Tech selloff amplifies the case for diversification

We’ve been asked one question more than any other in 2025. Is the AI trade a bubble? It’s a question that naturally calls for a yes-or-no answer, but it needs more nuance. In short, though, we don’t think this is a classic bubble — yet. We appreciate that there are some bubble-like characteristics but see significant differences between today’s market and the dot-com bubble of the late 1990s, for example.

Last week’s weakness in some large-cap US names has some investors questioning whether we’ve seen the peak in these stocks. We don’t think we have. But we do urge investors to think carefully about diversification today.

Resilience in the mega-cap space

In her book Grit, American academic and psychologist Angela Duckworth says, “Enthusiasm is common, endurance is rare.” There has been much enthusiasm towards the mega-cap tech names and the AI narrative in recent years. But there’s endurance there too.

The largest companies in the US today are engaging in enormous capital spending programs — and they keep expanding those plans. That might be a concern if they were funding that spending with debt and basing their plans on hopes for highly enthusiastic demand conditions in the future. That was the case in the late 1990s. Today, however, many of these companies cannot build data centers fast enough to keep up with current demand and are, for the most part, funding these projects from operating cash flow and cash reserves. Some companies are taking on debt, and some will likely struggle in the future. But broadly across the sector, leverage remains low compared to prior bubbles.

The multiples that these companies trade on are high by some historical standards but aren’t as high as during the dot-com period. The MSCI US Tech sector traded at nearly 50x 12-month forward earnings in early 2000.¹ Today, the same sector trades at about 32x 12-month forward earnings.² That’s because the performance of many of the largest US names has been closely matched by earnings growth. That wasn’t the case in the dot-com era.

Don’t fear the drawdown

The Nasdaq-100 Index® closed last week at 4.06% below its peak.³ The decline wasn’t universal across stocks. Microsoft, Nvidia, and Palantir fell, but Alphabet made a new high, and Apple was broadly flat.

Remember, drawdowns are a normal part of long-term investing, especially for stocks. Sustained stock market drawdowns do happen, but this drawdown is unlikely to be sustained, in our view.

For us to become more bearish, we’d need to see conditions that cause earnings to contract. So far, the third-quarter earnings season has been pointing to around 80% of companies beating consensus forecasts.⁴ Year-over-year earnings growth for the S&P 500 and Nasdaq has been running well into double-digit territory.⁵ Analysts are revising their 2026 earnings growth forecasts higher.⁶

Also, the Federal Reserve (Fed) is back in cutting mode, and we expect broader economic conditions to improve into and through 2026. Again, these aren’t the typical conditions for a deep and sustained stock sell-off.

Chumbawamba said it more colorfully than Angela Duckworth: “I get knocked down, but I get up again.” We suspect US tech stocks will potentially make new highs in relatively short order.

Diversification is prudent

Just because we’re not convinced that the AI-bubble scaremongers are right, that doesn’t mean we think this is the only place — or even the best place — to find opportunity. Rather, this pullback highlights our core message: It’s time, we believe, to diversify beyond mega-cap tech. We see significant opportunities in US cyclical, smaller-cap, and value stocks, and non-US markets.

Concentration can be exciting, and for a period, sometimes long periods, it can make investors feel like heroes. But in the long term, building resilient, well-diversified portfolios is essential. Now more than ever, we think that’s the case.