Though the Nasdaq was roughly flat on Wednesday, the Philadelphia Semiconductor Index
Regardless of the exact factors, we're now at a point where chip developers and equipment makers with differentiated products, high gross margins, blue-chip customer bases and clear long-term growth drivers -- and which in some cases are buying back large amounts of stock -- can be had for 10 to 13 times consensus estimates for their 2023 fiscal years. Think companies such as NXP Semiconductors (NXPI) , Microchip Technology (MCHP) , Applied Materials (AMAT) and KLA (KLAC) .
The counterargument that semiconductor bears will make here is that 2023 sales and EPS estimates for these companies and their peers will have to be significantly cut, as the industry goes through a down-cycle. However, we've reached a point where even EPS estimate cuts of 25% to 30% arguably wouldn't make these stocks expensive, at least assuming a cyclical bottom is reached next year.
In other words, a standard down-cycle such as the one the industry saw in 2019 and early 2020 -- one featuring customer inventory corrections and some demand weakness here and there -- probably wouldn't be enough to make a lot of these companies expensive. We'd need something closer to what happened in 2008 and 2009, when a major recession caused significant demand destruction -- for that to be the case.
And for now, there's still little evidence we're poised to see a 2008/2009-type downturn. It's not visible in macro data such as employment levels or consumer spending figures, and it's also not visible in end-markets such as automotive and cloud data centers, where demand has been exceeding supply for a while.
With demand weakening in some chip industry end-markets, such as PCs and low-end smartphones, and with demand at risk of weakening in others, such as traditional enterprise servers and memory capital spending, there's a good case for being selective about which chip stocks one buys at a moment like this. But dismissing every chip stock that's fallen to single-digit or low-double-digit EPS multiples as a value trap, as some bears are doing right now, might amount to throwing the baby out with the bathwater.