Business investment has been a key driver of S&P 500 profits and performance over time (chart below). The story has two layers: a long-term structural trend of labor substitution through capital — a process underway for decades — and a sharper cyclical surge driven by AI investment today.

Source: WCA, Bloomberg

Over the past several decades, business investment in equipment and intellectual property has grown nearly twice as fast as the overall economy. But that long-run trend masks pronounced cycles, and history offers two useful reference points.

In the late 1990s, networking capital spending pushed business investment up roughly 14% per year from 1996 to 2000 — about double the economy’s growth rate. Corporate profits doubled as a share of GDP, rising from 5% to 8% between 1995 and 2000. Confidence in the future was running high. Then in 2001, capital spending slowed abruptly and unexpectedly; profits fell back to around 6% of GDP (a 25% relative decline), and the NASDAQ dropped 85%. Previous investments were questioned, and write-offs followed. The best opportunities for investors turned out to show up during the shakeout, not the run-up.

The mid-2000s repeated the pattern. Confidence returned, investment firmed, and profits rose again as a share of national income. Business investment reached about 8.5% of GDP by 2006, and profits advanced 70% through 2007 to a record 10% of the economy — up from under 6% in 2002. The five-year cycle ended suddenly when bad mortgage loans unexpectedly triggered the 2008–2009 financial crisis, and the S&P 500 fell nearly 60%. Again, the best opportunities came during the pullback.

With the brief exception of the COVID recession, we are now in an unprecedented cycle favoring capital over labor. U.S. public equity markets just hit $77 trillion, and private equity could add another $10–20 trillion. At a combined $87–97 trillion against roughly $32 trillion in national income, the value of equity capital markets relative to the economy is without precedent.

Much of this value creation traces back to a growing share of national income flowing to businesses, and much of that profitability traces back to business investment in equipment and, increasingly, intellectual property. The mega-trend reignited after the 2008–2009 crisis, continued through COVID, and has accelerated further with the AI boom that began in 2022–2023. Trends in S&P 500 net profit margins tie closely to business investment as a share of national income (see chart above). The glue that holds this whole process together is a sometimes fickle ingredient: confidence.

Trillions of dollars are now committed to building AI capacity, much of it through massive data centers and through complex revenue-sharing and ownership arrangements between partners and customers. A 2025 McKinsey study estimated that roughly $5 trillion will be invested globally to build out AI capacity by 2030*. Some of this will generate substantial profit. Some will not. As markets evaluate which is which, the process is likely to be volatile — as it was in past technological cycles.

The valuation picture reinforces this concern because it reveals very lofty expectations. To measure how lofty expectations are, we might look at implied growth rates at today’s prices. For example, by subtracting a company’s operating cash flow yield from its weighted average cost of capital, one can derive the market’s long-run implied growth rate for each stock. Many companies clustered around the AI trade currently trade at implied growth rates in the mid-to-upper teens — two or more standard deviations above their own 10-year averages. The question this raises is straightforward: can this many companies, simultaneously, meet expectations this elevated? Some will. But if investment moderates — as capital spending cycles invariably do — meeting these expectations will be a tall order for the cluster as a whole.

We must also point out that all the companies driving the market forward in recent weeks have been advancing on the exact same narrative. Every one of the top ten companies that have driven the S&P 500’s recent advance, for example, is directly tied to the technology / artificial intelligence theme (chart below). While long-run implied growth rates for these 10 companies have risen with the AI buildout, it is important to remember that they cannot grow faster than the economy into perpetuity. Conversely, other sectors of the economy have been neglected in this cycle. Where are the consumer companies? Or healthcare companies? Or industrials? These are conspicuously absent. The broader market has been largely left behind. As the chart below shows, 75% of the S&P 500’s recent gain has come from ten stocks — all of whom are advancing for similar reasons (AI investment). By the way, most of the other 25% of the market’s recent gains are also directly or indirectly attributable to the same theme. When one narrative theme is overwhelmingly responsible for moving the S&P 500, it is hard to consider an investment in “the market” as “well diversified.”

Instead, the S&P 500 clearly becomes a one-sided, narrowly focused, increasingly risky bet.

We remain bullish on the long-run productivity potential of AI and automation. But capital-spending-driven profit cycles can reverse abruptly, and they have historically ended with periods of disappointment before the next leg of productivity gains arrives.

This is why we stick with quality at a reasonable price. The best opportunities in past cycles revealed themselves during pullbacks, not during the run-up. Maintaining discipline on quality and price gives our strategy a good chance of delivering the steady returns and rising income our advisors and clients expect. Abandoning that discipline exposes portfolios to excessive and uncompensated risk. Applying risk management strategies through a cycle like this is not about giving up gains — it is about positioning to take advantage of quality and value when prices properly compensate for the risk taken. Additionally, risk management strategies are about cushioning the inevitable drawdown that will likely occur when the theme runs its course.

* McKinsey Quarterly: The Cost of Compute: A $7 Trillion Race to Scale Data Centers

Kevin R. Caron, CFA
Senior Portfolio Manager
973-549-4051

Chad Morganlander
Senior Portfolio Manager
973-549-4052

Steve Lerit, CFA
Head of Portfolio Risk
973-549-4028

Eric Needham
External Sales and Marketing
312-771-6010

Matthew Battipaglia
Portfolio Manager
973-549-4047

Jeffrey Battipaglia
Client Portfolio Manager
973-549-4031

Suzanne Ashley
Internal Relationship Manager
973-549-4168