I think that the lifespan of 16M, 64M, 256M will each be small.
The next four years (and more) are going to extremely challenging.
Companies will have to be really creative to recoup their overhead.
Building of fabs will almost be like a game of chess, with moves
and countermoves. Contrary to the way it seems, converting 4M to 16M
is going to work in Micron's favor - their next fab will be 64/256M.
To give you an idea of how short the lifespan will be, follow the
Samgung link from scc.net. Samsung has
announced 1 Gig SDRAM, engineering samples in 1997, mass production
after the turn of the century - Now that is just 4 years (max 5 years)
to recoup cost of fabs.
4-5 years is not enough time to go thru all the generations
(16,64,256 M) of fabs. At the same time a 64M plant now will be
practically useless. 16M is OK for now, will probably last thru
1997 when PCs may have average 128Mb (avg now is 12Mb).
Fortunately, people get paid big bucks at Micron to worry about
things like this - and they do it for a living ! |
