few thoughts from bulls at Needham:
09:23am EST 28-Jan-02 Needham & Company, Inc. (Dan K. Scovel 212-705-0322) AMD
Semiconductor Industry: Chip Shot
January 28, 2002 Dan K. Scovel 212-705-0322
Semiconductor Industry: Chip Shots
Advanced Micro Devices (AMD, $16.30 +0.17, Buy, Target Price $20)
1Q: ($0.05) on $904 million, 2002: $0.13 on $3.8 billion, 2003: $0.97 on
$4.8 billion
Market Share Gains: Mercury Research estimates AMD grew microprocessor unit
market share during 2001 to 20.2% from 16.7%, while Intel slid to 78.7% from
82.2%. However, AMD lost share sequentially 4Q to 18.5% from 20.5% during 3Q01,
while Intel gained during the quarter to 80.6% from 78.6%. Including Intel
Pentium III shipments into the Xbox during 4Q, Intel accounted for 81.2% and AMD
18.0%. AMD's fading share 4Q was more than offset financially by a richer mix
of more expensive devices at the high end. We consider this information detail
backfill to the 4Q earnings announcements from both companies.
Mobile Athlon Speed Bump: AMD announced the Athlon 4-1500+ for portable PC
applications at $525 this morning with a design win at Compaq (CPQ, $11.95,
Hold). We consider this good news as AMD continues to fill out the breadth of
its attractive Athlon processor architecture to faster growing and more
profitable market segments (portable and server, specifically).
Intel (INTC, $33.68 +0.48, Hold)
1Q: $0.15 on $6.9 billion, 2002: $0.70 on $29 billion, 2003: $0.87 on $33
billion
Barron's Mixed Bag: This week Barron's notes INTC shares may benefit from undue
pessimism in the Great Buys Revisited article, but later blasts the company for
making processors that consume far too much power in Bill Alpert's column. Our
guess is the positive comment will carry the day--by a nose.
64-bit Yamhill: Late last week the San Jose Mercury News published an article
based on comments from an unidentified former Intel engineer regarding a 64-bit
processor initiative in Oregon code named Yamhill. The Yamhill is a 64-bit
processor that is backward compatible to the existing 32-bit x86 architecture
that has spawned several generations of PC processors for Intel. The plan is to
build such capability into the upcoming Prescott device and to energize the
option, should it be deemed desirable or necessary. We are inclined to believe
the general claims, since they have been rumored in recent months. However, it
is not clear to us how such an option could be provided without a significant
die size/unit cost penalty.
Fundamentally, we consider this bad news for AMD and its
strategically important 64-bit Hammer family initiative due out later this year,
and good news for Intel as a contingency to address such a competitive threat.
However, the Yamhill program is also extremely awkward for Intel: first, it
would tacitly endorse the AMD Hammer product strategy and plans; and second, it
could undercut the significant effort of years and billions of dollars behind
the 64-bit Itanium family intended for high-end servers, that is not backward
compatible with the existing x86 software code. We have not been as bullish on
the Itanium as others due to limited unit market potential, but those that have
been more bullish could likely be very concerned that the Yamhill would
cannibalize some portion of the Itanium potential opportunity. We also believe
this awkward situation for Intel actually works in AMD's favor, to a degree:
Intel would likely not introduce the Yamhill unless and until the Hammer proves
successful, thereby ensuring AMD's market lead. But Yamhill would significantly
close the competitive lag.
Price Cut: Expected today across a variety of processor offerings. We consider
this business as usual. |
