Semiconductor-chip stocks have surged since late March, with the VanEck Semiconductor ETF gaining about 55%, names like Micron Technology and Advanced Micro Devices more than doubling, and Intel nearly tripling. Nvidia, which reported better-than-expected earnings Wednesday, has gained about 22%. The rally doesn't seem to be a redo of the dot-com craze: It has been fueled by real AI-driven demand for semiconductors and strong earnings results. But has the chip group gotten over its proverbial skis? For this week's Barron's Advisor Big Q, we put the question to a panel of investment pros.
Lisa Shalett, chief investment officer, Morgan Stanley Wealth Management: When you look at the money flows and the technicals over the past six weeks, a move that has us up nearly 60%, you're looking at the type of behavior that is analogous to 1998/1999. That doesn't mean we necessarily are calling for a peak. But what we are saying is a lot of people have made a lot of money in these names, and they are historically extraordinarily cyclical. So we are encouraging folks to continue to trim their positions as a share of the index.
This move has taken the semiconductor sector up to over 15% of the total index now. That's not saying semiconductors aren't important. But it's highly unlikely that they are going to be able to continue to grow that share. When we say to take profits in semis, we're encouraging folks to rotate to other parts of the AI supply chain. This might be an opportunity to add some of the more staid, secular growth parts of the story that may hold up a little bit better over the course of the summer.
Darrell Cronk, chief investment officer, wealth and investment management, Wells Fargo: Short term, we are probably overbought and overextended, by almost all the technical indicators. That said, the driver hasn't really been sentiment. It's been a reset of earnings estimates. As long as the rally is earnings-based, not multiples-driven, it could have some legs. I do think on that overbought condition you'll get some settling back, and I would probably use it as a buying opportunity.
Coincident with the earnings estimates going up, it's incredible what's happened with the capex expectations for the five largest hyperscalers. A year ago those five hyperscalers were supposed to spend $450 billion in 2026 and $450 billion in 2027. Now they're going to spend $800 billion in 2026 and $1.2 trillion in 2027. In my opinion the biggest risk is, can they effectively spend the dollar amounts they're committing to? It's not all about just having the cash on the balance sheet to spend and the cash flow to do it. It's also about having the industrial capacity to build. It's about having the energy to power compute.
Matt Morse, chief investment officer, Grimes and Co.: We have meaningful exposure across the AI ecosystem and semiconductor chips. At the margin we are de-risking and diversifying: We've been adding to stocks that are uncorrelated to semis in the AI buildout in pockets of industrials, healthcare, and consumer, to name a few. We've been opportunistic in reallocating and diversifying across the semi ecosystem, versus prior years where we owned fewer names.
As of today, semi stocks, including traditional semis plus memory and hardware, comprise about 19% of the S&P 500. It's rare for an industry to be north of 15%, and the semis are now bigger than the industrial and healthcare sectors combined. Semis have outperformed the S&P by almost 60%, and on average they are about 50% above their nine- and 12-month price trends. When stocks are this far above trend, they tend to correct and revert closer to trend in one of two ways. One is price correction, and two is time -- the stocks don't do much for a few months, or until their longer-term trend catches up to the price.
Unlike the new-economy era of the late '90s, though, today's semis and tech leadership have exceptional operating leverage and free cash flow generation. So one could argue that things are not nearly as extended now as they were then. And what we've seen in the past three years or so has really been earnings driven as opposed to significant multiple expansion. We expect more volatility in the coming months, which we think will ultimately present some opportunities.
Jamie Williams, senior vice president, investments, Raymond James & Associates: The chip sector has seen parabolic upward momentum. Could this continue? Corporate demand coupled with a desire to not be left behind is present and so yes, the buildout pace could stretch further. However, risks are elevated and caution is warranted. It is difficult to place valuation in a rapidly developing environment. On the fundamentals side, several of the parabolic gainers have single-digit forward price-to-earnings multiples, which would imply further upside potential, or lack of belief that the momentum can continue. On the technical side, the underlying support levels are well off current prices, so any hiccup would lead to a significant correction.
We're still holding on to the core names, but some of the outliers that have made straight upward moves I would take some off the table. In the majority of accounts we're going to hold on to existing [core-stock] positions. With new money, I would buy some of the core names on pullbacks.