“You know I’ve been to the edge
And there I stood and looked down
You know I lost a lot of friends there, baby
I got no time to mess around“
(From “Ain’t Talkin; ‘Bout Love,” by Van Halen, 1978)
By Scott Welch, CIMA®, CEPA®, Chief Investment Officer & Partner
Reviewed by Carter Mecham, CMA®, IACCP®
First, let’s acknowledge and celebrate the success of the stock market performance of Artificial Intelligence (AI) -oriented companies over the past 2+ years. We can proxy the AI mania of roughly the past two years using the often-cited “Magnificent Seven” stocks (we’ll make it eight): Meta (Facebook), Apple, Amazon, Netflix, Microsoft, Nvidia, Tesla, and Alphabet (Google).
Source: Ycharts, data from Jan 2, 2024, through October 6, 2025. Past performance is no guarantee of future results.
Let’s also acknowledge that this amazing price rally has been accompanied by a corresponding (1) increase in the market capitalization percentage of these stocks within the S&P 500 index – an unprecedented almost 35% (first chart), and (2) the forward-looking P/E ratios of these stocks, which are “frothy” to say the least (second chart).
Source for both charts: Yardeni Research, as of September 30, 2025. You cannot invest in an index and past performance is no guarantee of future results.
One interpretation is that these stocks must continue to generate robust earnings growth to justify/maintain their lofty stock prices. The consensus view (at least for now) is that they can/will do so.
Source: Zacks Earnings Analysis, as of October 3, 2025. Hashed bars are estimates and subject to change as additional data come in. You cannot invest in an index and past performance is no guarantee of future results.
What is interesting to us about this chart is that tech earnings are expected to grow despite relatively flat revenue growth. Clearly, the expansion of AI and the accompanying productivity growth are being priced into the current market, along with an expectation of an accommodative Fed going forward.
And the rally has not been limited just to the AI tech stocks themselves. The semiconductor companies that will have to produce all those chips to power the AI surge have benefited as well.
Source: Bloomberg Markets Daily, as of October 2, 2025. You cannot invest in an index and past performance is no guarantee of future results.
History Does Not Repeat Itself, But Sometimes It Rhymes
But let’s remind ourselves of what can happen when stocks are “priced for perfection.” We need to go no further back than early this year, after the “DeepSeek” news broke (which potentially challenged the dominance of US-based AI) and the “Liberation Day” tariff announcements of early April.
Source: Ycharts, from January 2, 2025, through April 30, 2025. Past performance is no guarantee of future results.
This is not to pick on the mega-cap tech stocks – they are solid companies with excellent future prospects. The question is whether their prices are overly optimistic.
We’ve been around long enough to have seen this movie before (or, at least, a variation of it).
Here is the performance of two prototypical “Mag 7” stocks – Meta (Facebook) and Netflix – over the past three years.
Source: Ycharts, 3-year performance through October 6, 2025. Past performance is no guarantee of future results.
And here is the three-year performance of Cisco heading into the tech crash of the early 2000s.
Source: Ycharts, 3-year data from January 2, 1997, through January 3, 2000. Past performance is no guarantee of future results.
And, of course, here is what happened to Cisco in the following fifteen months, through March 2001.
Source: Ycharts, data from January 2, 2000, through March 31, 2001. Past performance is no guarantee of future results.
We are not suggesting this is what will happen to today’s mega-cap tech stocks – only that it could.
We are reminded of Pascal’s Wager.[1] Blaise Pascal was a 15th century French philosopher and mathematician, who also happened to be deeply religious. He was frustrated by his inability to mathematically “prove” that God existed, but he eventually settled on this compromise: it was better to believe in God and live a good life than to not believe in God and live a bad life.
If it turned out to be true that God existed and you had led a good life, then it all works out for you in the end. But if you chose to believe that God did not exist and you led a bad life – and you were wrong – then you were faced with eternal damnation. So, it was better to live a good life.
