The index holds up while individual stocks fall sharply
AI and semiconductor stocks, which had risen since the start of the year, now include many names that are down 30–40%. The broad index, by contrast, has not fallen much. A handful of mega-cap stocks are holding the index up while many individual names trade below their January levels, and the gap between the index and its constituents is widening.
Profit-taking is common at this time of year, and valuations have also come down from overheated levels toward their historical averages. Because earnings forecasts themselves remain intact, we think it is closer to the truth to say that the price investors pay for the same earnings has changed, rather than that earnings have deteriorated.
The constraint is funding, not technology
AI capital spending is long-term investment with a broad base: data centers, power, cooling and the grid. Some generating equipment and large power components have lead times of several years, so even when construction goes to plan, it takes a long time to come on line. In several places, local opposition has also halted construction.
Who provides the money in the meantime, and on what terms? In the early phase, large companies with ample cash funded the build-out themselves. As the scale has grown, reliance on loans and bonds has increased. For newer companies that specialize in operating data centers, funding conditions directly determine whether the business can continue.
Competition is hard to stop, so rates set the pace
The capital spending plans of the large players keep being revised upward. Because any company that eases off risks dropping out of the race, each one keeps investing even though it knows the risks — a prisoner's dilemma. We think the race itself is likely to continue until winners and losers become clear.
Since competition is unlikely to stop, what ultimately sets the pace of investment is the price of money — interest rates and credit conditions — more than technological progress. When rates stay high and spreads widen, the difference between companies with strong balance sheets and those that depend on borrowing becomes clear. Whether a company's leverage fits its size is an important yardstick when looking at individual stocks.
Implications for asset allocation
These are general considerations, not recommendations of specific securities.
- Consider longer-dated bonds as an option If a slowdown in AI investment leads to a broader economic slowdown, interest rates tend to fall. Adding longer-dated bonds is one way to soften the impact of equity declines.
- Act according to your equity weight For investors who mainly hold bonds and have a low equity weight, a correction can be an opportunity to add equities gradually. Investors with a high equity weight, on the other hand, may want to review their allocation with protection in mind.
- Do not concentrate in one issuer's bonds We sometimes see portfolios that hold only the higher-yielding bonds of a single well-known company. A high yield is the flip side of credit risk. Diversifying across issuers, currencies and maturities is the basic rule.