Advanced Micro Devices Inc (AMD)

[KeithDust2000]

Merrill Lynch upgrades AMD to "Buy", target price raised to $22

Here´s the report, thanks to skeptically/yahoo

Comment Semiconductors 11 November 2004
Joseph Osha Sidney Ho
Advanced Micro Devices
Hitting the product cycle BUY
Volatility Risk:HIGH
Reason for Report: Upgrade to Buy

12-Month Price Objective: $22.00
Date Established: 11-Nov-2004
Estimates (Dec) 2003A 2004E 2005E
EPS: -$0.79 $0.51 $0.70
P/E: NM 33.9x 24.7x
GAAP EPS: -$0.79 $0.50 $0.70
GAAP P/E: NM 34.6x 24.7x
EPS Change (YoY): NM 37.3%
Consensus EPS: $0.52 $0.68
(First Call: 04-Nov-2004)
Q4 EPS (Dec): $0.12 $0.18
Cash Flow/Share: $0.66 $2.65 $4.27
Price/Cash Flow: 26.2x 6.5x 4.0x
Dividend Rate: Nil Nil Nil
Dividend Yield: Nil Nil Nil
Opinion & Financial Data
Investment Opinion: C-2-9 to C-1-9
Mkt. Value / Shares Outstanding (mn): $7,227.2 / 418
Book Value/Share (Sep-2004): $6.16
Price/Book Ratio: 2.8x
ROE 2004E Average: 7.7%
Total Debt / Capital: 44.3%
Est. 5 Year EPS Growth: 15.0%
Est. 5 Year Dividend Growth: NA
Stock Data
52-Week Range: $18.50-$10.76
Symbol / Exchange: AMD / New York
Institutional Ownership-Vickers: 68.5%
Brokers Covering (First Call): 29

Highlights:
• We’re upgrading our investment recommendation for
Advanced Micro Devices from Neutral to Buy
(C-1-9), and establishing a $22 price target for the
stock.

• We believe that AMD is in the midst of a
microprocessor product cycle that should allow the
company to continue expanding profit margin and
gaining market share against Intel for at least another
year. In contrast, Intel’s own microprocessor roadmap
for the next twelve months looks troubled.

• AMD’s stock is already up 59% off the bottom, and
our shift in stance may look late. However, a look at
how the stock has performed in the past, when the
product cycle is operating in AMD’s favor, suggests
plenty of additional upside.

• Intel is beginning to put pressure on AMD in the flash
memory business, and we think that AMD’s flash
revenues and margin could decline. A successful
ramp of K8 should more than compensate for weaker
flash revenues, though. We note that AMD took a
20% sequential hit in flash revenues during Q3
without an overall decline in gross margin.

• Our $22 price target represents the peak cycle
valuation suggested by our long-term DCF model.
Given our belief that AMD is entering a period of
margin expansion and market share gain we think that
the peak-cycle target is appropriate.

http://finance.messages.yahoo.com/bbs?.mm=FN&action=m&board=4687810&tid=amd&sid=4687...

Here´s the full report:

Taking AMD to a buy – product cycle
accelerating
We’re upgrading our investment recommendation for
Advanced Micro Devices from Neutral to Buy, and
establishing a $22 price target for the stock. We believe
that AMD is in the midst of a microprocessor product
cycle that should allow the company to continue
expanding profit margin and gaining market share against
Intel for at least another year. In contrast, Intel’s own
microprocessor roadmap for the next twelve months looks
troubled. Our $22 price target represents the peak cycle
valuation suggested by our long-term DCF model. Given
our belief that AMD is entering a period of margin
expansion and market share gain we think that the peakcycle
target is appropriate.


Intel looks likely to continue struggling with
Prescott through 1H 2005
In the microprocessor business, Intel looks likely to spend
2005 living with the consequences of several bad
decisions, and we do not think that the company can do
much to address its problems during the coming year. In
the desktop processor business, Intel has largely run out of
ability to improve performance by increasing processor
clock speeds. Intel’s decision to focus on clock speed
above all else when designing the Netburst core currently
used in Prescott version of P4 has left the company with a
processor that generates too much heat to be run any faster.
Intel’s only remaining option for the first half of 2005 is
adding performance by adding cache memory to the chip,
which is less appealing for Intel’s customers and negative
for Intel’s own margin structure.


