonable returns elsewhere. Emerging-market equities, emerging-market bonds, and REITs exhibited valuations that suggested decent returns over the medium term. While not particularly inspiring, even TIPS and cash were yielding 4% and 2% above inflation, respectively. In essence, investors who were willing to re-allocate based on relative valuations were not backed into a corner: they were not forced to endure meaningfully subpar returns by avoiding what appeared to be (and were subsequently proven to be) overvalued assets.
The Everything Bubble (2007-8): Nowhere to Hide except Bonds
The “Everything Bubble” of 2007-8 and the global financial crisis that followed were entirely different animals. By the time the good times had peaked in 2007, practically all risky equity markets had become overpriced, foreshadowing negative returns over the next several years regardless of country or region.
Unlike the case with the dotcom bubble, only safe-harbour assets such as TIPS, government bonds, and cash offered positive albeit meagre returns. The only way to avoid significant losses was to liquidate all risk assets (as opposed to reallocating within them), hide in safe assets, and accept lackluster albeit positive returns. While doing so would not have been ideal, it nonetheless would have been the least bad alternative.
The Duration Bubble (2021): Nowhere to Hide
The duration bubble of 2021 bore a far greater resemblance to its 2007-8 predecessor than to the dotcom bubble, and in some respects was even more problematic. Investors were stuck between the “rock” of overvalued equities and the “hard place” of bond yields that were substantially below inflation levels. Only cash, which was the only asset that didn’t suffer losses, failed to keep pace with inflation.
At the end of 2021, the S&P 500 hit its highest ever cyclically adjusted PE ratio, save for the period preceding the dotcom crash. U.S. stocks were far from alone from a lofty valuation standpoint. Based on historical relationships between valuation and return, equities in nearly every region had negative expected returns over the next few years, thereby leaving investors with few choices to take refuge.
Further complicating matters, government bonds were also grossly overvalued and thus could not be construed as a safe harbour. Fixed-income markets remained complacent even as inflation became untethered. Specifically, 10-year Treasury yields stood at a paltry1.5%, which was far below inflation rates.
AI: If It’s a Bubble, it’s of the better Variety
The exceptional performance of mega-cap, AI-related stocks has propelled the S&P 500 to a cyclically adjusted valuation that stands above its 2021 peak, and which is only 10% below its all time high prior to the dotcom crash.
Fortunately, as was the case in late 1999/early 2000, there are plenty of other games in town. Both international developed and emerging-market equities are valued at levels that suggest attractive returns for the foreseeable future, while REITs offer returns that, while lower, are nonetheless comfortably above inflation.
For those with balanced portfolios, both U.S. and emerging-market bonds currently offer yields that are healthily positive on a net-of-inflation basis. As such, investors can reduce their exposure to large-cap U.S. stocks, mitigate the risk of severe losses, and still expect to achieve reasonable returns over the next several years.
To be clear, there is always a chance that AI-related stocks are not in a bubble. There is also a possibility that, even if a bubble exists, AI darlings nonetheless continue to deliver above average returns for a period that is sufficiently long to result in meaningful opportunity costs for those who avoid them.
However, the high expectations and seemingly unlimited possibilities of companies in the AI ecosystem are far from a well-kept secret, which is reflected in their valuations. Relatedly, while a crash is by no means imminent or guaranteed, it seems unlikely that they will perform as strongly as they have over the past several years. When juxtaposed against the reasonable valuations and expected returns of other assets, this suggests that there is limited risk from an opportunity cost perspective of rebalancing away from AI-related companies and U.S. large-cap stocks in general.
It is also a risk that I am perfectly happy to assume to mitigate possible losses from having excessive exposure to assets that may very well prove to be in a bubble.
Noah Solomon is Chief Investment Officer for Outcome Metric Asset Management Limited Partnership. From 2008 to 2016, Noah was CEO and CIO of GenFund Management Inc. (formerly Genuity Fund Management), where he designed and managed data-driven, statistically-based equity funds.
Between 2002 and 2008, Noah was a proprietary trader in the equities division of Goldman Sachs, where he deployed the firm’s capital in several quantitatively-driven investment strategies.
Prior to joining Goldman, Noah worked at Citibank and Lehman Brothers. Noah holds an MBA from the Wharton School of Business at the University of Pennsylvania, where he graduated as a Palmer Scholar (top 5% of graduating class). He also holds a BA from McGill University (magna cum laude). Noah is frequently featured in the media including a regular column in the Financial Post and appearances on BNN.
This blog originally appeared in the December 2025 Outcome newsletter and is republished on Findependence Hub with permission.
