Please Do Not Read The Text Below.

Download the original file at

Morgan Stanley Global Chipwatch Part 1 of 1 (47 pages)

We copied the scrambled text from pdf file (that's how it gets copied, sorry we had no control) so the keywords would be searchable.

Comment
Please see analyst certification and other important disclosures starting on page 42.
Page 1
Industry
Equity Research
Global
Global Semiconductor Research Team Monthly January 12, 2005
Morgan Stanley Global
ChipWatch — January 2005
GLOBAL BEST IDEA
Samsung Electronics (005930.KS, W548,000, O)
REGIONAL BEST IDEAS
Semiconductors - US
National Semiconductor (NSM, US$22, O-V)
NVIDIA (NVDA, US$36, O-V)
Semiconductors - Japan
NEC Electronics (6723.T, ¥6,200, O)
Semiconductors - Europe
ASML Holding (ASML.AS, €14, O-V)
Semiconductor Capital Equipment - US
KLA Tencor (KLAC, US$43.25, O)
Semiconductor Production Equipment - Japan
Tokyo Electron (8035.T, ¥6,900, O)
Prices shown above are price targets.
Stock ratings are relative to the individual analyst’s
industry coverage universe or relevant country
index and should be viewed in context with their
industry views.
• Near-term fundamentals remain challenging and earnings risk remains high
Although we believe that end-market demand generally met expectations in the fourth
quarter, a broad-based inventory correction and the need for a high level of turns
business continued to promote a large number of negative earnings surprises, and we
believe that the bias to earnings estimates is still negative.
• Inventories need to decline further, but seasonally weaker demand makes this
more challenging
Despite seeing a meaningful reduction in supply-chain inventories during the fourth
quarter, we believe that the average semiconductor company still out-shipped the key
end markets by 10–15 points in 2004, and we expect the inventory correction to
continue to pressure results in the first quarter.
• Excess capacity should pressure semiconductor prices and capital spending
A slowdown in global GDP growth and the ongoing inventory correction has led to
excess capacity, and we expect it to promote a more aggressive pricing environment
once the inventory correction is completed. In addition, it should pressure
semiconductor capital spending, which we expect to decline 11% in 2005.
• Fundamentals should bottom in 1H05 and we expect 8–12% semiconductor
revenue growth in 2006
Despite our near-term concerns about growth rates, margins, and negative earnings
surprises, we expect easier comparisons and a recovery in unit demand (once the
inventory correction is completed) to support a reacceleration of revenue growth in
2H05 and 2006.
• Recent decline suggests stocks have reconnected with fundamentals again
After reaching a 52-week low at 351 on September 8, the SOX index rallied sharply
and outperformed the overall stock market until early December, even though
fundamentals were deteriorating. Since then, the SOX has declined sharply and
underperformed the broad market, but we view SOX 400 as an equilibrium level.
• Global Industry Views
Please see page 8 of this report for a complete list of our global industry views related
to semiconductors. Our favorite global stock idea is Samsung Electronics.
Global Semiconductors
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 2
Morgan Stanley Global ChipWatch — January 2005
Morgan Stanley
Global Semiconductor
Research Team
Asia/Pacific
Dickson Ho +886 2 2730 2999
Sunil Gupta +65 6834 6732
Praveen Choudhary +65 6834 6744
Max Lee +886 2 2730 2863
Jasmine Lu +886 2 2730 2870
Bonnie Chang +886 2 2730 2861
Sharon Shih +886 2 2730 2991
Ellen Tseng +886 2 2730 2864
Howard Kang +886 2 2730 2811
Frank Wang +886 2 2730 2869
Keon Han +822 399 4933
Jonathan Rhee +822 399 4946
Shawn Kim +822 399 4939
So-Young Ahn +822 399-4937
Europe
Stuart Adrian +44 20 7425 3299
Kirsten Parker +44 20 7425 8617
Dieter Weber +44 20 7425 9895
Japan
Naoki Sato +81 3 5424 5927
Kazuo Yoshikawa +81 3 5424 5389
North America
Mark Edelstone +1 415 576 2381
Louis Gerhardy +1 415 576 2391
Harlan Sur +1 415 576 2359
Sonia Kimotsuki +1 415 576 2388
Jay Iyer +1 415 576 2607
Sanjay Devgan +1 415 576-2382
Timm Schulze-Melander +1 415-576-2324
Bernie Mahon +1 212 761 6274
Aaron Husock +1 212 761 7678
Tatiana Feldman +1 212 761 7134
Table of Contents
Global Industry Summary ………………………………………………………….3
Industry Views ……………………………………………………………………8
Global Best Idea. ……………………………………………………………………9
Regional Best Ideas...………………………………………………………………10
Semiconductors — Asia/Pacific…………………………………………………...12
Semiconductors — Japan…………………………………………………………..15
Semiconductors — Europe………………………………………………………...17
US Semiconductors — Electronic Design Automation (EDA)……………………19
Three-Month Semiconductor Billings……………………………………………..20
US Semiconductor Capital Equipment…………………………………………….21
US Semiconductor Equipment Book-to-Bill Ratios……………………………….23
US Semiconductor Equipment Bookings and Billings ……………………………24
US Semiconductor Equipment Bookings and Billings (YoY Change)……………25
Semiconductor Production Equipment — Japan…………………………………..26
PC Systems & Hardware— Taiwan……..………………………………………...29
Electronics Supply Chain…...……………………………………...………………34
Global Valuation Table…………………………………………………………….36
Global Semiconductor Capital Spending…………………………………………..40
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 3
Global Industry Summary
Summary and Investment Conclusion
Last year was challenging for semiconductor stock investors,
and we believe that some of the holdover effects from 2004
will likely cause near-term fundamentals to remain
challenging and difficult to accurately forecast. Key factors
causing us to hold this view as we enter 2005 include our
expectations for a continued deceleration in revenue growth
in the semiconductor industry and an actual decline in
semiconductor capital equipment spending; more aggressive
pricing trends; margin pressure; limited visibility caused by
low backlog levels and an increased dependency on turns
bookings (orders that are received and shipped for revenue
in the same quarter); and a high degree of earnings risk.
Although we estimate that channel and end-market
inventories were meaningfully reduced during the second
half of 2004, expectations for seasonally weak end-market
demand in the first quarter suggest that additional declines
in inventory levels should be expected. In addition,
semiconductor company inventories remain too high, and
when combined with expectations for slower global GDP
growth, we expect semiconductor capital spending to
decline in 2005 after two years of solid growth in 2003 and
2004. Once channel and overall semiconductor company
inventories have been completely purged, we expect excess
capacity to promote a more aggressive pricing environment
as 2005 unfolds and new orders are recorded.
In order to reflect current valuation parameters and
expectations for a challenging fundamental environment
this year, most of the analysts that contribute to the Global
ChipWatch report have In-Line industry views on their
individual sectors. The two exceptions are Korean
Semiconductors (Attractive industry view) and U.S.
Semiconductor Capital Equipment (Cautious industry view).
The Attractive view on Korean Semiconductors is primarily
driven by expectations that Samsung will continue to
expand margins and benefit from its leadership positions in
memory, cell phones, and flat-panel displays. The Cautious
view on U.S. Semiconductor Capital Equipment is driven
by the belief that the valuations of the equipment stocks will
decline once investors fully discount the anticipated depth
and duration of the current downturn in capital spending.
Similar to last year, we believe that the key to
outperforming the semiconductor sector and the overall
market this year will be through individual stock selection.
Companies that can gain share by benefiting from new
and/or strong product cycles, as well as those companies
with reasonable valuation parameters that can hold or
improve their margins this year should be in a position to
provide solid relative performance. Companies that are
either losing market share or destined to suffer from a
meaningful decline in margins should be expected to
underperform.
Semiconductor stock price performance improved in 4Q
but the stocks underperformed in 2004 overall
Although final data will not be available until the end of the
month, we estimate that the global semiconductor industry
grew 28% in 2004 versus 18% growth in 2003 and 1%
growth in 2002. Furthermore, three-month-average
revenues reached a new record in November. Despite these
strong results, the MSCI World Semiconductor and
Semiconductor Equipment index declined 20% versus a
15% decline in the SOX index and gains of 9% in the S&P
500 and NASDAQ Composite indices in 2004 (Exhibit 1).
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 4
Exhibit 1
Semiconductor stocks have started to underperform again and reconnect with weak near-term fundamentals
(as of January 12)
MSCI - Technology Sectors (US) 2002 1Q03 2Q03 3Q03 4Q03 2003 1Q04 2Q04 3Q04 4Q04 2004 1Q05 QTD
Communications Equipment -51.9% 9.4% 27.9% 12.5% 16.6% 84.5% 8.5% -1.4% -12.3% 18.1% 8.6% -5.1%
Computers & Peripherals -33.5% 4.6% 28.8% 2.5% 11.3% 52.1% 5.6% 1.0% -0.8% 27.4% 41.2% -4.6%
Electronic Equipment & Instruments -47.2% -7.0% 35.0% 20.7% 19.3% 78.9% 3.2% 0.4% -16.5% 13.7% -1.6% -6.2%
IT Services -33.8% -10.3% 23.7% 4.5% 8.9% 24.5% -2.5% 5.0% -8.0% 7.4% 0.9% -5.0%
Internet Software & Services -43.4% 27.9% 47.0% 2.9% 24.2% 139.3% 4.7% 35.1% -3.5% 40.1% 86.7% -7.5%
Office Electronics -22.7% 8.1% 21.7% -3.1% 34.5% 71.4% 5.6% -0.5% -2.9% 20.8% 23.3% -4.7%
Semiconductors & Semiconductor Equipment -43.3% 1.0% 28.4% 16.7% 22.8% 81.2% -3.4% -1.3% -22.0% 13.8% -16.7% -7.3%
Software -33.2% -3.0% 28.9% 7.4% 13.3% 51.2% -5.8% -1.7% -13.5% 24.9% -1.3% -6.8%
MSCI Information Technology (US) -40.5% -0.1% 28.3% 11.1% 17.2% 65.7% -0.5% 1.0% -13.9% 16.9% 1.5% -6.3%
MSCI Semi & Semi Equipment (World) -39.2% -5.7% 26.4% 23.9% 14.1% 67.9% -0.4% -8.9% -18.2% 8.9% -20.0% -5.8%
S&P 500 -23.4% -3.6% 14.9% 2.2% 11.6% 26.4% 1.3% 1.3% -2.3% 8.7% 9.0% -2.0%
NASDAQ -31.5% 0.4% 21.0% 10.1% 12.1% 50.0% -0.5% 2.7% -7.4% 14.7% 8.6% -3.8%
SOX -44.6% 2.4% 21.4% 16.7% 21.0% 75.7% -4.1% -0.4% -20.8% 12.8% -14.7% -7.3%
Morgan Stanley US Semi Universe -49.8% 1.9% 45.0% 22.7% 23.7% 117.2% -0.7% -8.5% -23.8% 17.1% -20.5% -9.7%
Source: Morgan Stanley research.
Some of the underperformance in 2004 can be attributed to
the strong absolute and relative gains recorded in 2003,
when the MSCI World Semiconductor and Semiconductor
Equipment index advanced 68% versus gains of 26% and
50% in the S&P 500 and NASDAQ Composite indices,
respectively. However, we believe that much of the
weakness was caused by a sharp deterioration in
fundamentals that began in the second quarter and became
overwhelming in the third quarter. If not for strong gains
during the last four months of 2004, the underperformance
of the average semiconductor stock for the overall year
would have been far worse. In the fourth quarter, the MSCI
World Semiconductor and Semiconductor Equipment index
increased 9%, versus gains of 9% in the S&P 500 and 15%
in the NASDAQ Composite. Although limited data are
available, the semiconductor stocks have underperformed
the broad market averages so far in 2005.