AMD’s ramp of 90nm manufacturing should
go more smoothly
Meanwhile, we expect AMD’s own ramp of the Athlon
desktop architecture on the 90nm manufacturing process to
go much more smoothly than Intel’s ramp did. AMD’s
Athlon core was designed with less emphasis on raw clock
speed and more emphasis on size and efficiency, which in
retrospect looks like a smart decision. We think that AMD
should be able to offer improving performance on Athlon
64 through most of 2005 as a result. Much has been made
of AMD’s advantage in 64-bit computing, but frankly we
don’t think that it matters much in the desktop processor
market. The real advantage that AMD brings to the party
is a smaller, more power-efficient core.

Intel has dual-core coming, but we don’t think
that’s going to help until 2006
Intel has of course responded to its roadmap problems in
the desktop market by pulling in plans for dual-core
processors, first scheduled for release in 2006, to mid
2005. We do believe that multiple processor cores are the
right direction for the industry to take, but we also believe
that the speed with which Intel is attempting to execute the
transition will leave the company struggling to offer a
compelling value proposition for dual-core processors in
2005. We think that it will take until 2006, when Intel
moves dual-core processors to 65nm manufacturing and
the PC OEM community has enough time to take
advantage of the new architecture, for Intel’s dual-core
products to gain traction with PC buyers.


AMD has jumped on several Intel mistakes in
the server MPU market
In the server microprocessor market, AMD’s Opteron
product has taken advantage of several strategic mistakes
from Intel. Intel’s decision to continue focusing on the
high-end IA-64 architecture, first rolled out in the Itanium
series of processors, caused to company to postpone
offering important changes to the IA-32 architecture that
supports the Pentium and Xeon processor lines. An
extended 64-bit instruction set was the most important
missing piece, but Intel also neglected to address problems
with system performance scaling in 2 and 4 processor
systems, while at the same time continuing to charge very
high prices.
AMD jumped on the resulting opportunity with Opteron, a
product that is easy to implement, highly competitive in 2-
processor and 4-processor systems, and priced
competitively with Intel’s products. Intel can always drop
prices for its server products, but only at the expense of
hurting overall margin, while AMD could still benefit from
boosting server processor revenues even at lower prices.
We don’t think there’s any workable strategy that Intel can
use to counter AMD’s share gain in the server processor
market.

New relationship with Chartered is a step in
the right direction
We wrote a week ago that it was important for AMD to
secure additional manufacturing capacity. The company’s
existing 200mm manufacturing facility is full, and AMD’s
own 300mm facility is not likely to begin producing output
for revenue until early 2006. To that extent we think that
the recently announced relationship with Chartered
Semiconductor is important. Chartered is a member of the
group of companies, also including AMD, which has codeveloped
manufacturing process technology with IBM.
Ramping 90nm manufacturing will be a challenge for
Chartered, but with support from both AMD and IBM we
think it’s feasible. The result should be another source of
much-needed supply for AMD, and we think it’s possible
for Chartered to have product available before the end of
2005 although neither Chartered nor AMD are saying that
right now. Until then, AMD’s biggest limitation will be its
own manufacturing capacity.


Balance sheet is a challenge, but we think AMD
can squeak by
AMD’s capital expenditures through 2007 for Fab 36 (the
new 300mm facility) are expected to be approximately
$2.5 billion. At first glance, that may seem implausible
given AMD’s debt-laden balance sheet. However, with
expected cash flow from operations of about $400-500
million per quarter for the next few quarters, along with
the reduced debt payment requirement in the next two
years, we can get the numbers to work, but there is little
margin for error. AMD is also expecting to get some help
from other sources. The table below sets forth the
financing arrangement for Fab 36 – we note that
approximately $600 million of the planned spending for
2004 is targeted for the facility.
AMD has also taken steps to reduce intermediate-term
liquidity challenges. In October, the company sold $600
million of senior notes due in 2012, and paid off a term
loan to German creditors that was due in 2005 and 2006.
That reduces the required debt repayment to approximately
$100-120 million in each of the next two years. Further,
AMD converted $130 million of its $403 million
convertible notes due in 2007 to 19 million shares of
common stock.
Table 1: Financing arrangement for Fab 36
(in $ million) Amount
Contribution from AMD $ 718
Investment from Leipziger Messe (unaffiliated) 246
Investment from Fab 36 Beteiligungs (unaffiliated) 147
Subsidies from Germany and the State of Saxony 666
Loans from a consortium of banks 859
Total funding $2,636
Source: Company information

