|
Amir, I would buy a 1/3 put position on Xylinx and Altera right now
if I didn't already have a position. With Micron, it was simply a
case of having lived through several semiconductor cycles, having
visited Asia, and knowing that the funddumbs had the "semis are not
cyclical any more" argument 100 percent wrong. I really wasn't looking
for a big downturn in semi demand, though Waterwindows 95 not being
able to match the hype was a help. My main key was that everyone was
announcing capacity additions in Dram. And much more capacity was
being added on the sly, without an announcement. Also, yields at
currently operating fabs were going up. The last straw was that the
Asian fabs were not 4 Mbit, they were 16 Mbit. So, you were getting
a lot more chips, but geometrically more Mbits. When I posted in
December
on
Prodigy that Samsung and NEC would be producing 23 million 16 Mbit
chips by the end of March, a gizmo nerd wrote to tell me that I was
lying and that was totally impossible.
As it turned out, they produced more than that and the industry
was simply not ready for the supply that hit all at once. Every
chipmaker was low-balling his own production to try to keep competitors
from building more. Entire countries, like Taiwan and Singapore, were
getting into Dram. TXN said that there was no chance that enough DRAM
capacity could be built in 1996 to meet demand. Knowing their record
of totally wrong market projections,
MU puts were a layup. So, really, though I hoped demand growth
would decline
to speed the process, and it did.
I think supply would have killed Dram prices all
by itself. Supply is the key to all the semiconductor prices right
now. There is too much of just about everything and capacity is
being built at a quick pace in everything not in glut right now.
Good luck, MB |