Valuations are reasonable but negative surprise
potential remains high
Based on our 2005 estimates, the average stock in our U.S.
semiconductor company universe is currently trading at
3.4x sales, which is close to the mid-point of the 2.5-5.0x
range that we have long expected the average
semiconductor stock to trade at on an ongoing basis. We
currently view the 400 level on the SOX index as an
equilibrium point, and generally expect the index to trade in
a 350-450 range unless fundamentals either strengthen or
deteriorate beyond current expectations.
Prior to the start of the fourth-quarter earnings-reporting
season, about 20 U.S. semiconductor companies
preannounced negative surprises, with an average revenue
miss of about 9%. In contrast, more than 40 companies
preannounced negative surprises in the third quarter, with
an average revenue miss of around 13%. While there were
virtually no positive preannouncements in the third quarter,
Intel, LSI Logic, NVIDIA, and SigmaTel preannounced
fourth-quarter revenue expectations that were better than
their original guidance. While the less negative surprise
profile of the fourth quarter is encouraging, it appears to
have been propped up by sharp downward revisions in
expectations and a seasonally stronger demand environment.
However, we still believe that consensus earnings estimates
should have a downward bias, especially as companies face
the seasonally weaker demand trends normally associated
with the key end markets in the first quarter.
We estimate that approximately 45% of all U.S.
semiconductor companies will have either preannounced a
negative surprise or report actual results that fall short of
expectations in the fourth quarter (Exhibit 2). In the third
quarter of 2004, about 75% of all U.S. semiconductor
companies failed to meet expectations. This was
comparable to the record-high number of misses in the first
quarter of 2001, and the sequential increase in the number
of negative surprises was the largest we have noticed since
we began keeping statistics in 1994.
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 5
Exhibit 2
The number of negative EPS surprises increased significantly in the third quarter
(Quarterly percentage of semi companies with earnings surprises and expanding operating margins)
0%
20%
40%
60%
80%
100%
1Q94 3Q94 1Q95 3Q95 1Q96 3Q96 1Q97 3Q97 1Q98 3Q98 1Q99 3Q99 1Q00 3Q00 1Q01 3Q01 1Q02 3Q02 1Q03 3Q03 1Q04 3Q04
Calendar Year
Companies With Earnings Surprises
and Expanding Margins
0
200
400
600
800
1000
1200
Quarterly SOX Average
Positive No Surprise Negative Expanding Operating Margins Quarterly SOX Index
Source: Morgan Stanley research.
While it was hard to determine a pattern to the types of
companies that missed expectations in the third quarter, the
companies that preannounced negative surprises in the
fourth quarter tended to be very dependant on turns
bookings, the distribution channel, or the digital consumer
end market. We view this phenomenon as further proof that
the primary reason behind the downturn in industry
fundamentals was due to overly optimistic expectations for
end-market demand earlier in the year and an overbuilding
of inventory throughout the supply chain to react to an
increase in lead times in late 2003/early 2004. Going
forward, weaker pricing trends caused by excess capacity
will likely become a growing reason behind potential
earnings shortfalls. Nevertheless, following two quarters
where a significant percentage of all companies
disappointed, we see the potential for earning expectations
to finally be lowered to levels that are at or below reality
sometime during the first quarter.
From an end-market perspective, demand within the key
computer, communications, and digital consumer markets
was seasonally strong. Similar to the third quarter, there
were fairly few negative surprises within the hardware
sectors in the fourth quarter. Exceptions included a few
semiconductor capital equipment companies, as well as
ADTRAN, 3Com, and UTStarcom, while Emulex, Network
Appliance, Maxtor, and Seagate preannounced positive
surprises. With numerous upsides throughout the storage
supply chain, it appears that the inventory correction that
plagued this sector in the first half of 2004 is over, and end
demand was seasonally strong in the fourth quarter. Overall
PC and cellular phone demand also appeared to be
seasonally strong in the fourth quarter. While certain
segments of the digital consumer end markets were strong,
such as digital TVs and digital audio players, the rest of the
retail sector was somewhat mixed, with a sharp deceleration
in desktop PCs.
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 6
Inventories will likely need to correct further in a
seasonally weak first quarter
In addition to seasonally weaker demand trends that should
be expected to be evident in the first quarter, we still believe
that the overall semiconductor industry out-shipped the key
computer and communications end markets by 10-15
percentage points in 2004 (Exhibit 3). Provided our fourthquarter
estimates prove to be accurate, then Agere, Altera,
AMCC, Analog Devices, Atheros, Broadcom, Cypress,
Conexant, Intersil, Lattice, National, ON, PMC-Sierra,
Silicon Laboratories, and Xilinx will have experienced a
double-digit decline in their quarterly revenue run rates
(from their most recent revenue peaks).
Exhibit 3
Semiconductor shipments have meaningfully exceeded
end-market sales during the first three quarters of 2004
(Semiconductor industry and end-market revenues
normalized to 1.0 in 3Q02)
1.00
1.10
1.20
1.30
1.40
1.50
1.60
3Q02 4Q02 1Q03 2Q03 3Q03 4Q03 1Q04 2Q04 3Q04 4Q04E
Normalized End Mkt Revenue Normalized Semiconductor Revenue
Source: Morgan Stanley research.
In order to continue to work through excess channel
inventories, we expect the overall semiconductor industry to
decline at least 10% from the record highs reached in
November, and many other semiconductor companies will
need to experience double-digit declines before sustainable
growth can resume. Based on this expectation, we believe
that the bias for revenue and earnings revisions will likely
be lower as the fourth-quarter reporting season unfolds. We
currently expect the global semiconductor industry to be flat
to down 2% sequentially in the fourth quarter of 2004, and
this should at least be followed by a mid single-digit
sequential decline in the first quarter of 2005. On an annual
basis, we estimate that global industry revenues grew about
14% in the fourth quarter versus 28% growth in the third
quarter and a cyclical peak of 40% growth in the second
quarter.
Excess capacity should pressure semiconductor prices
and capital spending in 2005…
Due to difficult comparisons, slower economic growth, and
a less favorable pricing environment, we expect year-overyear
revenue growth to continue to decline through the
summer, with a trough around no growth sometime between
June and August. However, we would not be surprised to
see trough comparisons end up being 5-10 percentage points
lower. Provided end-market demand holds up reasonably
well (excluding normal seasonal weakness in the first
quarter, then we would expect the current inventory
correction within the overall semiconductor supply chain to
be completed sometime during the next few months. Once
the inventory correction is completed, we believe that
excess capacity within the overall semiconductor industry
will promote a more aggressive pricing environment for
new orders until capacity utilization rates increase back
above the 80–85% level.
Due to a significant cutback in capital spending in 2001 and
2002 and an acceleration in global GDP in 2003, capacity
constraints began to develop toward the end of 2003. This
phenomenon led to an increase in lead times and the
incentive for OEMs and their channel partners to build
inventory of semiconductor components last year. By the
middle of last year, global semiconductor industry capacity
utilization increased to around 95%, and as the utilization
rates increased, capital-spending growth accelerated to
about 46% in 2004 versus 16% in 2003. Due to the
combination of a slowdown in global GDP and the ill
effects of the ongoing inventory correction, we estimate that
overall capacity utilization rates have declined sharply since
then.
Given an increase in the level of excess capacity, we expect
capital spending to decline in 2005. However, we have
adjusted our global semiconductor capital spending forecast
to reflect a 11% decline rather than an 18% decline
previously, as several of the major memory manufacturers
(such as Micron and Samsung) have increased their
spending forecasts. In addition, we believe that Intel has
already accrued nearly all of the capital efficiency benefits
of moving to 300-mm production, and a sharp increase in
their 2005 spending plans help to confirm this view.
Regardless, the members of our global semiconductor
capital equipment team still expect equipment orders to
decline sequentially until a trough is reached in the second
quarter of this year.
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 7
…But a better environment and higher growth should be
recorded in 2006
Although the near-term challenges associated with the
semiconductor industry’s fundamentals are significant and
they make it difficult to feel confident about longer-term
projections, we believe that the environment will improve
next year. Our initial estimate for global semiconductor
industry revenue growth in 2006 is 8–12%, and we believe
that easy comparisons and a more benign pricing
environment (vis-à-vis 2005) caused by higher capacity
utilization levels should help to support stronger revenue
growth (Exhibit 4). From an end-market perspective, the
transition to higher-content 3G cell phones, digital TVs and
other high-end digital consumer electronics products, and a
more aggressive roll-out of triple play (voice, video, and
data) services within the digital home should represent some
of the demand drivers.
Exhibit 4
Semiconductor industry revenue growth has slowed but
should recover in 2006
(Year-over-year semiconductor industry revenue growth)
-35%
-25%
-15%
-5%
5%
15%
25%
35%
45%
55%
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
00
01
02
03
04E
05E
06E
% Growth
Semiconductor Growth
10-Year CAGR
Source: Morgan Stanley research.
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 8
Industry Views
US – Semiconductor Industry View: In-Line
Although valuations are reasonable, we believe that near- to intermediate-term fundamentals will remain challenging, and
the potential for additional reductions in forward revenue, margin, and earnings estimates remains high.
Taiwan – Foundry Industry View: In-Line
Near term cyclical concerns are largely factored in. We believe industry valuations are reasonable.
Korea – Semiconductor Industry View: Attractive
Past investments are paying off in the form of operating leverage and earnings growth, mostly for Samsung Electronics.
Singapore – Technology Industry View: In-Line
Cyclical recovery is well priced in and inventory concerns are growing.
China – Technology View: In-Line
The China technology sector benefits from large potential market but suffers from potential overcapacity.
Europe – Semiconductor Industry View: In-line
The upside to fair value for our Universe of stocks (market-cap weighted) is now in-line with our strategists forecast for 7%
upside to MSCI Europe over the next 12 months.
US – Semiconductor Capital Equipment Industry View: Cautious
We believe we are at the start of an industry downturn. Although anticipated, we conclude the depth and duration of the
downturn is greater than currently discounted by the market. Investors will be disappointed at the lackluster rate of growth
once the industry stabilizes, in our view. Slower industry growth means slower recovery from excess capacity.
Japan – Semiconductor Production Equipment Industry View: In-Line
We expect SPE orders will bottom out around the April-June quarter, and to recovery through the latter half of the year.
However, as there do not appear to be any final products that will become major drivers of semiconductor demand, we
expect only a gradual recovery in semiconductor capex.
Taiwan – System & PC Hardware Industry View: In-Line
Morgan Stanley expects the global PC industry to report 9% unit growth in 2005. Pricing environment and margin
contraction remain the main issues.
Japan – Industrial Electronics Industry View: In-Line
We are maintaining our In-Line industry view. Sluggish demand for domestic heavy electrical/industrial equipment and IT
software/hardware, as well as sluggish semiconductor business, should be already discounted in the share prices. Looking for
bargain hunting.
Taiwan – PC-Related Chip Design Industry View: In-Line
Companies face foundry price increase pressure and an intensifying competitive landscape due to single-digit PC growth.
US – Electronics Supply Chain Industry View: In-Line
Consensus 2005 revenue estimates appear aggressive in light of cyclical pressures. However, EMS companies should at
least preserve margins as we expect a milder trough combined with restructuring benefits.
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 9
Global Best Idea
Samsung Electronics
(005930.KS, W 435,000, Overweight; target W548,000)
Samsung seems well positioned to weather a downturn in
the DRAM market. Based on its technology roadmap, we
believe it can withstand 30% ASP erosion in the DRAM
business and 40% ASP erosion in the NAND flash
business without sacrificing profit margins significantly.
Samsung tends to outperform various global tech sectors
during cyclical downturns due to its neutral beta and
diversified profit structure. DRAM industry risks are on
the positive side, in our view. While industry laggards are
catching up on the 110 nano process, the speed of DRAM
chips has not improved vastly. Difficulty for Samsung in
ramping up its 90 nano DRAM process could be another
early indication that supply side could see further
constraints in 2005. Bottlenecks in DDR-2 tester capacity
favor integrated DRAM producers such as Samsung. In
the NAND industry, end-market demand is responding
well to producer price cuts, manifesting good price
elasticity. The productivity increase was much better than
we expected in 3Q04, with 90 nano process technology
comprising 65% of NAND output versus only 20% in the
previous quarter. The product remains highly profitable
despite the sharp price cuts.