We’ll be keeping an eye on flash memory . . .
One potential problem for AMD’s profitability outlook is
the flash memory business. Market share held by the
Spansion joint venture, which AMD controls, rose to #1
during the first half of 2004, but Intel’s abundant excess
capacity and need to re-establish presence in the wireless
end market are a problem. It is clear that Intel is beginning
to take share back, and we think that the result could be a
decline in both revenue and margin for AMD in flash. It’s
worth pointing out, though, that AMD absorbed a 20%
sequential decline in flash revenues during Q3 2004
without any deterioration in the overall margin structure.
Intel may be able to inflict pain on AMD in the flash
memory business, but if AMD is able to successfully ramp
the much higher-margin microprocessor business, flash
won’t matter.
Chart 3: AMD revenue breakdown – Q3 2004
All Other
2%
Microprocessors
55%
Flash Memory
43%
Source: Company Information


We may be late, but there’s plenty of room for
additional returns
AMD’s stock is up by 59% from the $10.86 low in
September, and the stock is up by 16% since the beginning
of the year as compared to a 20% decline for the
Philadelphia Semiconductor Index. Viewed in that context
it might seem that our upgrade of the stock comes too late.
However, we’d encourage investors to look at what
AMD’s stock has done in the past, both in absolute terms
and relative to Intel, when AMD really has a product
advantage against Intel.


Stock performance tends to follow product
cycles
The short answer is that AMD’s stock price has
historically responded well to product cycles. The stock
outperformed INTC during the K7 ramp between 1999 and
2001, as the chart below illustrates. AMD stock hit a high
of $48.50 (post-split) in June 2000 when 50% of revenue
came from K7, but its relative performance against INTC
peaked four quarters later when the entire product line was
switched to K7. After that, AMD’s stock more or less
tracked Intel’s until the recent bout of outperformance.
We’re seeing a similar pattern for K8 – AMD’s stock has
outperformed Intel already, but history suggests that there
is a good deal more outperformance left as AMD
completes its transition to K8.

We see upside potential in gross margin
We’d also point out that AMD is one of the few companies
in our universe of coverage that looks set to show positive
rather than negative margin leverage during 2005.
Although we are maintaining our 2005 earnings estimate
of $0.70 per share for now, we do see potential upside to
the number. Competitive products and better pricing could
keep ASP from deteriorating, and we think the upside will
mainly come from the gross margin line. Currently K8
products carry gross margins that are 5-10 percentage
points higher than that of K7, which should become
apparent as the company finishes transitioning to K8 over
the next few quarters. Our sensitivity analysis suggests that
a 1 percentage point increase in gross margin in 2005
would increase EPS by $0.10 – we regard that kind of
improvement as entirely plausible.


Price objective and risks
As AMD enters a new product cycle, we believe the
company deserves a peak cycle valuation despite our more
conservative stance on the semiconductor business overall.
We are establishing our 12-month price objective for
AMD of $22, which is supported by our proprietary
returns model. On a price-to-earnings basis, our price
objective suggests a P/E ratio of 31x based on 2005E EPS,
which seems pricey. However, unlike most of the
companies in our coverage universe, we see an upward
rather than a downward bias to our estimates.
As we learned several times in the past, there are risks
relating to AMD’s execution. First and foremost is AMD’s
ability to build Fab 36 within budget and targeted time
frame. Second, we think that AMD will continue to see
pressure from Intel in the microprocessor business, which
could be in the form of pricing or competitive products.
Third, Intel has becoming more aggressive in the flash
business. Finally, we note that AMD is susceptible to the
overall demand for PCs and, increasingly, enterprise
spending budgets.