- There are many interesting tech companies which Warren Buffett doesn't touch.
- I show why one such company, SanDisk, seemingly very profitable and not very expensive, is still avoided by Warren Buffett.
- The problem has to do both with cyclicality and the risk of obsolescence. I show one example which could put SanDisk at risk in the not so distant future.
SanDisk (NSNDK) is a nice, profitable company. On top of that, with the massive adoption of smartphones and tablets, flash memory has seen tremendous demand. It's thus not much of a surprise that SanDisk stock did magnificently over the last 5 years.
So why does Warren Buffett not buy into companies like these where the favorable trends are so obvious and strong? There are two main reasons.
First, these companies are usually cyclicals
Demand for memory might be ever-growing except in rare, deep, recessions. But the provisioning of capacity to build this memory is not very granular. The factories are huge and the product is commoditized and sold on specifications.
This means that when the product gets very profitable, new capacity usually emerges in large chunks, and the market quickly goes from undersupplied and firm to oversupplied and chaotic.
That's not necessarily the case today, though it might be at some point in the future when the smartphone market matures. Still, the very situation is a bit alien to Warren Buffett, and it makes things hard for investors since when these companies are the most profitable and with the lowest valuation multiples is usually when they're the most dangerous.
Second, you can be obsolesced quickly
More importantly for Warren Buffett, he likes to look at his companies and be able to tell that they'll be around decades into the future. Merge this persistence with secular growth and a fair to cheap valuation when you buy, and it's hard to do badly.
Such characteristic, however, is not present in tech companies like SanDisk. You cannot be sure that SanDisk will be around 20 years from now. It's easy to show how for instance take this Engadget article, "Future phones could house a terabyte of storage":
You may think that the 3GB of memory in your new smartphone is hot stuff, but that pales in comparison with what Rice University has in store. Its scientists have detailed a form of resistive RAM (RRAM) that can be made using regular equipment at room temperatures, making it practical for everyday gadgets. The trick is the use of porous silicon oxide where metals (such as gold or platinum) fill the gaps. Using the silicon material doesn't just give manufacturers something familiar to work with; it requires much less power than previous techniques, can last through 100 times as many uses and isn't fazed by heat. It's also far denser than earlier RRAM, storing nine bits per cell where even conventional flash storage stops at three. The result should be an easy-to-make RAM chip with the kind of capacity that you'd normally expect from much larger permanent storage, like an SSD - as the company Crossbar hinted when it first discussed this approach, you could stuff 1TB into a component the size of a postage stamp.
If this RRAM fulfills its promises, the kind of memory SanDisk produces will instantly be obsolesced. Furthermore, in the chaos which would follow, several new memory producers would emerge, and someone else would have the intellectual property to this new RRAM. It is likely that such a technological development would see SanDisk as a victim.
So there you have it, even a few years into the future there's massive technological risk for companies like SanDisk. This alone drives away Warren Buffett from them. And it ought to lead sensible investors to require much lower earnings multiples as well, given the risk. Yet, at present SanDisk trades for 13.8 times 2015 earnings estimates.
While many might find SanDisk interesting as an investment, its intrinsic qualities will forever keep investors like Warren Buffett away. Both its cyclicality and technological risks are potentially massive.
If I were to be long SanDisk, I'd keep an eye on the RRAM development. It looks like a potential asteroid if it comes to fruition.