The handset business should turn around in 1Q05. New
product launch misses and weak domestic demand led to a
shortfall in 4Q04 ASP and margins. The company plans to
launch 15 new handset models at higher ASPs that are
likely to reverse the margins trends in its handset business.
We estimate that, at 9.2x 2005 earnings and 1.5x trailing
book value, Samsung’s stock fully discounts the 32% EPS
decline we expect in 2005. Margins at the LCD and
handset divisions look set to turn around soon, and we
expect this to set a more positive tone for the share price.
Past investments are paying off in the form of better cost
structure and faster technology migration, especially for
Samsung Electronics. Strong global memory
semiconductor demand favors Korea. With the likelihood
of investors returning to global cyclical stocks at an early
date and appealing valuations, we hold an Attractive view
on the South Korea Semiconductor industry and maintain
our Overweight rating on Samsung Electronics shares.
Upside risk to our target could come from higher global
PC units shipments than our expectations and delays in
flash chips by Renesas and Infineon. Downside risks could
come from re-emergence of Hynix as a formidable
opponent in the DRAM and NAND, and DSC inventory
buildup.
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 10
Regional Best Ideas
Semiconductors – US
National Semiconductor
(NSM, $16.95, Overweight-V; target $22)
Due to an inventory correction in the distribution channel
and somewhat weaker-than-expected end demand for cell
phones, PCs, displays, and consumer electronics, National
Semiconductor’s November quarter represented the
company’s second consecutive negative earnings surprise.
Now that the company’s revenues are 21% below the level
reached two quarters earlier, we believe that most of the risk
has been eliminated, and we believe that the company’s
overall revenues are in the process of approaching a
sustainable trough. Despite the decline in revenues,
National has managed to maintain solid margins, and
management’s actions suggest to us that the company is
more focused on the bottom line than ever before. National
has a strong balance sheet, with about $2.30 of net cash per
share, and the stock is currently trading at a slight discount
to the average 2005E P/S multiple of the stocks in our
universe, while its profit margins suggest to us that it should
trade at a premium. New product execution and the health
of the overall economy are the primary risks we see to our
earnings estimates and overall investment thesis on NSM.
NVIDIA
(NVDA, $22.03, Overweight-V; target $36)
We continue to believe that NVIDIA is better positioned
now than at any time during the last couple of years. The
company reported a large positive earnings surprise on
strong revenue growth, expanding gross margins, and tight
operating expense controls in the October quarter. A strong
ramp in the company’s GeForce 6000 GPU product line
drove the positive surprise, and when combined with strong
growth in core logic chipsets and media processors for
cellular phones, we expect additional momentum to be
achieved in the January quarter and positive surprise
potential should remain high. We expect accelerated
revenue growth and meaningful margin expansion during
the next 1–2 years thanks to license and royalty revenues for
Sony’s PlayStation 3, media processors for cellular phones,
and core logic chipsets for notebook PCs, servers, and Intelbased
processors. NVIDIA has a strong balance sheet
(about $3.40 of net cash per share), and NVDA is trading at
a 50% discount to the average 2005E P/S multiple of the
stocks in our universe. New product execution, the health
of the PC market and consumer demand, and yields on the
new GeForce 6000 products manufactured by TSMC and
IBM are the primary risks we see to our earnings estimates
and investment thesis on NVDA.
Industrial Electronics – Japan
NEC Electronics
(6723.T, ¥4,810, Overweight; target ¥6,200)
We expect that the semiconductor market will return to
growth over the next year. NEC Electronics, as a specialist
in this area, is a strong investment candidate, in our view.
Our target price for NEC Electronics is ¥6,200, 28% higher
than the current price. Risks include a delay in
semiconductor recovery and related inventory buildup/
sharp price erosion.
Semiconductors – Europe
ASML Holding (ASML.AS, €14.62, Overweight-V; target
€14)
Although the market-share gain and higher ASP story for
ASML remain compelling in the medium term, we believe
earnings risk remains due to the continuing caution in the
semiconductor food chain. As we believe that near-term
data points are likely to remain mixed, we expect a better
buying opportunity for stocks in the European
semiconductor group as a whole in the first half of 2005.
Semiconductor Capital Equipment – US
KLA Tencor (KLAC, US$42.97, Overweight; target
US$43.25)
Visibility for KLAC is about the best in the industry, in our
view. KLA-Tencor has managed its business for the sort of
cyclicality we are experiencing currently. The company has
enjoyed a Book-to-Shipment ratio above 1.0 since F3Q03,
and this has allowed the company to build a solid backlog
of orders, which are regularly “scrubbed” to weed out those
orders which management suspect may be liable to push-out
or cancellation. Coupled with significant deferred revenues
and stable service revenues, we conclude that the revenue
outlook is robust as KLA draws on its run rate of revenues
and deferred revenues to generate relatively resilient
earnings performance during this down-cycle.
We remain comfortable that KLA remains well placed in
the process control market and is likely to be a primary
beneficiary of Intel’s migration to the 65nm design node. It
is able to generate attractive levels of shareholder value and
we believe this value creation is a durable feature of the
KLA-Tencor business model and IP portfolio. We continue
to rate the stock Overweight, within our cautious view on
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 11
the SCE industry. Our residual income valuation
framework drives our price target of $43.25/share.
Semiconductor Production Equipment –
Japan
Tokyo Electron
(8035.T, ¥6,040, Overweight; target ¥6,900)
We maintain our Overweight rating on Tokyo Electron. The
semiconductor industry continues its transition to 300mm
wafer lines, and we think that TEL will benefit from this in
the medium to long term as it has taken the lead in making
300mm wafer equipment. In addition, TEL has not backed
off on structural reforms. Our target price values the stock
at a P/E of 26x (the previous two cycles’ peak) on our
F2006 estimates and a P/B of 3.7x on our F2004 estimates.
The main risk factor we see is a prolonged decline in orders.
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 12
Semiconductors — Asia/Pacific
Sunil Gupta (Sunil.Gupta@morganstanley.com)
Dickson Ho (Dickson.Ho@morganstanley.com)
Keon Han (Keon.Han@morganstanley.com)
Frank Wang (Frank.AY.Wang@morganstanley.com)
Key Points
Taiwan Foundry industry view remains In-Line.
Near-term cyclical concerns appear to be largely
factored in. We believe industry valuations are
reasonable.
Korea Semiconductor industry view remains
Attractive. Past investments are paying off in the form
of operating leverage and earnings growth, mostly for
Samsung Electronics. Strong NAND flash demand
globally favors Korea.
Singapore Technology industry view remains In-
Line. Stocks appear fairly valued, but the cycle is
peaking.
MSCI SECTOR TECHNOLOGY HARDWARE & EQUIPMENT
A/P Strategist 12.0%
MSCI AC Asia Pacific Weight 13.3%
Taiwan Foundry: Industry view – In-Line
TSMC (2330.TW, Overweight-V)
Decelerated Decline in December Sales
TSMC reported December sales revenue of NT$19.92
billion, down 5.1% MoM and up 5.0% YoY, roughly in-line
with our expectation. Compared to the 8.6% drop in
November sales, we believe that the decline in monthly
revenue had been essentially decelerated. 4Q sales
aggregated to NT$63.88 billion, down 8.4% sequentially,
slightly ahead of our forecast of 10% decline. We now
expect TSMC to deliver flat monthly revenue in January.
Rush orders exist but not significant enough
There has been some anticipation on the market of rush
orders from LCD drivers and GPUs TSMC recently, which
might suggest an earlier-than-anticipated recovery for
TSMC. According to our latest checks, we believe rush
orders from several different sectors did appear in
December, but it is more likely a normal practice that
happens every quarter-end. More importantly, the volume
was not ample enough to offset the weakness throughout all
market segments. We hence hold our expectations of mid
single-digit decline for 1Q utilization rate and 10% drop in
wafer shipments for TSMC. Furthermore, we believe
pricing pressure on TSMC will be relatively mild in 1Q,
thanks to its higher-than-average utilized capacities on
advanced technologies.
High utilization in advanced technologies differentiates
TSMC from peers
As we have often stated, we believe capacities for advanced
technologies (0.13u and below) at TSMC are still utilized at
a level exceeding 90%, especially copper/low-k production
lines (close to 110%). Moreover, with 90nm volume
production in place from 4Q04, we believe TSMC has been
getting decent orders from customers like TI. As a result,
we expect TSMC to double its 90nm wafer shipments, with
sales contribution greater than 2% in 1Q05. We believe the
superior customer base, technology leadership and cost
advantages on 12" wafers are the keys to secure highquality
orders despite on a deteriorating environment. This
is supportive to our statement earlier that TSMC is going to
further differentiate itself with other foundry competitors.
Overweight on TSMC
We believe that most of the negative cyclical news about
excess inventory, excess supply and pricing weakness is
well understood and discounted by the market. While the
risk of a cash call and sharp pricing deterioration remains
high at second tier Foundries, we expect TSMC to widen its
cost leadership due to its 12" Fab strategy, and improve
returns to shareholders via dividends and possibly buy
backs at some stage. We see any weakness in technology
stocks due to excess inventory in end markets post holiday
season as an opportunity to build a position in TSMC shares.
UMC (2303.TW, Equal-weight)
NT$ Appreciation Dragged December Sales Lower
UMC reported December sales of NT$8.88 billion, down
4.3% MoM and up 6.4% YoY, slightly worse than our
expectation. We believe actual wafer shipments in 4Q fell
in the lower end of the -15% to -17% range and ASP remion.
4Q sales aggregated to NT$28.23 billion, in-line with the
company’s guidance. More importantly, we believe it was
NT$ appreciation that caused the lower-than-expected
quarterly sales result.
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 13
Sluggish wafer demand in the near term
Based on our channel checks, we have noticed that order
cuts had been largely muted in December, which suggests a
flat to slightly down monthly revenue in January. Customer
order forecasts seemed to remain weak and conservative
due to excess inventories exiting the year-end. Fortunately,
wafer bookings appeared to stabilize across the board
recently. Hence, our anticipation of low-to-mid 60s
utilization in 1Q (and probably 2Q) remains unchanged.
Sell-through between the Christmas season last year and the
coming Lunar New Year will be rather crucial in the near
term.
Likely downside risk on sharper ASP decline
In order to stimulate demands for mature technologies, we
believe UMC is now offering volume discounts to some
major customers for products using 0.18u and 0.25u
geometries. While we are not sure about the magnitude of
these discounts, we believe it is possible to be at the range
of 7%-10% as what we have seen in 2001. As a result,
sharper ASP decline is likely to appear in 1Q05, earlier than
our expectation (we expected it to happen in 2Q05
previously). Hence, apart from lower utilization rate due to
excess capacity, we believe there could be some more
downside risks to ASP in 1Q05.
Reiterate Equal-weight rating on UMC
We believe the near-term risk for the stock is still on the
downside because of weak fundamentals and the limited
long-term visibility. NT dollar appreciation and pricing
pressures introduced by competitors might also impact on
profit margin and hence earnings. Our Taiwan Foundry
industry view is In-Line.
Singapore Technology: Industry view – In Line
United Test and Assembly (UTAC.SI, Overweight-V)
Ride with MP3 for Now, but Remember to Get Down!
UTAC announced commencement of full turnkey volume
production of Sigmatel’s ("SGTL") audio decoder SOC
MP3 chip (D Major, ~90% of 3Q04 SGTL revenue).
UTAC has previously been testing its Audio Codec chip (C
Major, ~5% of 3Q04 SGTL revenue). We believe the new
contract could add up to 5% additional revenue growth or
US$8m to our 2005 revenue estimates of $202m (+24%
YoY). However, we are keeping our estimates unchanged
since this growth could be neutralized by a slowdown in
orders from Freescale (CDMA chips) & UMCi (wafer
probing) and potential ASP pressure on memory testing
segment in 2005.
