Advanced Micro Devices - Moderated (AMD) Message Board

AMD :AMD Dual-Core 8-Inch Dilemma - Improving Mix Means Lower G
2006-03-31 09:06 (New York)

------------------------------------------------------------------------------
Moors & Cabot
Capital Markets COMPANY UPDATE
------------------------------------------------------------------------------

SEMICONDUCTOR DEVICES
Advanced Micro Devices
Hans Mosesmann, SVP (AMD - $33.59 - NYSE)
203.504.1602
hmosesmann@moorscabot.com
March 31, 2006
Kevin Cassidy
415.288.2247 Rating: SELL
kcassidy@moorscabot.com

------------------------------------------------------------------------------

AMD Dual-Core 8-Inch Dilemma - Improving Mix Means Lower Gross Margins

We believe we now understand why AMD's increasing mix of high performance
processors is detrimental to gross margins. In this note we believe we uncover
this mystery.

Advanced Micro Devices

Symbol: AMD
Price: $33.59
12-Month Price Target: $10.00
52 Week Range: $42.70 - $14.08
Shares O/S: (mm) 452.0
Market Cap: ($mm) $15,183
Average Daily Volume: 17,757,578

3-Yr. EPS CAGR: 15.0%

Fiscal Year End: December

EPS (Net)Diluted 2005A 2006E 2007E
Q1 ($0.04)A $0.31 $0.10
Prior: -- $0.25 --
Q2 $0.03A $0.21 $0.10
Prior: -- $0.15 --
Q3 $0.18A $0.19 $0.14
Prior: -- $0.21 --
Q4 $0.21A $0.21 $0.16
Prior: -- $0.23 --
Annual $0.38A $0.92 $0.50
P/E 88.4x 36.5x 67.2x

Revenue($M)Basic 2005A 2006E 2007E
Q1 $1,227A $1,450 $1,300
Prior: -- $1,350 --
Q2 $1,260A $1,400 $1,300
Prior: -- $1,300 --
Q3 $1,523A $1,350 $1,400
Q4 $1,838A $1,400 $1,500
Annual $5,848A $5,600 $5,500
Mkt Cap/Rev 2.6x 2.7x 2.8x

Source: Moors & Cabot and Bloomberg.
WWW.MOORSCABOT.COM

* We now believe we have resolved the gross margin dilemma posed by the
combination of lower QoQ gross margins in Q405 and AMD's 10-K commentary
that increasing mix of high performance processors were detrimental to gross
margins in '05.

* AMD's choice of a monolithic dual core solution (2 cores on a single sliver
of silicon), combined with 8 inch wafers and a faster than expected dual
core ramp, is adversely impacting gross margins.

* When one of the "cores" of the chip is defective, AMD may need to scrap the
entire chip. While AMD 90nm defect densities are world class, the actual
yields of dual core "chips" vs. single core "chips" are dramatically lower,
in our opinion.

* AMD's move to 300mm/90nm may help AMD's cost structure in 2H06, however its
new AM2 solutions will likely be larger, AMD's entire product (desktops/
notebooks) are transitioning to dual core, and ASPs are going down.

* We are increasing our Q106 topline by $100 million to $1.45 billion on
strong server momentum, however our EPS estimate goes up to $0.31 from
$0.25, well below the $0.40 EPS such a topline upside would suggest under
normal circumstances (70%-80% incremental margins).

-------------------------------------------------------------------------------

Please see the important disclosures at the end of this report.

Over the past several weeks we have attempted to resolve the declining gross
margin dynamic seen in Q405, where ex-depreciation, gross margins declined 200
bps sequentially off of a 30% sequential top-line growth.

The nagging curiosity became a true mystery when AMD released its 2005 10-K a
month ago, in which on page 32 it indicated that its 1.5% gross margin
improvement in 2005 had been aided by a 9% increase in microprocessor ASPs,
offset by an increasing mix of high performance microprocessors. Industry and
street observers when provided this information have been essentially baffled,
given that intuitively a better mix should lead to better gross margins,
however the consensus being that some kind of yield problem had developed. We
now believe we can explain the dynamic at hand at AMD.

