Head in the Clouds, and Liking What I See

Fear seems to have not gone completely out of fashion after all, and I’m blaming the Bush Administration. In trying to make sense of the unrest sweeping through various and sundry of the non-developed world’s hell holes, one cannot escape the very real possibility that sowing the notion that “freedom is not only your desire, it is your destiny” will, in time, produce a crop of all kinds unpredictable outcomes. Now it is hard to imagine that Libya ends up a more benighted place than its been for the past, oh, hundred or so decades, but it is possible. That possibility, that clutch of uncertainties that a month ago did not rise to a level where they entered our attention span, is what has spooked investors in these waning days of February. None of the countries that have thus far experienced this upwelling of repressed humanity carry much weight in terms of the global economy, but we are all haunted by the very real uncertainty that we don’t know how far this goes before it runs its course. (Mr. Bret Stephens of the WSJ has his finger on this in a 2/22 column explaining why the protests in Bahrain are a “shot across the bow” for Beijing.)

President Bush et al were absolutely right that people everywhere yearn for freedom, for at least a modicum of the dignity of personhood as opposed to de facto chattel. This yearning has morphed into a contagion, opening up possibilities which may in the long run end up very positive, but in the short term are thoroughly inscrutable uncertainties. It is quite unlikely, in my opinion, to really catch on in China, at least not this time around, but we all know in our guts that the grievances of Libyans, Egyptians and Bahrainians are the grievances of many, many Chinese. Freedom matters. Throw in the specter of elevated energy costs smothering a still tenuous recovery in consumer spending, and it gets a lot easier for worry to drown out greed. (No wait, never mind. “Volatile food and energy prices”, despite their importance in households near the margin of viability, aren’t counted in the monthly inflation figures, which is why “we don’t have inflation.”)

Events like these rat hole countries being turned upside down are so hard to handicap its really not worth trying. I suspect that this episode has been going on long enough that whatever countries are going to boil up have already started to do so. The substantive threat of Libyan oil production being interrupted should be understood in terms of Libya being ranked #18 in production, between Angola and the UK, and plenty of spare capacity among its larger, less dysfunctional OPEC brethren. I would favor the odds that this crisis du jour is behind us in a matter of weeks, and greed finds its way back into the driver’s seat, but with matters like this, you never know until its over.

The prospects for a meaningful pullback in share prices definitely has our attention, but lately the biggest focus for my attention has been a Market sub-sector that has not needed a pullback. In the last Musings, I briefly touched on what seems to me to be excess pessimism, as expressed in excessively compressed valuation, of the hard disk drive (HDD) sector. In this edition, I will take issue with the apparent basis for this pessimism. As previously noted, the HDD industry has consolidated down to just five producers. It is dominated by Western Digital (WDC) and Seagate (STX), which together have about two-thirds of the market. The other participants include Hitachi (about 18%) and increasingly tenuous positions by Toshiba and Samsung, the two companies that, not coincidentally dominate that emergent form of data storage that is NAND memory. It is the specter of this newer, in some ways superior technology that more than any other factor is scaring investors away from the HDD stocks. We are being told that “The HDD is dead!”, much the way the emergence of disk storage elicited cries that “Tape storage is dead”, about twenty years before tape based storage actually quit growing. Now it is a pretty safe bet that HDD storage will eventually be supplanted by something else and go into decline, but based on what is knowable at the moment, it is NOT going to be because of NAND based solid state storage devices (SSD) and not in the next five years at least. The HDD makers might struggle with economic jitters and their PC maker customers trying to sort out just where tablet devices fit in (it’s always tough to tell how much of the initial demand is due to the fashion statement some gadgets attain), but a look at the physics of storage technology tells us that the SSD threat is greatly overblown. Indeed, what I see is two very complementary technologies that are going to meet the data storage needs of a billion or so “personal clouds” for a good long time to come.

The purported rivalry between HDD and SSD comes down to the same as issue for any competing mass storage media, that value proposition of being able to deliver more utility for less cost, over time. (Note the mass, there are specialized applications where other factors trump value.) An increasingly digital economy has an ever growing number of households generating, editing, sharing and storing digital content, growing at a rate of at least 40%. The only way to keep up with this, for it to be possible at all, is to make mass storage steadily and grossly cheaper. This has been accomplished primarily by increasing the areal density (AD), the number of bits of info that can be stored in a given area, of storage media over time. Throughout its fifty year life, the HDD industry has been able to keep up, albeit in a discontinuous manner of rapid breakthroughs followed by interludes of deceleration. Silicon-based NAND storage has come along more recently, riding the lithography advances that have as much as anything driven Moore’s Law. NAND memory reads much faster than HDD and has no mechanical parts, but among other disadvantages is, so far, much more expensive on a per gigabyte (GB) basis in all but that smallest capacity (e.g., smart phones, cameras) applications. The bear case for HDD seems to rest on the premise that if you plot the cost per Gb over time, the lines say that eventually SSD is going to catch up with HDD.

