DRAM Diatribe

Taiwan Memory Corporation, the state bad DRAM bank, has chosen Elpida as their primary technology partner.  The release suggests Micron is still in the running to be included in the conglomerate, yet Micron’s technology is based on different architecture so this seems unlikely.  Besides Elpida, it is unclear which vendors will wind up part of Taiwan Memory.  Nonetheless, DRAM shares shot higher around the world with Elpida, Hynix, Powerchip and others closing up limit.  Spot prices apparently moved down.

DRAMexchange put out a forecast suggesting DRAM supply will grow 3% (down from prior estimate of 17.82%) and demand will grow 13.84% (down from prior 19.82%).  As a result, they predict shortages in the second half of 2009, though they say at present OEMs seem to be loaded with over a month of low-cost inventory, which must be eaten through first.  They speculate prices will rise to $1.20-$1.50 if the inventory reduction goes well.

NAND is expected to see a ramp of production in 2Q in anticipation of the second half consumer electronics build.  They don’t hazard a guess at supply/demand dynamics here.

The Skew:  Resolution of Taiwan Memory would lead to some visibility as it will have some kind of budget.  The counter argument is that even if they do, there’s so much idle capacity that can be brought back online that it won’t matter.  In the meanwhile, capital spending is on hold as players await a clearer picture of the competitive landscape.  Many semiconductor capital equipment companies are seeing levels of business that indicate a total freeze in the market place.  Taiwan Semi (TSM) is seeing utilization as high as 90% according to some press reports.  Without demand forecast increases by PC and cellphone manufacturers its unlikely these rates of recovery will be sustained.  If demand stays constant for another quarter, though, the foundries could be forced to actually buy some equipment.

Analysts and semicap equipment companies have been guiding for several quarters of dragging along the bottom.  The stocks trade mostly at under 2* book, having troughed close to book.  I generally favor buying semicaps on weakness as companies are aggressively cutting costs.  They’re telling you business will suck for the next year and a half.  I’ll tell you what sucks – semiconductor capital equipment vendor ability to predict the future – they’re the worst.  The longer their forecast gets, the better off you are betting against it.  They say several quarters of bottom dragging and that probably means you should buy the stocks.