Brief Background In Dual Core x86 Dynamics In 2005

In early 2005 AMD had signaled to the industry that it would introduce the
first dual-core x86 server processors in the world for a May launch. The chip,
a true dual-core chip with 2 cores on a single sliver of silicon, was an
impressive feat technologically and competitively given that AMD knew Intel was
months away from a server dual-core solution. However, Intel actually rained on
the AMD dual-core party to some degree, by launching a desktop dual-core
solution 1 week before the AMD launch. One characterization could be Intel kind
of cheated (if you happen to be a processor architect or purist), in that its
dual-core solution used 2 single core processors that were basically bolted
together, packaged, and Voil`! Intel was first to dual core.

AMD's Dual Core Solution Meets 8 Inch Wafers

While AMD's dual-core solution was the more elegant and technically more
compelling dual-core implementation compared to Intel's quick and dirty early
solution; 4 dynamics conspired against AMD: 1. AMD's 8 inch wafer
manufacturing, 2. a large dual core die size, 3. a monolithic dual-core
approach, and 4. faster than expected transition of dual core processors in the
market.

1. Most major semiconductor logic vendors that produced 90nm chips in 2005 used
300mm wafers, such as Intel, Xilinx, Altera, ATI, etc. AMD however is
noteworthy in its 200mm/90nm manufacturing in 2005.

2. AMD 90nm dual core solutions are roughly twice the size of its single core
brethren, coming in at 199 mm2. AMD essentially gets half as many "chips"
theoretically when going to dual-core. By comparison, Intel's Netburst
solutions use 2-75 mm2 single cores (on 300 mm wafers).

3. A monolithic dual-core approach means that when one of the cores is not
functional, you have to scrap the entire chip (or possibly sell it as a single-
core). So, not only is AMD getting half the chips per 8 inch wafer as seen in
#2 above, they are scrapping potentially good chips on top of this. An
expensive proposition particularly when compared to Intel's approach, where a
defective single core is merely scrapped while its intended pair is then paired
with a non-defective single core.

4. By Q305, AMD's server dual-core mix had reached 20%-25%, per information
provided by AMD during its November analyst day. The 20%-25% mix points to
truly impressive growth given that the product class had been introduced in
mid-Q205. We suspect that the bulk of the server growth in Q405 was of the
dual-core variety and may have approached 50% of the mix. We think AMD will hit
its estimated 90% plus dual core mix by Q406, 2-3 quarters ahead of time.

Does AMD Have A Yield Problem? Yes and No

We believe AMD's defect densities at 90nm are world class, however given the
combination of 8 inch manufacturing and the large dual-core die size
(monolithic approach), we believe the company is simply scrapping too much
silicon. Hence, the more high performance processors (dual core) the company
sells, the more pressure on gross margins.

300mm Manufacturing Does Not Come In Soon Enough

By going to 300mm (90nm) as 2006 progresses, AMD may be able to lessen the
dual-core blow somewhat. However, AMD's migration to the new DDR2 compatible
AM2 socket (Socket F), will very likely mean bigger die sizes not smaller.
Furthermore, look for the end markets to demand dual core for all classes of
processors as the year progresses. AMD's die size is still very large compared
to Intel's new dual-core monolithic solutions (Core Duo for example); and we
suspect that as AMD moves its desktop and notebook solutions to dual core in
Q206, the larger die sizes could cause margins to decline.

Q106 Topline Upside Driven By Dual-Core Servers

AMD is seeing continued momentum in servers in Q106, however rather than the
70%-80% incremental gross margins investors would expect from an upside of the
magnitude of $100 million ($0.15 plus incremental EPS), we believe gross
margins are materially below these levels (as was seen in Q405).

Changes to Estimates

We are increasing our topline estimate in Q106 by $100 million to $1.45 billion
from $1.35 billion as we believe AMD's server momentum has continued into Q1.
Our Q106 GAAP EPS estimate increases 6 cents to $0.31 from $0.25. Gross margins
are estimated to decline 100 bps to 52.5% from 53.5% as we do not believe AMD
is gaining leverage with the incremental revenue. We estimate OpEx will
increase roughly $10 million to $548 million from $538 million. Our revenue
estimates for 2006 increase $200 million to $5.6 billion from $5.4 billion
based on server momentum in 1H06, though our 2006 GAAP EPS estimate only
increases 7 cents to $0.92 from $0.85. Gross margin estimates for 2006 decline
80 bps to 52.6% from 53.4%. Our 2007 estimates remain the same with topline at
$5.5 billion and bottom-line at $0.50.