We all know that extrapolation of recent trends into an indeterminant future can get us into all kinds of trouble. What no one, save a few folks I have spoken within the industry, seems to understand is just how untenable such extrapolation is when it comes to AD. It comes down to the relative ability of silicon versus the ferritic material used in HDD to reliably hold a charge. Both media can be thought of as containing unthinkably large numbers of cells that can be manipulated to hold an electrical charge in a way that denotes the 0’s and 1’s of digital info. It turns out there is a world of difference between what you can do with the two media as it gets more crowded (the technical term seems to be “coercivity”). Silicon is a lot “slipperier” than ferritic particles, especially when you crowd the cells with the multi-level configuration that has been the primary driver of improved price/performance. Lose charge from a cell when it is supposed to stay charged and you have bad data. The closer together, more likely it is that one cell will alter the charge on another. So while the lithography to shrink the node sizes in silicon still has miles to go (see Cymer), it is increasingly doubtful that memory, which is all about reliability over time, will be able to continue to reap the benefit of this the same way logic chips will. The HDD makers are having a harder time advancing AD too, as the breakthrough that was PMR grows long in the tooth (from 40%+ to about 25%), but their issues are no where as fundamental as it is going to be for SSD. There are grave doubts about being able to use NAND to produce reliable (as in, “Dude, where’s my data?”) storage at sub 32nm with the multi level schemes needed to bring the cost down at a rate anywhere close to what HDD can do.

So what should we expect? The HDD makers are presently under earning (though contra past cycles, the leaders are remaining profitable), which is more than anything a result of last year’s aggressive pricing. This will subside as a function of new product introductions later in the year. Unless something derails the way Web 2.0 is enabling more and more people to generate more and more data, the volume growth is there to soak up whatever excess capacity developed in Q2 10. SSD will grow as well, but reliability concerns and high cost will greatly limit whatever effects its growth has on HDD. Volume and new product driven price improvement should be the basis for earnings momentum, probably in H2 11 and all but certainly in 2012. From current levels, there is, at the very least, a decent trade in these stocks, just waiting to be made. However, I believe that it is finally time in the evolution of the HDD industry for investors to consider the possibility of a much more protracted and remunerative involvement. I have considered long and hard what the future, five or so years out, is likely to hold for these storage technologies, and it looks to me like these survivors of the 40+ year fight that got them to today have a very promising future.

First of all, there will probably only be three or perhaps four HDD makers. Everyone I talk to in the industry seems to agree that the consolidation among drive makers is not over. The companies that are here five years from now will be providing lots of disks to the “public cloud” providers. SSD will have a piece of this too, but only a piece. HDD will also be the backbone of a whole bunch of personal clouds. Like maybe a billion or so. Some of these will manage all their storage needs by “renting” space on HDDs in a public cloud, a few will keep it all on their own HDD. One suspects that most will being doing both. That’s the richest billion or so households. Another billion or so are likely to have become users and owners of digital content. It is unlikely that their storage needs will be met on terms resembling what we now think of as a data plan (“Hey, I feed my family and my cattle for that much!”). Not when $50 buys a multi Terabyte storage device that last 5+ years. It also quite possible that a preponderance of the devices tethered to the storage backbone of all these personal and corporate clouds will be able to get by with the modest amount of storage best provided by SSD, but many will not. What I envision is a hybrid solution with a value so compelling that it will be hard for all but the most casual users to say no to it. Think an array of SLC NAND big enough (32GB?, 64GB?) to handle the boot up and otherwise mimic an SSD (except for the folks who think they can notice the extra 0.05 second seek time when they go to view a photo) and a single platter providing several TB of mass storage, profitably delivered for $50. The ubiquity of mobile drives will of course be affected by how seamlessly and at what cost device owners can synch into their personal or corporate clouds (don’t expect an absence of bumps in that road!). In any case, the number of drives will continue to grow for many years to come, even as the number of companies that can make HDD dwindles.

The best way to invest in this future is probably WDC. Look at how far this company has come in ten years. About the most perplexing thing I have seen in the investment world of late is that a company that has grown like WDC has, made a profit in every quarter since 2002, including the great global freeze-up that was late 2008, can sell at only five times last year’s earnings. And the company literally has an embarrassment of riches. Having piled up over $3B in cash, management is trying to figure out how to solve this “problem”, which was inconceivable only a few years ago.

Based on what I was able to understand when I started studying HDD 5+ years ago, STX looked like the better company to me, but I was wrong. (I did not understand the margin impact of the Enterprise segment, or that WD had been working its way out of a deep hole.) It got to be a large position for me by virtue of the great price that developed at the bottom of the Crash, when this fool rushed in with both hands. Over the past three years at least, STX has disappointed everyone, including themselves, and needs to show the world that it has not squandered its once vaunted technological leadership. As a value investor who has made more on “things turning out not as bad as they seem” than on the “wonderful getting more so” so many growth chasers seek after, STX just might have more juice, but for the superior, way more than ample, “heartburn adjusted” rate of return, WDC would be not only the better of the two but one of the better stocks one could own for the foreseeable future.

And speaking of “heart burn”, etc., I also have a large position in component maker Hutchinson Technology. After a strong comeback in 2009, HTCH had a horrible 2010, as the challenges of a new manufacturing process exacerbated the effects of the volume slowdown that characterized H2 10. These challenges have been resolved, and while still not back to profitable, HTCH is on the road to recovery. It could take a couple of more quarters for the necessary volume recovery to accelerate in a way that returns them to profitability (though positive CF is closer at hand), but when once it kicks in, their product line-up is such that the ramp should be rapid and enduring. It is also intriguing that as HDD makers grapple with a difficult and expensive “materials” solution to diminishing AD gains from PMR, HTCH is the most capable and dependable provider of “mechanical” solutions. Not for the faint of heart, but as I have noted in the past, there is a basis for that “stock of a lifetime” appreciation in HTCH. That is, if my vision of “personal clouds” five or so years from now is even remotely on target.