One of my early bosses (three levels up from me as a portfolio manager) was very fond of saying, “Trees never grow to the sky.” It mean, of course, that even the most powerful upward impetus for a given stock has its limits. Where that upper bound may be is not set through a flat-out rule–or at least not by any rule that I know of.

When you find a stock that you don’t own yet, but which your analysis says has incredible growth potential, you typically look for something to sell in order to buy it. The target “something” may be your largest position (that you’ve allowed to run past your normal position size limits), or a stock that has already had a very good run and is closing in on a sky-high valuation, or the inevitable clunker that your eye somehow skips over when running down the list of your holdings and performance.

Sometimes, though, even if the position size and surprisingly good recent performance are telling you to trim, you don’t do so. For me, it’s partly that I don’t like to have a lot of cash in what is intended to be a portfolio of stocks and partly that I haven’t found that next good thing. It could also be that I haven’t seen a sharp signal that the situation that is allowing the company to make extraordinary profits is about to change.

My guess is that we’re seeing this kind of warning signal now in the memory chip business. The basic story, as I understand it, is that adding capacity in memory chips is extremely expensive, takes several years to come on line and is the domain of a small number of specialized semiconductor manufacturers. As larger chunks of this relatively fixed capacity are directed to making AI-related product, a shortage of regular old DRAM has developed. Prices have tripled or quadrupled and new capacity now being planned is a long time away.

…but Apple has just indicated that it is exploring using Chinese-made chips in its cellphones

This may come to nothing, although the fact that it’s Apple and that the company has announced this publicly seems to imply that Chinese output is at least a plausible substitute for US or Korean chip offerings. It’s hard to know how all this will play out, however. In the meantime, I think this possibility will act as a drag on all the DRAM stocks. (I’d been scaling out of MU before the Apple announcement and have since sold the rest of my position), In an adjacent arena, NVDA has been pointing out for some time that denying China access to the most advanced AI chips will ultimately backfire, because it will simply incentivize that country to develop its own substitutes. Still, one can see how the trajectory of NVDA shares has flattened since this development. My guess is that the same combination of strong recent performance and new questions about the underlying story will act as a brake on MU.

For what it’s worth, I also think that the amazing strength of IT last year and this is in part the market reaction to government policies that, intentionally or not, mimic those of developing economies. The tried and true formula in that arena is to emphasize export-oriented manufacturing and avoid the domestic economy. My sense is that Wall Street is shifting away from that stance. Whether this is stretched valuations or anticipation of future change isn’t clear. But that’s what current price action is telling me.