This became known as Pascal’s Wager and, in investment terms it can be described as the probability of being wrong can be very different than the consequences of being wrong.
Put in statistical terms, this describes a “low probability – high consequence” event.
Back in the late 1990s, the internet phenomenon was in full swing, and most believed there was an unending demand for Cisco’s hardware. That turned out to not be true or, more accurately, the expectation of the demand for Cisco’s hardware dramatically exceeded the actual demandfor its hardware – at least for a while.
There are other historical precedents for industry supply (as measured by Capital Expenditures as a percentage of EBITDA) exceeding expected but not actual demand: the railroad industry in the 1860s and 1870s, the telecom industry from 1995 – 2000, and the energy industry in 2014 – 2015.
Source: CQG Partners, “Dotcom on Steroids,” September 11, 2025. For illustration purposes only and does not represent investment advice.
And we can also see what happened to the earnings of the high-flying “dotcom” companies when demand did not keep up with supply.
Source: CQG Partners, “Dotcom on Steroids,” September 11, 2025. You cannot invest in an index and past performance is no guarantee of future results.
So, What Does Any of This Have to Do with Dr. Hyman Minsky?
Dr. Hyman Minsky was a late-20th century economist who was not especially renowned or acclaimed in his own lifetime (he died in 1996).
His primary contribution to economic thought was his research into and explanation of financial crises, summarized in his working paper, “The Financial Instability Hypothesis.”[2] In his own words, he believed, “Instability is an inherent and inescapable flaw of capitalism.”
Specifically, Minsky segmented borrowers into three categories:
- Hedge borrowers, whose cash flow can cover both interest and principal on any debt incurred;
- Speculative borrowers, whose cash flow can cover interest but not principal; and
- Ponzi borrowers, whose cash flow cannot cover either interest or principal and who must rely on rising asset prices and increased borrowing to survive.
Minsky’s theory suggests that in strong economic times the risk of failure is “forgotten”, leading to increased borrowing and a slow (but inevitable) “flow” from hedge borrowing through speculative borrowing and finally to Ponzi borrowing, at which point a market bubble has evolved that will soon, just as inevitably, pop.
During periods of economic strength and seeming stability “within horizon” risk is forgotten or ignored, and stability is extrapolated into perpetuity (the “trees grow to the sky” phenomenon).
Or, to paraphrase into Panglossian-like terms: “All is for the best in this, the best of all possible worlds.”[3]
[But we should not forget from our 10th grade literature class that, in the novel “Candide”, the fictional antagonist Pangloss was one of the few survivors of the (actual) 1755 earthquake in Lisbon.[4]
This leads hedge borrowers to morph slowly into speculative borrowers and speculative borrowers to morph into Ponzi borrowers precariously balanced on a proverbial house of cards held together only by easy credit and rising asset values.
In other words, periods of market stability mask underlying elevated levels of market instability until a “Minsky Moment” occurs – a tipping point when the credit bubble pops, the leveraged house of cards collapses, and financial markets crater.
This helps to explain why so many investors do not see – or choose to ignore – any warning signals that may precede a market collapse – as a collective body we forget that failure is an option.
We can see, perhaps, an indication of this market complacency (“failure is not an option”) in looking at today’s overall market’s valuation.
A “rule of thumb” widely tracked by many analysts is the “Buffett Indicator,” which measures the total equity market cap as a percentage of US GDP. Any ratio greater than 200% is viewed as an indicator of an overvalued and overbought stock market or, in Buffett’s own words, “playing with fire”.
As of early October, that ratio sat at 217%.
Source: CNBC, as of October 1, 2025. Past performance is no guarantee of future results.
Again, we are not forecasting doom and gloom to happen to today’s mega-cap tech stocks, but here are some things that give us pause.