Companies mentioned in this report:
Advanced Micro Devices (AMD; $33.59; SELL)
Intel Corporation (INTC; $19.70; BUY)

Valuation
AMD Valuation. Our 12-month price target of $10 for AMD shares is based on a
20x multiple to our 2007 EPS estimate of $0.50. The 20x multiple is an 11%
premium to the company's historical 18x P/E ratio and is actually the same
multiple we are applying to INTC shares, which is justified by our view given
that AMD is a significantly better managed company today, under CEO and
Chairman Hector Ruiz.

Intel Valuation. Our 12-month price target of $30 based on 20x our 2007
estimate of $1.50.

Risk
AMD Risk. Risks that would prevent AMD shares from achieving our $10 price
target include better than anticipated success of AMD 64 microprocessor
products, development of new products that would spur demand beyond current
expectations, accelerated transition to more advanced manufacturing
technologies and the company's near-term ability to gain share from Intel while
maintaining stable ASPs.

Intel Risk. Risks that could potentially impact Intel's results include margin
pressure due to underutilization of fabs, competition from AMD's Opteron/
Athlon-64 chips, PC and communications end-market weakness, poor execution of
manufacturing upgrades, and slowdown in semiconductor growth.

IMPORTANT DISCLOSURES

Analyst Certification
The research analyst(s) responsible for the preparation of this research report
certifies that the opinion(s) expressed herein accurately reflect his or her
personal view about the subject companies and securities. Their compensation is
based upon various factors, including quality of research, investor client
feedback, stock picking, competitive factors, firm revenues, and investment
banking revenues.

Companies mentioned in this report:
Advanced Micro Devices (AMD; SELL; $33.59)
Intel Corporation (INTC; BUY; $19.70)
The basis for recommendations and price targets, investment risks, historical
price charts and other important disclosures for each subject company that
includes a recommendation can be found at mac.bluematrix.com
bluematrix/Disclosure or through your Moors & Cabot sales representative.
Rating definitions and allocations
BUY: Moors & Cabot expects the stock to outperform the market over the next 12-
18 months.

HOLD: Moors & Cabot expects the stock price to perform in-line with the market
over the next 12-18 months.

SELL: Moors & Cabot expects the stock to underperform the market over the next
12-18 months.

NOT RATED: Stock does not have an investment rating by Moors & Cabot.

% of coverage companies for which compensation has been received for products
or services other than Investment Banking services in the last 12 months:
BUY: 6.4%

HOLD: 18.2%

SELL: 14.3%
The information contained herein, including any expression of opinion, has been
obtained from, or is based upon, sources believed by us to be reliable, but is
not guaranteed as to accuracy or completeness. This is not intended to be an
offer to buy or sell or a solicitation of an offer to buy or sell, the
securities or commodities, if any, referred to herein. Moors & Cabot acts as a
principal in any security mentioned herein. Our firm and/or its officers and
employees have positions in one or more of the securities mentioned herein. The
firm or one of its affiliates perform investment banking or other services for,
or solicit banking or other business from companies mentioned in this report.

Moors & Cabot in no way guarantees the accuracy or completeness of such
information. This report has been prepared for informational purposes only and
was issued by Moors & Cabot Capital Markets for distribution to our market
professional and institutional investor customers. Recipients who are not
market professional or institutional investor customers should seek the advice
of their financial advisor prior to making any investment decisions based on
its contents. Nothing in this report constitutes investment, legal, accounting
or tax advice or a representation that any investment or strategy is suitable
or appropriate to your individual circumstances, nor does this constitute a
personal recommendation to you. Opinions expressed are subject to change
without notice and past performance is no guarantee of future results.

TRADING DESKS
Boston, MA 800.732.4824

New York, NY 888.566.9239

San Francisco, CA 800.655.3583
www.bluematrix.com

Provider ID: 00007082
-0- Mar/31/2006 14:06 GMT