Software Scare Is a Wake-Up Call: Time to Diversify

• The software selloff exposed dangerous portfolio concentration, but it didn’t come out of nowhere — it revealed how overexposed many investors are to expensive technology names, driven more by fear of missing out than by fundamentals.
• The market rotation in early 2026 has reinforced a basic but often ignored truth: diversification remains the “only free lunch” in investing. Navigating the rotation doesn’t require genius, just discipline and valuation awareness.
• As AI challenges digital business models, companies with physical assets, predictable cash flows, dividends and durable moats — such as infrastructure, industrials, consumer staples, select financials and energy — stand out as better positioned in a volatile, high valuation market.

The huge selloff in software stocks in 2026 has shocked the market, but the only real surprise is that so many investors were caught off guard. Developers have been using Cursor and other AI tools to expedite writing code since 2023: was it such a leap to imagine that AI would start building software itself?

It is easy to throw stones after the fact; we certainly did not predict that the release of Claude Cowork would cause software stocks to swoon. But it was not hard to see this coming. Anthropic’s latest release was the match, but it did not light the fire.

Claude Cowork uncovered the big risk hiding in plain sight: many investors are overexposed to expensive technology stocks. Market concentration is no secret — many of us have warned about it for years — but, as long as the market rose, greed trumped fear.

Diversification is literally the first rule of prudent investment. Even before we’ve heard of the stock market, as children we’re warned not to put all our eggs in one basket. And diversification is more than just common sense — it is Nobel Prize–winning stuff! Harry Markowitz won the Nobel Prize in 1990 for his work on portfolio optimization by using fancy math to prove what we learned as kids: diversification works! Markowitz’s research famously led him to quip that “diversification is the only free lunch in investing.”

So why are most U.S. equity portfolios so phenomenally concentrated? Did folks forget that it is risky to put half their portfolios in technology stocks? Coming into the year at around 35x¹ earnings, did software stocks really screen so incredibly cheap?

Of course not. Every sophisticated investor knows the risks of concentration. Nobody was under the illusion that 35x earnings represents value. What really explains this reckless positioning is fear — fear of missing out. People watched AI stocks soar most of the year, and they didn’t want to miss out; professional investors were afraid to take the career risk of being underweight a rising sector.

Fast forward to February 2026 and we have seen a notable rotation (Exhibit 1). Those who are underweight software stocks — and overweight elsewhere — now look prescient. But it did not take a genius to be positioned accordingly; it just required that one remain diversified and measured in one’s exposure to very expensive stocks.

Exhibit 1: A Software Selloff Drives Market Rotation As of Feb. 26, 2026. Source: ClearBridge Investments, FactSet.

While it certainly feels bad out there, the S&P 500 Index is nominally up for the year. Investors have taken the proceeds from their software sales and bought other, more staid sectors of the market. What are these? Value continues to look cheap relative to growth, and high-quality dividend growers, in particular, stand out.

Dividend growers offer three benefits in today’s market: 1) dividends tend to limit downside in volatile markets; 2) they provide an offset when capital appreciation is restrained; and 3) dividend growth protects purchasing power amid persistent and pernicious inflation. Despite offering all these well-known advantages, dividend growth stocks are attractively priced relative to the broader market (Exhibit 2). They also have a history of outperforming during periods of stable or falling market concentration (Exhibit 3).

Exhibit 2: Dividend Growers Closing a Large Underperformance Gap

Exhibit 3: Dividend Growers Fare Well as Market Concentration Fades

This is not to say software stocks are bad — only that, in an environment of rapid technological change where some business models are easily called into question, emphasizing businesses with predictable moats is a sound strategy. Moats of non-digital assets, for example, are less subject to disintermediation than those of digital assets. Hard assets also offer a complementary margin of safety alongside valuation and often pay attractive dividends.

We have always emphasized investments in physical businesses because it is imperative that we be able to underwrite investments with conviction. This includes companies such as Air Products and Linde, which provide industrial gases, Freeport-McMoRan, a copper miner, and Vulcan Materials, which provides gravel for road building. We like consumer staples companies because we know that AI will not replace coffee, chocolate, laundry detergent or soft drinks. We like infrastructure businesses – cable, cellular towers, wireless companies, railroads, gas pipelines, utilities and waste — because they provide essential services that cannot be replaced by digital advancements.

Amid the selloff in software, many investors will seek out opportunity. Given the profound changes possible with AI, however, it may make sense to focus on opportunities adjacent to the software disruption. Several financials fall into this category. Alternative asset managers such as Blackstone and Apollo have traded down over concerns of their financing software companies — they own some software debt — but that is a fraction of their businesses, alongside real estate, infrastructure and private equity. Marsh & McLennan, an insurance broker, has also sold off on disintermediation fears, but given its vast, proprietary data and market position, we believe the risk of being displaced by AI is limited.

After being out of favor for many years, energy stocks have surged of late. Oil producers, like Exxon Mobil, have risen on geopolitical concerns. Natural gas infrastructure companies like Williams have soared due to the phenomenal power needs of the growing data center sector. It is impossible to forecast whether torrid semiconductor growth represents a sustainable demand trajectory or a bubble, but 10 years from now Williams will still be collecting its payments due to the long-term nature of its contracts.

We have begun to see a new continuum of businesses in the market: not so much value and growth, but what could be displaced by AI and what cannot. Judging by the market rotation we’ve seen so far in 2026, focusing on more stable companies seems a prudent course of action.

Given the huge quantum of money invested in technology, both sides of this trade could have further to run. For those unnerved by recent volatility, it is not too late to act. With market valuations near all-time highs and the world buffeted by countless risks (both geopolitical and technological), it is high time to pursue a more balanced approach.