- We have seen more analysis and commentary in the past month worrying about an AI “bubble” than in any time since the phenomenon began two years ago. Specifically, a constant theme is the notion of “circular investing.”[5] This captures the idea that, while these wildly successful mega-cap tech companies are pledging to invest trillions of dollars into AI and its adjacent services (e.g., energy production and transmission, data centers, etc.), for the most part they are planning to invest in each other.
Source: Bloomberg Markets Daily, as of October 7, 2025. For illustration purposes only.
In other words, it appears to be an increasingly insular ecosystem, which increases the risk that supply will dramatically exceed demand, at least in the near future. (Remember, demand did eventually catch up to supply in railroads, energy, telecom, and computer hardware, but it took a while and there were disruptions along the way.)
2. Everyone is leaning in the same direction. At the request of a prospective client, we recently surveyed analysts’ recommendations for 10-15 of the stocks that would be included in most people’s “mega-cap tech” grouping. At least 90%+ of the analysts’ recommendations were “recommend”, “buy”, “strong buy” or, at worst, “hold’.
We are reminded of legendary investment strategist Bob Farrell and his iconic “Ten Rules for Investing.”[6] As it pertains to the overwhelmingly positive consensus regarding mega-cap tech stocks, we focus specifically on Rule #9:
Rule #1: Markets tend to mean revert over time
Rule #2: Excess moves in one direction will lead to an opposite excess in the other direction
Rule #3: There are no new eras – excesses are never permanent
Rule #7: Markets are strongest when they are broad and weakest when they narrow to a few blue-chip names
Rule #9: When all the experts and forecasters agree – something else is going to happen
Let’s also illustrate Rule #7 – Markets are strongest when they are broad and weakest when they narrow to a few blue-chip names.
We certainly can make an argument that that the performance of the S&P 500 index over the past few years, driven almost entirely by the mega-cap tech stocks, is the embodiment of a “narrow” stock market.
Source: Yardeni Research, as of October 3, 2025. You cannot invest in an index and past performance is no guarantee of future results.
What does “priced for perfection” mean? One definition is that it refers to investments that carry a high valuation based on optimistic expectations for future earnings. That certainly is true for today’s mega-cap tech stocks.
Another definition, however, is that current valuations are based on an assumption of only positive outcomes, such as interest rates, economic conditions, investor appetite and sentiment, and minimizing the risk of a technological “leapfrog” that may create seismic shifts in the current landscape.
The release of the DeepSeek news early this year, combined with the “Liberation Day” tariff announcements, is an example of what can happen to “perfectly priced” companies when events go “off script”, even if that particular phase proved to be fairly short-lived (see the above chart of mega-cap tech performance from January through April of this year).
Summary and Interpretation
We are not Chicken Little, and we are not suggesting that the AI “sky” is falling. These are quality companies at the cusp of a generational quantum leap in technological and productivity advancement.
But we’ve been here before with railroads, telecom, energy, and the internet, to name a few other generational quantum leaps.
What we witnessed in each of those eras was a period when excitement and supply dramatically exceeded demand (which took a while to catch up) – and the unfortunate failure of many companies along the way.
So, we are not suggesting in any sense abandoning AI-oriented or large-cap growth stocks. This asset style belongs in any diversified portfolio.
We simply suggest that advisors and investors be forewarned that there are always hiccups along the way to the future, and to build your portfolio accordingly.
At the risk of being repetitive, we continue to recommend diversification, patience, discipline, and a focus on a longer-term time horizon.
As always, we welcome your questions and feedback.
[1] Pascal’s wager – Wikipedia
[2] “The Financial Instability Hypothesis”, by Hyman Minsky, Working Paper #74, prepared for The Handbook of Radical Political Economy, edited by Philip Arestis and Malcolm Sawyer, Edward Elgar: Aldershot, 1993. (http://www.levy.org/pubs/wp74.pdf).
[3] Pangloss Character Analysis in Candide | SparkNotes
[4] Why is there an earthquake in Candide? | Adventures on the Bookshelf
[5] The AI boom’s reliance on circular deals is raising fears of a bubble