Additionally, we expect UTAC to revise up its 3Q04 and
4Q04 revenue due to inclusion of revenue from full turnkey
project from 1" HDD for Seagate. But it does not have any
impact on our net income expectations.
Positive Takeaways
Pragmatic to ride MP3 wave: At the time when wireless
segment is struggling with inventory issues, UTAC has
quickly leveraged its mixed signal expertise to tap highest
growing electronics category (MP3 segment) with the help
of its historical performance with the existing customer.
UTAC is also seeing higher volume from another customer,
Synaptics (makes touchpad for notebooks and wheelpad for
iPod).
Customer acquisition is the key to success: UTAC in the last
six months has successfully acquired more than five quality
new customers and successfully retained them. We believe
UTAC has been successful in not only the diversification of
customers but also transforming itself into the Tier 1
supplier for the existing customers (Broadcom, Nanya).
No new investment required for this customer: Since the
company has the requisite equipment and skills to do the
full turnkey solutions for the SGTL chip, it will help
increase the utilization without adding new capex.
Key Risks
Competition from existing testing partners for SGTL:
Currently, SGTL outsources its chip assembly and testing
services to four subcontractors including ASAT, UTAC,
ASE and Signetics. Management is expecting $10–20m of
revenue from SGTL, which is optimistic, in our view. Our
estimate of $8m revenue for 05 is also based on the
assumption of UTAC getting 30% of total outsourcing.
Considering that the other three ATS players are equally
competent, in our view, there could be risk to our and
management expectations.
Share overhang and option dilution: Despite strong
operational performance, UTAC shares have been diving
down in the last six months due to share overhang (roughly
40% shares with VC and private equity). We believe this
could remain a drag on the stock performance. In addition,
last quarter the company issued 21 million options to 242
employees of the group at a strike price of S$0.59 in this
quarter. Currently, there are 69.5m options are outstanding,
which could dilute EPS by 8% if exercised. Half of the
total options outstanding (36.7m) are currently deeply in the
money with an exercise price of US$0.1875.
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 14
Investment Conclusion
UTAC is currently trading at 1.0x P/BV 04E and 16.0x P/E
based on 2005E. The stock has already rallied 20% since it
hit the bottom of $0.44 last month. We remain constructive
on the stock with an Overweight-V rating, but we caution
investors to watch out for consensus earnings coming down
due to high depreciation and pricing pressure.
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 15
Semiconductors — Japan
Naoki Sato (Naoki.Sato@morganstanley.com)
Kazuo Yoshikawa (Kazuo.Yoshikawa@morganstanley.com)
Key Points
Japan industrial electronics industry view is In-Line
We are maintaining our In-Line industry view. Sluggish
demand for domestic heavy electrical/industrial
equipment and IT software/hardware, as well as
sluggish semiconductor business, should be already
discounted in the share prices. Looking for bargain
hunting.
Sequential recovery hoped for, but uncertain in 4Q
All industry firms hope to recover in 4Q from inventory
corrections but there is a risk of an earnings shortfall if
the recovery is delayed.
F1H04 aggregate semiconductor sales for Japan’s five
industrial electronics companies (Toshiba, NEC, Fujitsu,
Mitsubishi Electric and Hitachi) increased 10.0% YoY.
Growth was much stronger in real terms since, in F1H03,
sales that were not yet transferred to Renesas remained with
Hitachi and Mitsubishi, while Fujitsu had consolidated its
JV subsidiary with AMD.
Logic, flash memory, LCD drivers, image sensors, and
discrete semiconductors, which were used in cell phones,
personal computers, and digital appliances all saw major
YoY growth.
Segment operating income more than doubled from ¥61.6
billion in the same period a year earlier to ¥128.2 billion
thanks to higher utilization of fabs and a favorable price
environment. Improved profitability of semiconductor
equipment at Hitachi High-Technologies also helped.
On a quarterly basis, however, the YoY revenue growth
dropped from almost 12% in April-June to under 9% in
July-September and operating income slumped from about
¥70 billion to roughly ¥58 billion. The reasons for this
include (1) an inventory adjustment in cell phones in Japan
and abroad, (2) decreased speed in drivers as oversupply of
LCDs developed, and (3) production adjustments in
recordable DVD players and drives. Profit margins dropped
for flash memory as prices plunged toward end-quarter.
Semiconductor earnings are likely to deteriorate markedly
in 2H. We forecast sales to dip 1% YoY and operating
income to drop by nearly half. The decline is similar to 1H,
which was about the same level as F2H03 earnings.
3Q looks to be especially tough, as a full-blown adjustment
seems likely. Drivers for large LCD panels and logic LSIs
for cell phones and digital appliances are likely to feel the
effects of inventory corrections in final products. Flash
memory should be strong in volume, but the decline in
average unit price inherited from the 2Q will erode
profitability.
All firms, however, expect to see gains in 4Q. NEC
Electronics guidance projects a 4% YoY revenue drop in
3Q but a 5% YoY gain in 4Q. Behind this projection is
sentiment that final product inventories are not that large
and the market is headed for a recovery as the adjustment
starts promptly.
We agree as far as direction is concerned. We believe that
this adjustment and recovery will behave like the curve
shown in Exhibit 5.
Exhibit 5
Global Demand for Final Products vs. Components
Supplies
1-3 4-6 7-9 10-12 1-3 4-6
04 05
Dem ands for
Final Products
C om ponents
Production/Supply
Adjustment
R ecovery
Source: Morgan Stanley Research
Evidence backing up a pattern like the exhibit includes the
fact that, since last year, semiconductor monthly shipment
trends have not departed far from ordinary seasonality, and
the fact that US high-tech inventories are low compared to
past levels, although they have somewhat increased.
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 16
Final product demand at the beginning of this year had
strength not seen in recent years. Not only were the markets
for notebook PCs and high-end cell phones strong, sales of
digital appliances also grew strongly, helped by the
Olympics and launches of new products. This kept
semiconductor demand so strong that many chipmakers had
a hard time keeping up with orders, namely in image
processing LSIs and LCD drivers. They thus increased
production capacity.
Inventories started to grow, however, for some products
between early spring and summer. This happened in cell
phones (China), digital cameras (Japan and Taiwan), large
LCD panels (Taiwan and South Korea), and DVD recorders
(China and Taiwan).
As a result, customers brake hard for orders in cell phone
LSIs, drivers for large panels LCDs, digital appliance LSIs,
and flash memory; the start of an adjustment in Aug-Sep.
Although demand for final products slumped, it should not
be huge in our view. We expect that when inventories of
parts and products accumulated in Jul-Sep are to be
exhausted, part orders will recover to the level of actual
demand for final product.
The key question is the timing and extent of the recovery.
If personal computers, one of the major applications that
uses semiconductors, retains its relative strength, it is hard
to see the semiconductor market overall falling any further.
We also have been seeing expectations for year-end demand
somewhat subdued this time. Judging from the Oct-Dec
guidance of many high-tech companies, neither product nor
component vendors are going all out in production and
shipment. If this holds up, the inventory adjustment should
work out smoothly, so long as there is not a major drop in
final demand. If demand instead rises even a little,
manufacturers could run out of inventory at year end.
In this scenario, orders for components could return to the
level of actual demand for final products quickly. This
would occur in Jan-Mar at the earliest or Apr-Jun at the
latest, in our view.
If the recovery does not take place in Jan-Mar period,
however, there is a risk that Japanese chipmakers may warn
in the coming months of earnings shortfalls for F04 ending
March. Should this happen and share prices decline hereby,
we view this as a good opportunity to accumulate stocks
highly leveraged to semiconductors.
Exhibit 6
Semiconductors Segment: Net Sales & Earnings Trends
(¥ billion) FY03 FY04 YoY YoY
FY 1H 2H 1H 2He 1H 2He FY03 FY04e FY05e FY04e FY05e
Semicon
Consolidated Sales YoY % change YoY % change
NEC 360.4 363.9 376.8 336.2 4.6% -7.6% 724.3 713.0 750.0 -1.6% 5.2%
Fujitsu 182.3 221.6 219.9 210.1 20.6% -5.2% 403.9 430.0 453.0 6.5% 5.3%
Hitachi 35.0 36.2 35.5 31.5 1.4% -13.0% 71.2 67.0 70.0 -5.9% 4.5%
Toshiba 432.5 466.3 489.5 462.5 13.2% -0.8% 898.8 952.0 1,029.0 5.9% 8.1%
Melco 65.5 58.9 62.0 74.0 -5.4% 25.6% 124.4 136.0 143.0 9.3% 5.1%
Total 1,075.7 1,146.9 1,183.7 1,114.3 10.0% -2.8% 2,222.6 2,298.0 2,445.0 3.4% 6.4%
Consolidated OP YoY % change YoY % change
NEC 24.6 32.9 30.7 5.3 24.4% -83.8% 57.5 36.0 51.0 -37.4% 41.7%
Fujitsu (6.7) 15.7 23.0 8.0 NM -49.0% 9.0 31.0 35.0 244.4% 12.9%
Hitachi (1.0) 2.2 6.4 0.6 NM -72.7% 1.2 7.0 9.0 483.3% 28.6%
Toshiba 44.0 74.4 64.5 35.5 46.6% -52.3% 118.4 100.0 133.0 -15.5% 33.0%
Melco 0.7 (0.7) 3.7 4.4 458.3% NM 0.0 8.1 11.0 NM 35.7%
Total 61.6 124.5 128.2 53.9 108.2% -56.7% 186.1 182.1 239.0 -2.1% 31.2%
Consolidated OP Margin (%) YoY change (pp) YoY change (pp)
NEC 6.8% 9.0% 8.1% 1.6% 1.3% -7.4% 7.9% 5.0% 6.8% -2.9% 1.8%
Fujitsu -3.7% 7.1% 10.5% 3.8% 14.1% -3.3% 2.2% 7.2% 7.7% 5.0% 0.5%
Hitachi -2.9% 6.1% 18.0% 1.9% 20.9% -4.2% 1.7% 10.4% 12.9% 8.8% 2.4%
Toshiba 10.2% 16.0% 13.2% 7.7% 3.0% -8.3% 13.2% 10.5% 12.9% -2.7% 2.4%
Melco 1.0% -1.1% 5.9% 6.0% 4.9% 7.1% 0.0% 6.0% 7.7% 6.0% 1.7%
Total 5.7% 10.9% 10.8% 4.8% 5.1% -6.0% 8.4% 7.9% 9.8% -0.4% 1.9%
Source: Company, Morgan Stanley Research
E=Morgan Stanley Research estimates
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 17
Semiconductors — Europe
Stuart Adrian (Stuart.Adrian@morganstanley.com)
Kirsten Parker (Kirsten.Parker@morganstanley.com)
Dieter Weber (Dieter.Weber@morganstanley.com)
Key Points
ASML Holding: We believe ASML is likely to make a
decision to enter into the LCD stepper market within the
next few months.
In case of an affirmative decision, we believe the nearterm
financial impact will be limited. In our “mostlikely
” case scenario, we expect the new LCD segment
to account for only 6% of group revenues.
GICS SECTOR INFORMATION TECHNOLOGY
EU Strategist Weight 3.5%
MSCI Europe Weight 4.0%
ASML Holding: To Enter or Not to Enter?
ASML has indicated on numerous occasions that it is
considering a potential entry into the market for LCD
steppers. We believe the company is likely to make a
decision some time over the next few months. While we
originally expected the company to make a decision before
the end of 2004, we believe the arrival of its new CEO, Eric
Maurice, in September 2004 has potentially pushed out the
announcement of the official decision by a few weeks.
Nevertheless, we expect investors continue to be focused on
this topic.
In our latest report on ASML entitled “ASML Holding: To
Enter or Not to Enter?” dated December 1, 2004, we
looked in detail at the potential impact on ASML’s financial
performance as well as the company’s competitive position
and the potential drivers for ASML in case of an affirmative
decision.
