I will respond with a copy of something mentioned to a private
investor group that I correspond with:
...
On another line, y'all may want to think real hard about shorting
Intel, with the idea that late this year we may see them
showing a significant slowdown. Not enough to do them in but
enough to take 'em quite a ways off of their current run rate.
There are at least five factors weighing in against them:
1) The battle for the eyeballs is very short on software and
doesn't need anywhere near the horsepower they are selling.
Even a lowly P133 with MMX can do serious video so their high
end chips are not required. It depends nearly 100% on effective
marketing.
2) A slew of competitors is now out there. National, IDT,
Cyrix, AMD and probably others are all doing both serious
direct competition in the high end and completely destroying
the low end. National has a P133 type that supposedly can
be sold for 30 bucks. If Intel wants to answer, margins will
suffer.
3) The patent lawsuit from DEC is for real. Recent patent
case rulings by the supreme court strengthen the
"doctrine of equivalents" which may pose a serious hazard
to Intel. DEC did get there first and their stuff is the
best and probably pretty critical to the internal operation
of the P6/P-II lineup. It will takes years to litigate but
DEC could actually either get billions in revenue or a seat
at the table with the HP/Intel/Merced project. In either
case Intel is going to lose some revenue.
4) Microsoft's recent accession to the validity of multi-user
NT with attached NC's. As the server grows in power and
networks grow in bandwidth, the need for local computing power
is reduced. It really does not makes sense for a business
today to put in a conventional PC based computing network for
things such as office automation, word processing, spreadsheets,
contact management, accounting, etc. A server with NC's will
cost less to buy and operate by a significant margin. Now that
MSFT has agreed to lock in both Citrix and Prologue technology
for the multi-user kernel and make that kernel themselves, the
concept has been validated and if any player stands to lose it
is Intel. MSFT is supposed to deliver the multi-user 4.0 by
the end of this year. (I won't be buying until mid 1998 since
the 1.0 MU kernel will probably be a piece junk!).
5) Recent indications are that the growth rate of the home PC
market is slowing down. Not surprising since a certain amount
of saturation has occurred at various economic strata. With
some new inexpensive TV set top web surfing boxes coming to
market, the need for buying a PC is even further reduced for
some consumers. Now if ISP's provide simple server based
e-mail the need for a PC is nearly eliminated. What would
happen if AOL could run on a $600 Java client?
For the really long term planners out there, rapid growth of
xDSL or other high bandwidth "server-to-curb" (tm) technology
will also cut into the need for everyone to buy a PC. The
advent of Java (and its associated JVM) is a further diversion
for Intel. Java is currently the best language/environment
out there for truly distributed applications. Nothing MSFT
makes even comes close (largely due to MSFT security problems).
Since the JVM is inherently an abstraction, it is processor
independent, another reason to skip Intel's high priced
chips.
These are just a few reasons that I would not invest in INTC
as a growth stock. Sooner or later they are going to run out
of gas.
Spence |
