Re: Informed bear
With over half a lifetime of experience dealing with adversity (among other things: held up at gunpoint, struck by a car, stuck a couple of hundred miles offshore in a 45-footer during a Force 9 storm, dealing with my ex-wife for over 10 years), I really can't complain about some of the responses my posts have engendered. So if some want to fall back on emotion rather than engage in somewhat higher level discourse, so be it (patently transparent attempt to seize the high ground!).
About this supply and demand balance thing - it probably doesn't exist. But what has been going on now is really perverse: OEMs don't want to hold inventory at the same time that the DRAM market has become more differentiated. This is a pretty potent brew, and it's no wonder that some prices go through exaggerated swings (e.g., 16Mb EDO parts). How in the world can DRAM manufacturers be expected to divine the whims of the customer base? Of course, you can claim that they have only responded to market pressures, but that goes back to my earlier statement about commodity suppliers reacting more than anticipating. C'est la guerre.
About this bull/bear thing - I believe that some of the problem on the thread may be due to people confusing their feelings about the company with their take on the stock (we'd all like to buy Cisco at 8 times this year's EPS, but situations like this seldom occur during waking hours). The DRAM market is unruly, and Micron has, until recently, been a perceptible laggard to many of the leaders in this market. This opens the door to a wide dispersion of perceptions about the industry and company fundamentals, which probably accounts for some of the more vehement posturing on this thread.
Another contributor to dissonance is the rate of change in the DRAM market in terms of technology, manufacturing capacity and market shares. This leads to us having multiple perspectives (past, present and future for just those three variables creates nine separate considerations, and we're just getting started).
My point in several of my posts is that there appears to be a lack of widespread awareness of some of the "easy" stuff, particularly past and present market shares. At the very least, MU investors should have a grasp of some of the more reality-attached data. Then a discussion of the future, which is intrinsically more difficult, has half a chance of starting off correctly.
Now, as an "informed bear," two points/observations:
1) Skeeter's comments about rapid ramps in CapEx appear to be correct, but only to a point. As a former equipment analyst (at H&Q and Bear, Stearns), I pay close attention to equipment spending. In fact, I operate on the assumption that the typical viewpoint is 90-degrees wrong - semiconductor industry performance doesn't affect capital spending nearly as much as capital spending sets the tone for semiconductor industry fundamentals (it's not 180-degrees wrong because the relationship is extremely symbiotic). And yes, Skeeter's right, the data does indicate that equipment purchases are trending up. But it's important to keep in mind that this upward trend is off a bottom last summer/fall that was a pretty good imitation of nuclear winter. Even with the recent increase, CapEx appears to be well below what should be considered norms.
And that's good, because it means that the engineers' ability to continue to shrink designs is becoming limited. Eventually, this reduces supply growth and leads to firmer pricing and better returns on investment.
2) Skeeter and Earlie make some good points about PC growth, which leads to a really key question: if DRAM pricing reflects tighter supply over this cycle (particularly next year and 2001), will OEMs let their DRAM budgets increase (Carl R's question)? After all, with DRAM bit growth running at 80-90% for the last several quarters in the face of 15% unit growth in PCs (with a definite skew towards the lower end price points that utilize less memory), it's apparent that the DRAM per box has been trending up at a pretty hefty rate. Assuming that we are moving rapidly to the 64MB minimum (only the cheapest PCs I see offered have 32MB), won't most OEMs freeze their DRAM/box in the face of firming DRAM prices. I find it hard to believe that DRAM budgets could move up in the face of continuous PC price declines.
There may be an out to this situation. Throughout 1997 and 1998, much of the price declines in the PC market were accommodated by component cost reductions ex the MPU. It was only towards the end of last year that microprocessor pricing began to decline. If turnabout is fair play, perhaps we are entering a period when declining MPU costs offset pricing firmness for components such as DRAM.
Personally, I have a hard time buying into such a scenario, but that's a bias on my part due to the accumulated baggage of history. This happened, to a certain extent, in 1995, and the response of OEMs was to ship PCs with a whopping 4MB of memory knowing that the buyers would have to go into the aftermarket to bring their machines up to minimum memory levels sufficient to run Windows. But as current machines are starting from a much higher memory configuration, I don't foresee the same aftermarket effect.
So,
I believe that there is a major upcycle in the semiconductor industry lurking out there and that memory in general and DRAM, specifically, should outperform the semiconductor industry in such an environment. It appears that the drivers for this cycle are likely to be different from 1995-95. As always, the key questions remain timing and magnitude. Then the question becomes the extent to which Micron participates in the upcycle and is able to take that participation to net income. All of us involved in the stock should be spending our energies on these questions, which is made easier if the starting data is agreed upon.
What offsets much of my enthusiasm for this scenario is the recognition that the consolidation of the DRAM industry is a reflection of the duress of the downcycle, and that a sustainable upcycle will bring in additional suppliers as well as incremental supply from the existing vendors. This is the nature of the beast, and investors who ignore it generally get to pay for their education. It may well be that the determinant of the next cycle is the equipment industry's ability to respond to a ramp in demand for additional tools. - Tad LaFountain |