We believe that the financial impact on ASML from a
move into LCD steppers is likely to be limited in the
medium term. During the period from the initial decision
(2005-1H06), ASML would only manufacture a very small
number of test (or beta) tools and focus on the set-up of its
supply chain. However, even by 2006-08, we believe the
financial impact would remain relatively limited. We
model three different scenarios for potential outcomes if
ASML decides to enter the LCD stepper market. In our
‘Likely Scenario’ we model revenues from the LCD
segment of €270 million in 2008 and breakeven on the
operating profit level — see Exhibit 7 (our three-stage
scenario model is available in our December 1 report
mentioned above). This would represent about 6% of total
group revenues and would reduce the overall group EBIT
margin by just 100bps in 2008.
In addition, we believe ASML is also likely to evaluate
various funding options, such as government grants or
funding from major clients, to limit its financial
commitments.
Exhibit 7
ASML: Most Likely Scenario for the Financial
Performance of the LCD Segment and ASML Group
Most Likely Scenario
2006e 2007e 2008e
LCD Stepper Segment
LCD Model 2006 2007 2008
ASML Units 3 15 30
~ Market Share (%) 2.4 8.0 17.1
ASP (€)
ASML 6.0 7.5 9.0
~ Premium 44.0 50.0 50.0
Revenues 18.0 90.0 270.0
Gross Profit 0.9 11.3 60.8
Gross Margin (%) 5.0 15.5 25.5
Opex 25.0 35.0 60.0
Operating Income (24.1) (23.8) 0.8
Operating Margin (%) -134 -26 0
IC Stepper Segment (Forecasts for this segment are independent of
LCD scenario)
Revenues 2,705.5 3,310.5 3,891.4
Gross Margin (%) 38.0 38.5 38.0
Operating Income 498.9 582.8 623.2
Operating Margin (%) 18.4 17.6 16.0
1,843.9 1,760.4 1,601.4
Combined Entity
Revenues 2,723.5 3,400.5 4,161.4
Gross Profit 1,029.0 1,585.8 1,539.5
Gross Margin (%) 37.8 37.8 33.0
Operating Income 474.8 559.0 623.9
Operating Margin (%) 17.4 16.4 15.0
Impact on Operating Margin (%) -1.0 -1.2 -1.0
Source: Morgan Stanley Research estimates
We believe the two major risks in entering into the LCD
stepper market are, first, the potential for poor internal
execution leading to operating losses and, second, the
likelihood that the rate of technological change (specifically
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 18
the substrate size) will slow beyond 7G/7.5G. We believe
there is the likelihood that 7G substrates represent the most
viable solution for LCD panel manufacturers as 7G/7.5G
substrates are likely to be large enough to address 90% of
the display market. This would eliminate one of the drivers
for LCD panel manufacturers to upgrade their lithography
tools. Additional tool demand would be driven only by
capacity rather than technology as well.
Why is ASML considering a move into the LCD
lithography market?
As the market for LCD lithography equipment is
meaningfully smaller than the market for semiconductor
lithography equipment (in value as well as in unit terms),
one of the major questions is why entry into this market
would be attractive. Although ASML has not discussed the
reasons behind its plans, we believe there are a number of
drivers that have prompted ASML to look at this
opportunity.
• The LCD stepper market is profitable. We believe
that the LCD stepper business is the main driver for
Nikon’s profitability target for its “Precision
Equipment” segment (this combines its IC and LCD
steppers) of ¥0.9 billion in FY2004. Although Nikon
has historically not provided a detailed breakdown of
the profitability by product, it has indicated that it
expects the LCD business segment, despite its smaller
size, to offset a “no-better-than-break-even” situation in
its IC stepper segment. In addition, we also believe
Canon enjoys a better profitability structure in its LCD
stepper business than in its IC stepper division.
• Solid outlook for long-term growth rate of end
markets. Our Taiwan LCD analyst Frank Wang is
positive on the long-term outlook for TFT LCD
technology. He believes that this is the most
economically scaleable flat panel display technology
for end-products such as notebooks, monitors, and TVs.
Specifically, he believes that over the next 15 years
LCD TVs are likely to replace most CRT TVs.
• Both markets have a strong focus on optical and tool
engineering. As outlined later in this report, LCD
lithography has many similarities to the technology
used in IC lithography. We believe ASML will benefit
from its strong experience in optics and tool
engineering.
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 19
Semiconductors — Electronic Design Automation (EDA)
Harlan Sur (Harlan.Sur@morganstanley.com)
Mark Edelstone (Mark.Edelstone@morganstanley.com)
Louis Gerhardy (Louis.Gerhardy@morganstanley.com)
Key Points
Weak 2H04 semiconductor trends lead to cautious
semi R&D spending and weak EDA fundamentals.
For 2004, semi R&D spending should grow about 10–
11% and EDA industry revenue growth should be flat to
up slightly.
EDA industry revenues should grow by low to midsingle
digits in 2005, similar to semi R&D spending
growth. Top fabless semiconductor companies are
beginning to make an accelerated move to the 90 nm
node, and overall we expect 90 nm to be a larger driver
of EDA fundamentals in 2005.
Cadence on track to deliver an in-line December
quarter. We believe Cadence can deliver a positive
book-to-bill, 13% se-quential revenue growth, and EPS
of $0.25 per share.
GICS SECTOR INFORMATION TECHNOLOGY
US Strategist Weight 13.6%
S&P 500 Weight 15.8%
Weak EDA fundamentals in 2004. We believe weak
second-half fundamentals within the semiconductor industry
have resulted in tighter R&D spending controls and
the elimination of discretionary spending on new chipdesign
tools. Although 90 nm chip-design activity
continues to accelerate, the technology transition is in its
infancy, and we expect 90 nm to be a larger driver of EDA
fundamentals in 2005. Given the weak semiconductor
trends in the second half of 2004, we believe that EDA
industry revenue growth will be flat to up slightly in 2004.
Exhibit 8
90 nm Chip-Design Activity Is Accelerating
0
50
100
150
200
250
300
350
400
F1Q03 F2Q03 F3Q03 F4Q03 F1Q04 F2Q04 F3Q04 F4Q04
Source: Synopsys
90 nm and 65 nm chip-design activity should help drive
EDA industry fundamentals in 2005. Advanced chipdesign
activity appears to be fairly robust (see Exhibit 8);
commentary from all four of the major EDA suppliers
indicates that both 90 nm and 65 nm design activity continues
at a healthy pace. Top fabless semiconductor companies
are beginning to make an accelerated move to the 90
nm node, and overall we expect 90 nm to be a larger driver
of EDA fundamentals in 2005. Our current 2005 R&D
spending growth forecast for the 31 companies in our
semiconductor universe is 4–5%, and we believe EDA
industry revenue/bookings should grow by a similar amount
in 2005.
Cadence Design continues to be our favorite name in the
EDA space. Cadence (CDN, $13.34, Equal-weight-V,
target $18) is set to report December-quarter results on
February 3, and we expect the company to deliver a positive
book-to-bill, 13% sequential revenue growth, and EPS of
$0.25 per share. Led by a new management team, Cadence
appears to be benefiting from its focus on digital integrated
circuit (IC) designs at the 90 nm node, and a solid renewal
cycle should help buffer the company from near-term
perturbations in semiconductor industry R&D spending.
Overall, we believe the company will grow bookings by 5–
6% in 2004. In ad-dition to a solid bookings and earnings
outlook, CDN’s valuations are attractive, we believe. CDN
is trading at about 16.5 times our C2005 adjusted EPS
estimate of $0.80 — a 40% discount to the average stock in
our universe.
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 20
Three-Month Semiconductor Billings: Jan. 2001 – Nov. 2004 (US$, ‘000)
NORTH AMERICA EUROPE JAPAN ASIA PACIFIC WORLD
Billings % Change Billings % Change Billings % Change Billings % Change Billings % Change
January-2001 $5,122,554 13% $3,567,697 13% $3,529,684 7% $2,715,677 -30% $16,629,419 12%
February $4,747,327 6% $3,344,747 6% $3,766,898 15% $3,623,750 -3% $15,482,723 6%
March $4,113,802 -11% $3,279,523 -1% $3,594,299 7% $3,424,762 -10% $14,412,386 -4%
April $3,763,964 -19% $3,075,073 -8% $3,358,671 -2% $3,539,883 -9% $13,737,590 -10%
May $3,352,084 -32% $2,818,591 -17% $3,138,873 -11% $3,398,475 -16% $12,708,022 -20%
June $2,901,611 -45% $2,502,046 -28% $2,917,393 -21% $3,220,260 -24% $11,541,310 -31%
July $2,689,406 -51% $2,252,960 -35% $2,771,781 -29% $3,103,732 -30% $10,817,879 -37%
August $2,600,449 -55% $2,110,313 -41% $2,571,961 -36% $3,162,307 -32% $10,445,030 -42%
September $2,438,631 -59% $2,094,494 -42% $2,411,495 -43% $3,243,225 -31% $10,187,845 -45%
October $2,514,867 -57% $2,201,804 -42% $2,348,456 -46% $3,373,972 -27% $10,439,099 -44%
November $2,501,564 -55% $2,318,728 -39% $2,324,242 -47% $3,459,533 -22% $10,604,067 -42%
December $2,472,105 -55% $2,196,026 -41% $2,126,085 -51% $3,385,114 -21% $10,179,329 -43%
January 2002 $2,471,215 -52% $2,126,767 -40% $2,020,810 -43% $3,399,361 25% $10,018,153 -40%
February $2,521,352 -47% $2,097,270 -37% $2,003,091 -47% $3,406,325 -6% $10,028,037 -35%
March $2,611,087 -37% $2,256,698 -31% $2,109,193 -41% $3,756,138 10% $10,733,117 -26%
April $2,615,317 -31% $2,282,053 -26% $2,201,244 -34% $3,969,853 12% $11,068,467 -19%
May $2,648,872 -21% $2,238,253 -21% $2,326,822 -26% $4,161,414 22% $11,375,361 -10%
June $2,584,874 -11% $2,146,477 -14% $2,452,241 -16% $4,170,099 29% $11,353,691 -2%
July $2,597,560 -3% $2,154,642 -4% $2,660,508 -4% $4,262,696 37% $11,675,406 8%
August $2,580,077 -1% $2,213,996 5% $2,754,351 7% $4,377,584 38% $11,926,008 14%
September $2,632,933 8% $2,316,440 11% $2,828,289 17% $4,511,825 39% $12,289,487 21%
October $2,646,277 5% $2,459,336 12% $2,851,703 21% $4,555,107 35% $12,512,423 20%
November $2,693,427 8% $2,608,623 13% $2,847,383 23% $4,678,461 35% $12,827,893 21%
December $2,596,265 5% $2,543,211 16% $2,774,827 31% $4,613,843 36% $12,528,146 23%
January 2003 $2,557,871 4% $2,471,243 16% $2,726,265 35% $4,525,972 33% $12,281,351 23%
February $2,440,849 -3% $2,452,177 17% $2,720,105 36% $4,302,668 26% $11,915,799 19%
March $2,458,861 -6% $2,528,085 12% $2,879,576 37% $4,432,454 18% $12,298,975 15%
April $2,446,378 -6% $2,543,167 11% $2,876,221 31% $4,534,173 14% $12,399,939 12%
May $2,527,425 -5% $2,466,328 10% $2,981,235 28% $4,675,222 12% $12,650,209 11%
June $2,534,256 -2% $2,423,888 13% $3,024,594 23% $4,714,323 13% $12,697,061 12%
July $2,554,711 -2% $2,478,215 15% $3,165,964 19% $4,854,508 14% $13,053,398 12%
August $2,610,276 1% $2,543,919 15% $3,212,108 17% $5,197,826 19% $13,564,128 14%
September $2,751,306 4% $2,712,851 17% $3,361,248 19% $5,620,410 25% $14,445,816 18%
October $2,936,871 11% $2,961,638 20% $3,558,398 25% $5,971,596 31% $15,428,503 23%
November $3,055,105 13% $3,147,874 21% $3,719,463 31% $6,194,396 32% $16,116,837 26%
December $3,032,471 17% $3,105,185 22% $3,715,314 34% $6,180,351 34% $16,033,321 28%
January 2004 $2,934,760 15% $2,941,629 19% $3,614,778 33% $6,063,900 34% $15,555,067 27%
February $2,931,305 20% $2,961,724 21% $3,555,620 31% $6,142,970 43% $15,591,619 31%
March $3,076,135 25% $3,144,343 24% $3,606,558 25% $6,459,900 46% $16,286,936 32%
April $3,191,415 30% $3,223,835 27% $3,669,933 28% $6,902,617 52% $16,987,800 37%
May $3,197,703 27% $3,186,391 29% $3,721,921 25% $7,231,822 55% $17,337,838 37%
June $3,287,891 30% $3,123,128 29% $3,834,987 27% $7,590,354 61% $17,836,360 40%
July $3,295,095 29% $3,163,242 28% $3,915,310 24% $7,646,082 58% $18,019,729 38%
August $3,417,598 31% $3,217,855 26% $3,975,814 24% $7,671,071 48% $18,282,338 35%
September $3,420,532 24% $3,348,456 23% $3,953,481 18% $7,795,491 39% $18,517,960 28%
October $3,512,622 20% $3,463,804 17% $3,882,417 9% $7,925,989 33% $18,784,833 22%
November $3,454,990 13% $3,609,033 15% $3,913,203 5% $8,045,341 30% $19,022,568 18%
Source: World Semiconductor Trade Statistics (WSTS)
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 21
Semiconductor Capital Equipment — US
Muted recovery from 2Q05 trough may pressure valuation multiples
Timm Schulze-Melander, CFA
(Timm.Schulze-Melander@morganstanley.com)
Conclusion
We remain concerned as to the valuation implications
of a muted capex recovery in 2H05 going into 2006.
Correspondingly we anticipate downside risk to stock
prices from here and retain our cautious industry rating.
However, DRAM and flash memory capex in 2005
appears higher than we thought. 300mm expansion
plans by memory makers are being underpinned by
relatively robust DRAM pricing. In our survey estimate
memory capex accounts for 33-34% of spending
compared with 23-24% usually. As a result, new orders
and revenues in the capital equipment sector are
potentially more exposed to changes in DRAM pricing
through 2005.
In addition, Intel has launched a sizeable investment
program ($5.1bln) to launch 65nm production in 2005.
We prefer exposure to the process control segment of
the semi cap industry as it will continue to benefit from
Moore’s Law and chipmakers pushing mature
technologies into new process nodes. KLAC (O/W PT
$43.25) remains our top pick.
What’s new?
We estimate that semi cap equip bookings have
declined 15-20% QoQ through December 2004 and are
half way done. We calculate that order bookings have
another 10-20% downside from here. We are not
changing our outlook for a trough in order bookings in
2Q05 and a muted recovery through into 2006
Our proprietary 2005 Global Semi Capex survey of 90
chipmakers has been revised up by 8.3% from $37.8bln
to $40.9bln. This still equates to a 10.5% decline YoY,
despite the $2.9bln increase in our survey estimate.
GICS SECTOR INFORMATION TECHNOLOGY
US Strategist Weight 13.6%
S&P 500 Weight 15.8%
Summary
We conclude that the semiconductor capital equipment
industry is approaching the trough in order bookings at we
go through 2Q05. This decline in bookings activity should
bring the chip industry into equilibrium after it built excess
capacity through 2H04.
From the trough, we calculate that the order bookings
recovery will be relatively muted because low capacity
utilization rates will provide plenty of headroom for the
chip industry to grow into. Specifically, we expect that the
2004 peak in capital spending will not be recovered until the
end of 2006, going into 2007.
In our opinion, this argues for a downward re-appraisal of
valuation multiples for the group and correspondingly, we
anticipate downside risk to current share prices.
Risks to Our Forecasts
The risks to our forecast are predominantly macro factors.
Strong demand in end markets would drive higher
semiconductor demand. This would lift capacity utilization
rates more quickly than we currently anticipate and would
likely result in more accelerated capital spending plans.
Also, Japan capital spending may prove to be more robust
than we currently anticipate.
Similarly, if end market demand is weaker than we expect,
then even greater cuts to capex may occur, and the
subsequent recovery will be more muted than we currently
forecast.
Global Capex forecast: What has changed?
We have made a number of estimate revisions to our 2005
capex survey. Most recently, we have revised up our Intel
capital spending forecast by $1.4bn to $5.1bln, as the
company has launched an aggressive 65nm capacity ramp
plan. The wafer fab equipment (WFE) portion of Intel’s
capex has also increased from around a third to almost a
half for 2005.
In addition, we have reflected higher memory capacity
spending than we had expected early in 4Q04. It seems that
sustained DRAM pricing as well as constructive trends in
NAND flash are both supportive of higher capex trends than
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 22
he had previously anticipated. We have revised up our
AMD capex estimate to reflect better the build out of the
300mm Dresden facility and expansion of leading edge 90-
65nm capacity. Similarly, we have revised up our Micron
capex estimate to reflect a more sizeable build out of their
300mm capacity.
We have lowered our Infineon US$ capex forecasts to
reflect weaker USD/ EUR conversion, although the
underlying capex at the Richmond facility currently appears
stable. In our survey, ST Micro capex is slowing as Crolles
2 is built out and the 300mm facility at Catania remains unequipped.
Also, we have revised lower our Winbond capex
estimate for 2005.
Our largest upward revision is at Samsung. Their memory
expansion plans are significant, including the continued
equipping of line 13 as it doubles in size. Moreover, the
launch of line 14 to manufacture NAND flash (10k wspm
capacity) is now fully reflected in our 2005 estimates.
System LSI spending at Samsung is also larger than we had
thought before, as the company spends to convert older
DRAM fabs to the manufacturing of trailing edge logic
devices such as LCD drivers etc.
For Inotera, we model for some capex growth as deliveries
of previously ordered equipment result in higher cash
outgoings as the full fab equipping continues. However, we
believe that the peak in purchase orders has passed.
Powerchip spending also remains firm, as they equip their
second 300mm DRAM facility.
Currently, there are no changes to our Japan capex forecasts.
The aggregate of these adjustments results in a 2005 capex
forecast of around $40.9 bln in 2005 compared with the
$37.8 bln previously estimated. The net revision to our
estimates is essentially $1.4 bln of additional Intel capex
and a further $1.5 bln of incremental memory capex.
DRAM skew in capital spending is significant
The capex breakdown illustrated in our global capital
spending model shows that 2005 is unusually highly
skewed towards DRAM and flash memory capex. For the
full year we estimate that around 33-34% of capex will be
for memory applications up from a 23-24% average level
over the last 5 years. We conclude that the capex outlook
for any recovery in the second half of 2005 is exposed to
changes in the DRAM and flash memory pricing
environment and the knock on effect this could have on the
capital investment plans of the industry.
“Normal” demand patterns are a key assumption. Our
analysis of the inventory reduction process and modeling
the subsequent recovery in capital investment is influenced
significantly by our estimates of capacity utilization rates.
As a basis, we have assumed that the average sequential
trends in unit demand for semiconductors remains stable
going forward. We have used the average quarterly patterns
established since 1991 and have assumed these to be
“normal”. This equates to around 12% YoY growth in unit
demand. Having assumed normal demand patterns, we
have also assumed that chipmakers reach their days of
inventory (DoI) targets in 1H05. Using these inputs we
have also plugged in our capacity expansion estimates,
based on our revised capital spending forecasts. This allows
us to model the corresponding fab output and utilization
rates.
Valuation
Our residual-income valuation methodology suggests a fair
value price target of $43.25, using an estimated WACC of
11.1% and a terminal growth rate of 5%.
Risks to Our Rating and Price Target
We see a number of strategic and tactical risks that we
believe are relevant to our investment thesis.
1. Weaker semi cycle than expected.
If the semiconductor cycle turns down faster or
more deeply than we anticipate, this would have a
significantly negative effect on the revenue and
earnings prospects of the company.
2. Pricing pressure. If pricing pressure for new
semiconductor capital equipment is worse than we
expect, this might also result in disappointing
profitability and hence a lower share price.
3. Market-share loss. KLA-Tencor currently
dominates most of the market sub-segments in
which it operates. With competitors seeking to
gain share, this could result in lower growth for
KLA than we currently anticipate.
4. We may have misjudged the correct discount
rate at which to value the stock, or overestimated
the company’s long-run growth potential. In both
instances, this would mean our price target and
investment rating are too positive.
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 23
US Semiconductor Equipment Book-to-Bill Ratios
Total U.S. Semiconductor Capital Equipment Book-To-Bill Ratios
0.35
0.55
0.75
0.95
1.15
1.35
92 93 94 95 96 97 98 99 00 01 02 03 04
Front-End Capital Equipment Book-To-Bill Ratios
0.35
0.55
0.75
0.95
1.15
1.35
92 93 94 95 96 97 98 99 00 01 02 03 04
Assembly & Test Equipment Book-To-Bill Ratios
0.35
0.55
0.75
0.95
1.15
1.35
92 93 94 95 96 97 98 99 00 01 02 03 04
.
Source: SEMI
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 24
US Semiconductor Equipment Bookings and Billings (US$ millions)
Total Equipment Three-Month Average Bookings
$0
$500
$1,000
$1,500
$2,000
$2,500
$3,000
$3,500
91 92 93 94 95 96 97 98 99 00 01 02 03 04
Front End Three-Month Average Bookings
$0
$500
$1,000
$1,500
$2,000
$2,500
$3,000
91 92 93 94 95 96 97 98 99 00 01 02 03 04
Assembly & Test Equipment Three-Month Average
Bookings
$0
$100
$200
$300
$400
$500
$600
$700
$800
91 92 93 94 95 96 97 98 99 00 01 02 03 04
Source: SEMI
Total Equipment Three-Month Average Billings
$0
$500
$1,000
$1,500
$2,000
$2,500
$3,000
91 92 93 94 95 96 97 98 99 00 01 02 03 04
Front End Three-Month Average Billings
$0
$500
$1,000
$1,500
$2,000
$2,500
91 92 93 94 95 96 97 98 99 00 01 02 03 04
Assembly & Test Equipment Three-Month Average
Billings
$0
$100
$200
$300
$400
$500
$600
$700
91 92 93 94 95 96 97 98 99 00 01 02 03 04
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 25
US Semiconductor Equipment Bookings and Billings (YoY Change)
Year-Over-Year Change In Total Equipment Bookings
-100%
-50%
0%
50%
100%
150%
200%
250%
91 92 93 94 95 96 97 98 99 00 01 02 03 04
Year-Over-Year Change In Front End Equipment
Bookings
-100%
-50%
0%
50%
100%
150%
200%
250%
91 92 93 94 95 96 97 98 99 00 01 02 03 04
Year-Over-Year Change In Assembly & Test
Equipment Bookings
-150%
-100%
-50%
0%
50%
100%
150%
200%
250%
91 92 93 94 95 96 97 98 99 00 01 02 03 04
Source: SEMI
Year-Over-Year Change In Total Equipment Billings
-80%
-60%
-40%
-20%
0%
20%
40%
60%
80%
100%
120%
91 92 93 94 95 96 97 98 99 00 01 02 03 04
Year-Over-Year Change In Front End Equipment
Billings
-80%
-60%
-40%
-20%
0%
20%
40%
60%
80%
100%
120%
91 92 93 94 95 96 97 98 99 00 01 02 03 04
Year-Over-Year Change In Assembly & Test
Equipment Billings
-100%
-50%
0%
50%
100%
150%
91 92 93 94 95 96 97 98 99 00 01 02 03 04
a
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 26
Semiconductor Production Equipment — Japan
Naoki Sato (Naoki.Sato@morganstanley.com)
Kazuo Yoshikawa (Kazuo.Yoshikawa@morganstanley.com)
Key Points
Our view of the semiconductor production equipment
industry remains In-Line. We expect SPE orders will
bottom out around the April-June quarter, and to
recovery through the latter half of the year. However, as
there do not appear to be any final products that will
become major drivers of semiconductor demand, we
expect only a gradual recovery in semiconductor capex.
The decline in orders in Oct-Dec was mild. We expect
Oct-Dec orders at semiconductor production equipment
companies under our coverage to be down moderately,
by some 10% sequentially. Investment by foundries in
low-end logic chips remained weak, but investment in
DRAMs, NAND flash memories, and MPUs was firm,
though hardly strong. If the DRAM market weakens,
there is risk that C1H05 could see a severe drop in
orders.
Our view of the semiconductor production equipment
industry remains In-Line
SPE orders have been declining since mid-year 2004, and
we expect that they will continue to do so in Jan-Mar and
Apr-Jun 2005, as we expect utilization rates to decline,
mainly at foundries. During this cycle, however, inventory
adjustment has begun early for both semiconductors and
final products. So we expect inventory adjustment will also
finish early. In addition, the industry has maintained a
cautious stance on semiconductor capex through this cycle,
and we have not seen the kind of excessive investment that
characterized previous cycles. So SPE orders are expected
to recover during the latter half of the year. However, as
there do not appear to be any final products that will
become major drivers of semiconductor demand, we expect
SPE capex to drop 11% in 2005, and then to rebound
modestly in 2006.
Japanese SPE stock prices have been rebounding off a
bottom reached in August 2004, and are currently trading at
an average P/B of 2.9 on our F2004 estimates. This
multiple does not look overvalued, but stock prices could go
through another period of weakening, given that we expect
orders to continue declining in Jan-Mar and Apr-Jun 2005.
Currently we expect only a moderate order recovery, and so
we see only limited upside for stock prices even at the time
of the rebound. For this reason, we maintain our In-Line
view.
Only moderate order decline expected for Oct-Dec 2004
Companies are slated to release Oct-Dec order results from
the middle of January, and we expect the firms to post
moderate sequential order declines, limited to some 10%,
due to firm investment in memories, primarily DRAMs.
However, if the DRAM market weakens, there is risk that
C1H05 could see a severe drop in orders.
Exhibit 9
Japanese Semiconductor Production Equipment: Order
Bookings and Book-to-Bill Ratio
0
25,000
50,000
75,000
100,000
125,000
150,000
175,000
200,000
225,000
250,000
94 95 96 97 98 99 00 01 02 03 04 (CY)
(Yen millions)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
Back-end Bookings
Front-end Bokkings
BB ratio
Source: SEAJ, Morgan Stanley Research
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 27
Exhibit 10
Japan Semiconductor Production Equipment: PB Ratio
2.9
1.5 1.8
2.3 2.1 2.1 2.1 2.0
1
2
3
4
5
6
7
8
90 92 94 96 98 00 02 04 (FY)
(x)
Historical Average: 3.3
Note: Market cap average of TEL, Advantest, Hitachi Kokusai and Nikon.
Source: FactSet, Morgan Stanley Research
Tokyo Electron: We think parent SPE-FPD orders for Oct-
Dec now look set to dip by about ¥13.0 billion to ¥120
billion, vs. our previous estimate of ¥110 billion, with SPE
orders down by ¥13.0 billion to ¥114 billion and flat panel
display production equipment (FPD) orders flat at about
¥6.0 billion. SPE orders are supported by ongoing active
investment in DRAMs, NAND flash memories and other
memories. On the other hand, foundry investment remains
lackluster, limited basically to leading edge technology.
We expect orders for Jan-Mar to dip about ¥20 billion
sequentially to ¥100 billion, as memory investment should
be somewhat weak.
Advantest: We look for Oct-Dec consolidated orders to be
about ¥43 billion (-4¥billion or so sequentially). Although
firms are still actively investing in memory (DRAM,
NAND Flash etc), we think investment in LCD drivers to
remain slack.
Advantest assumes Jan-Mar orders will be about ¥55-60
billion, due to increased investment in DDR-II DRAM
(+¥12-17 billion sequentially). We, however, look for the
slump to continue, and assume ¥40 billion or so (-¥3
billion).
Hitachi Kokusai Electric: Oct-Dec parent SPE orders look
to come in slightly above the company’s assumed ¥12
billion or so (+¥2 billion or so QoQ), due to active
investment by Korean DRAM manufacturers.
At the time of its Nov. 1H earnings announcement, the firm
estimated Jan-Mar parent SPE orders would be about ¥7
billion (-¥5 billion). Since this figure didn’t assume a
significant amount of orders from major Korean DRAM
makers, however, we see some upside here.
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 28
Exhibit 11
Japanese Semiconductor Production Equipment: Order Bookings and Sales
Annual Quarter Month
(¥ billion) FY02 FY03 FY04 2003 2004 2004
FY YTD 10-12 1-3 4-6 7-9 6 7 8 9 10 11
Order Bookings 894.0 1,415.4 1,091.9 434.2 402.9 472.8 382.4 164.5 153.2 118.5 110.7 114.3 122.3
Front-end Equipment 629.6 951.3 748.5 305.9 246.9 299.1 273.3 107.4 110.0 86.7 76.6 82.8 93.3
Back-end Equipment 264.4 464.1 499.4 128.3 156.0 173.7 109.1 57.1 43.2 31.8 34.0 31.5 29.0
Sales 857.5 1,167.1 1,030.4 267.4 447.0 397.7 437.6 157.0 142.4 127.2 168.1 84.3 110.8
Front-end Equipment 600.4 773.7 667.4 172.8 297.4 252.7 278.0 102.0 91.6 78.8 107.6 59.3 77.4
Back-end Equipment 257.1 393.5 362.9 94.6 149.6 145.0 159.6 55.0 50.8 48.3 60.5 25.0 33.4
YoY % change
Order Bookings 43% 58% 25% 135% 101% 79% 21% 80% 55% 16% -3% -21% -19%
Front-end Equipment 30% 51% 22% 147% 94% 70% 23% 72% 46% 22% 0% -18% -17%
Back-end Equipment 86% 76% 30% 111% 115% 97% 19% 96% 87% 2% -9% -29% -25%
Sales -3% 36% 70% 58% 58% 137% 54% 94% 84% 64% 29% 35% 20%
Front-end Equipment -14% 29% 64% 42% 52% 148% 38% 97% 65% 38% 21% 49% 20%
Back-end Equipment 36% 53% 82% 98% 73% 120% 91% 90% 135% 137% 46% 11% 21%
Sequential % change
Order Bookings - - - 38% -7% 17% -19% 7% -7% -23% -7% 3% 7%
Front-end Equipment - - - 37% -19% 21% -9% 6% 2% -21% -12% 8% 13%
Back-end Equipment - - - 40% 22% 11% -37% 8% -24% -26% 7% -7% -8%
Sales - - - -6% 67% -11% 10% 26% -9% -11% 32% -50% 32%
Front-end Equipment - - - -14% 72% -15% 10% 27% -10% -14% 36% -45% 31%
Back-end Equipment - - - 13% 58% -3% 10% 24% -8% -5% 25% -59% 34%
Book-to-Bill Ratio 1.04 1.21 1.06 1.62 0.90 1.19 0.87 1.19 1.11 1.02 0.87 0.91 0.96
Front-end Equipment 1.05 1.23 1.12 1.77 0.83 1.18 0.98 1.18 1.16 1.12 0.98 1.00 1.03
Back-end Equipment 1.03 1.18 0.95 1.36 1.04 1.20 0.68 1.20 1.02 0.86 0.68 0.73 0.80
Note: Book-to-Bill ratio is based on 3-month moving average bookings and billings.
Source: SEAJ, Morgan Stanley Research
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 29
PC Systems & Hardware — Taiwan
Ellen Tseng (Ellen.Tseng@morganstanley.com)
Howard Kang (Howard.Kang@morganstanley.com)
Key Points
The top four MB companies likely shipped 7.87 million
units of MBs in December, down 11% MoM and up
24% YoY; this falls within the historical range of -8%~-
14% MoM.
Shipments from top four MB companies grew at 15%
QoQ during 4Q, which is at the mid-lower part of the
historical range of 12-20%; this was mainly caused by
the less compelling platform migration despite
component shortage strengthen 4Q outlook to certain
extent.
We think double ordering resulting from component
shortage as well as mismatch between chipset and CPU
likely support strong 1Q for MB, we expect the
shipment from top 4 MB makers to drop by 8% QoQ.
January MB shipments from the top four in Taiwan
likely come in at 7.30 million units, down 7% MoM, up
10% YoY as a result of regular seasonality and pull-in
effect prior to Chinese New Year.
December shipments from top six NB makers should
decline by 7% MoM and up 35% YoY mainly due to
seasonal effect, the number ourperformed our previous
estimate of down 10% MoM given component supply
from Intel eased off in December thus make up upside.
Based on preliminary number, our 4Q NB model
suggests up 21% QoQ, falls in line with our previous
expectation of 20-23% growth.
We expect January Top 6 NB shipments from top six
will likely be 2.75 million, down 6% MoM and 50%
YoY, in line with the historical pattern. For 1Q05, the
NB shipments should decline by 13% QoQ due to
normal seasonality.
Motherboard Sector Update
From our preliminary datapoints, Asustek, Gigabyte,
Microstar and Elite (ECS) likely shipped 7.87 million MB
units in December, down 7% MoM and up 24% YoY. This
was in line with our prior expectation of 10% MoM decline,
but a better YoY increase volume (18% YoY prior estimate)
given a stronger November shipment base. The magnitude
of sequential decline in December is within the historical
range of -8~-14% MoM, suggesting that 4Q was performing
at a regular seasonal pattern. According to our data, 4Q MB
seasonality came in around the mid-lower end of the
historical range. In the past, top four MB makers 4Q QoQ
growth was in the range of 12-20%, but due to less than
compelling platform migration and low acceptance of Intel
915 platform, 4Q top four MB maker grew at 15% QoQ,
which is at mid-lower end of the historical range. The 4Q
number was consistent with our expectation.
Looking forward, we think that potential double booking
will likely support strong 1Q05 for the MB segment: Our
conversations with industry sources indicate that recent
Intel 865 chipset shortfall (given Intel push to migrate
more sales mix to Intel 915 chipset platform) and mismatch
between CPU socket type and chipset will likely see
double-ordering in 1Q05 to smooth 1Q05 shipment prospect.
MB makers mentioned they see wide availability for Intel
socket 478 CPUs in the market, which require greater
supply of Intel 865 chipset based MBs. Nonetheless, Intel
865 chipset supply still remains as major bottleneck. To
prevent potential 1Q05 shipment disorder and reduce
planning risk for platform migration risk, MB makers will
likely increase chipset inventory for both Intel 865 chipset
(and third party’s 865 equivalent chipset) for Intel socket
478CPU platform and Intel 915 chipset (for Intel LGA775
CPU platform). Meanwhile, the clone market will likely
place orders more aggressively toward both platforms to
build up some buffer.
Our industry sources suggest chipset inventory overbuild by
the end of 1Q05 of about 20–30%. Based on recent
observations, January MB orders will likely remain strong
(flat MoM) as a result of 4Q disorder in 865 shortage.
Despite 1Q05 visibility remaining poor, we think we are
likely to see strong 1Q05 MB shipment outlook (decline
high single digit), which is inflated. We expect top four
MB shipments in January to drop 7% sequentially on top of
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 30
some potential pull-in effect. We look for 1Q05 shipments
to decline by 8% — this is roughly within the historical
pattern of a 5–13% sequential decline for the top four MB
makers in Taiwan.
However, blended ASP for MB may see a significant
correction in 1Q05: MB ASP in 1Q05 will likely decline by
15% QoQ given the significant ASP erosion for 915-based
MBs (and Intel also launched 915PL-low cost version) to
accelerate 915 migration.
Asustek: The company shipped 4.03 million units of MB
during December, down 10% MoM and up 42% YoY,
outperforming our expectation of 3.89 million. Based on
our data collection, AsusTek 4Q MB shipment grew over
26% QoQ, outperforming other MB makers thanks to
noticeable market-share gain. According to our estimate,
January MB shipments should land at 3.9 million, down 3%
MoM, and up 18% YoY; 1Q05 shipments should decline by
7% QoQ.
Gigabyte: We estimate that the company shipped 1.23
million MBs in December or down 15% MoM and 17%
YoY, running in line with our previous expectation. We
estimate Gigabyte 4Q MB shipments to grow to 4.18
million MBs, up 11% QoQ and 14% YoY. We expect
Gigabyte January shipment to decline 10% MoM, or 1.44
million. For 1Q05, we estimated Gigabyte shipments to
decline by 10%.
Microstar: Microstar’s December MB shipments likely to
land on 1.10 million units, this is in line with our previous
expectation. We estimate MSI’s 4Q shipments to drop by
1% QoQ. For January, we expect shipments to be flat. Due
to regular seasonality, 1Q05 shipments should be similar to
its peers, down by 10%.
Elite (ECS): We estimate that December shipments will
drop by 6% MoM, to 1.50 million, better than our previous
expectation of 1.15 million. We estimate ECS 4Q
shipments to grow 7% QoQ. For January, we expect
shipments to drop 13% MoM or a shipment of 1.30 million
MBs.
Exhibit 12
Quarterly Shipments Growth by Taiwan's Top-Four
Motherboard Makers, 1Q 2002 - 1Q (E) 2005 (000 units)
QoQ Asustek Gigabyte Microstar Elite Total
1Q02 -13% -11% -17% -3% -11%
2Q02 -12% -13% -7% -19% -13%
3Q02 22% 39% 37% 30% 31%
4Q02 3% 24% 14% 29% 17%
1Q03 7% -7% 8% -21% -5%
2Q03 19% -18% -22% -15% -7%
3Q03 40% 21% 4% 10% 21%
4Q03 19% 4% 12% 16% 14%
1Q04 5% -1% -8% -14% -2%
2Q04 -17% -10% -11% -9% -13%
3Q04 22% 15% 19% 12% 18%
4Q04 26% 11% -1% 7% 15%
1Q05E -7% -10% -10% -10% -8%
Source: Company Data, Morgan Stanley Research
E=Morgan Stanley Estimates
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 31
Exhibit 13
Monthly Shipments by Taiwan's Top-Four Motherboard Makers, Jan 2002-Jan(E) 2005 (000 units)
(k units) Asustek MoM% Gigabyte MoM% MSI MoM% ECS MoM% Top 4 MoM% YoY%
02-01 1,449 6 950 30 925 -12 1,750 30 5,074 13 41
02-02 1,236 -15 800 -16 880 -5 1,200 -31 4,116 -19 -4
02-03 1,419 15 810 1 900 2 1,250 4 4,379 6 -11
02-04 1,261 -11 730 -10 800 -11 1,050 -16 3,841 -12 8
02-05 1,174 -7 780 7 820 2 1,200 14 3,974 3 23
02-06 1,190 1 710 -9 900 10 1,150 -4 3,950 -1 21
02-07 1,491 25 900 27 1,050 17 1,200 4 4,641 18 26
02-08 1,450 -3 1,050 17 1,100 5 1,440 20 5,040 9 15
02-09 1,490 3 1,130 8 1,300 18 1,790 24 5,710 13 20
02-10 1,850 24 1,360 20 1,550 19 2,030 13 6,790 19 25
02-11 1,400 -24 1,250 -8 1,400 -10 1,800 -11 5,850 -14 10
02-12 1,330 -5 1,200 -4 1,000 -29 1,880 4 5,410 -8 20
03-01 1,650 24 1,250 4 1,350 35 1,887 0 6,137 13 21
03-02 1,400 -15 1,000 -20 1,300 -4 1,400 -26 5,100 -17 24
03-03 1,850 32 1,300 30 1,600 23 1,200 -14 5,950 17 36
03-04 1,850 0 920 -29 1,180 -26 1,140 -5 5,090 -14 33
03-05 1,890 2 950 3 1,000 -15 1,350 18 5,190 2 31
03-06 2,100 11 1,050 11 1,150 15 1,340 -1 5,640 9 43
03-07 2,580 23 1,190 13 1,081 -6 1,408 5 6,259 11 35
03-08 2,560 -1 1,150 -3 1,167 8 1,300 -8 6,177 -1 23
03-09 3,020 18 1,185 3 1,203 3 1,495 15 6,902 12 21
03-10 3,780 25 1,374 16 1,500 25 1,710 14 8,364 21 23
03-11 3,100 -18 1,237 -10 1,300 -13 1,750 2 7,387 -12 26
03-12 2,840 -8 1,050 -15 1,050 -19 1,400 -20 6,340 -14 17
04-01 3,300 16 950 -10 1,100 5 1,260 -10 6,610 4 8
04-02 3,300 0 1,280 35 1,200 9 1,500 19 7,280 10 43
04-03 3,600 9 1,400 9 1,260 5 1,410 -6 7,670 5 29
04-04 3,000 -17 1,220 -13 1,071 -15 1,340 -5 6,631 -14 30
04-05 2,710 -10 1,050 -14 964 -10 1,273 -5 5,997 -10 16
04-06 2,800 3 1,000 -5 1,150 19 1,180 -7 6,130 2 9
04-07 3,100 11 1,100 10 1,200 4 1,260 7 6,660 9 6
04-08 3,400 10 1,210 10 1,100 -8 1,340 6 7,050 6 14
04-09 3,910 15 1,451 20 1,500 36 1,630 22 8,491 20 23
04-10 4,560 17 1,500 3 1,350 -10 1,440 -12 8,850 4 6
04-11 4,480 -2 1,450 -3 1,300 -4 1,600 11 8,830 0 20
04-12 4,032 -10 1,233 -15 1,105 -15 1,500 -6 7,870 -11 24
05-01E 3,900 -3 1,000 -19 1,100 0 1,300 -13 7,300 -7 10
Source: Company Data, Morgan Stanley Research E=Morgan Stanley Estimates
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 32
Notebook Segment Update
From our preliminary data points, the top-six notebook PC
makers — Quanta, Compal, Wistron, Inventec, Arima and
Asustek — likely shipped a total of 2.92 million notebook PC
units in December, down 7% MoM and up 35% YoY,
outperforming our earlier expectation of down 10% MoM.
Overall the NB segment has outperformed our expectation
during December, as the shortage constraints from Intel has
eased off in December, thus making up some upside for
December.
Based on our 4Q estimate, 4Q top six NB shipment should
grow over 21% QoQ and 40% YoY, in line with our previous
expectation of 20-23% QoQ. The Intel 865 chipset tightness
also resulted in aggressive ordering so far, and we think the
sell-through strength may kick in during the latter part of the
year.
For 1Q05, the visibily generally is poor so far especially
March outlook and so far, Christmas sales from the consumer
channel look soft but commercial/SMB segment look quite on
track. We think that combined top six NB shipments should
decline by 13%, in line with a normal seasonal pattern from
previous years. The risk we see for a huge inventory
correction for the NB segment is less severe compared to
1Q04, as ODM producers has lowered their expectations for
1Q05 following strong 4Q prospects. Although Intel will
likely launch the Sonoma platform with the Intel Alviso
chipset around January 20, ODM producers think this is
unlikely to trigger demand substantially in 1Q05 given the
initial price premium. The Sonoma major ramping may take
1-2 quarters to reach mainstream pricing points.
Quanta: December NB shipments dropped by 3% MoM and
were up by 36.7% YoY to 1.17 million units, a bit lower than
our previous expectation. This is mainly due to the shortage
of the Intel chipset. Based on our 4Q estimate, we believe
that 4Q shipments grew over 30% QoQ, in line with company
guidance. We expect January shipments to drop slightly, by
6% MoM, or shipment of 1.1 million, due to the regular
seasonality factor. The company guided that it would not be a
surprise if the 1Q05 shipment dropped by a magnitude of 10%
due to regular seasonality.
Compal: Compal shipped 760,000 NB units in December,
down 12% MoM, but up 7% YoY; the number was well
above our expectation of 728,000 NB units. This is because
the seasonality was more amplified than our expectation
previously as a result of less severe component shortage. The
strength is mainly supported by Dell, as we notice the
company barebone mix was up to 45% in December up from
40% in November. We expect January shipement to be down
by 5%, or a level of 720,000 units. Compal’s 4Q05 shipment
grew by 21% QoQ and this is a bit better than our expectation.
Overall, we expect 1Q05 NB shipments to drop by 15% QoQ
due to normal seasonality. Compal normally has higher
exposure to the consumer segment, so we believe the 1Q
correction shall be higher than at Inventec or Quanta.
Inventec: December NB shipments will likely come in at
300,000 units, flat from last month, but up 100% YoY; the
number was in line with our expectation. For January, the NB
shipment should land around 300,000 units, flat from
December. The company’s seasonal pattern is less amplified
compared to its peers, because its commercial NB percentage
in the product mix is higher than peers. For 4Q, the company
shipments should drop by 1% QoQ, again less amplified than
its peers. For January, we expect shipment to come in at
300K, flat from December, and 87.5% YoY. For 1Q05, the
company’s shipment should decline slightly, by 1%, as
Inventec clients aggressively prepare Alviso chipset platform
models.
Asustek: According to our company check, Asustek’s
December shipment is likely to be at 297,000 units, down by
10% MoM, but up 48.5% YoY, in line with the regular
seasonality factor. We expect 4Q shipments at AsusTek to
grow 27% QoQ. January shipments should drop by 10%
MoM. Overall, Asustek’s 1Q05 shipments should decline by
12%.
Wistron: Wistron shipped 310,000 units NB in December,
down 11% MoM, but up 55% YoY, in line with our previous
expectation. Wistron’s 4Q04 shipments grew 7% QoQ, less
than Taiwan Top 6 makers’ total growth of 21% MoM. For
January, we expect shipments to drop to 280K, or down -10%
MoM. Wistron’s 1Q05 shipments are likely to drop by 13%
QoQ as a result of the normal seasonality pattern.
Arima: December NB shipments should be 80K, down 20%
MoM and up 54% YoY, a bit less than our previous
expectation. 4Q05 shipment should come in around 260K,
37% QoQ, highly above peers’ growth rates due to the new
orders. We expect January shipments to stay flat with
December’s 80K. We expect 1Q05 NB shipments to drop by
2%, in line with the regular seasonal pattern.
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 33
Exhibit 14
Quarterly Shipments Growth by Taiwan's Top-Six Notebook PC Makers, 1Q2002–1Q (E) 2005
QoQ Quanta Wistron Inventec Compal Arima Asustek Top 6
1Q02 -12% 8% 22% 5% 57% NA 12%
2Q02 -2% 36% 4% 6% -7% -33% 2%
3Q02 21% -13% -12% 0% 11% 13% 5%
4Q02 43% 17% 27% 15% 23% 39% 28%
1Q03 4% -17% 5% -14% -51% 7% -9%
2Q03 10% -8% 9% 10% 20% -21% 7%
3Q03 23% 17% 3% 37% -48% 173% 26%
4Q03 3% 20% 24% 32% 4% 8% 14%
1Q04 -20% -3% 17% -23% -5% -14% -16%
2Q04 6% 18% 18% 10% 7% 30% 12%
3Q04 24% 37% 40% 20% 6% 0% 23%
4Q04 30% 7% -1% 21% 37% 27% 21%
1Q05E -14% -15% -1% -15% -8% -12% -13%
e = Morgan Stanley Research estimates
Source: Company Data, Morgan Stanley Research
Exhibit 15
Monthly Shipments by Taiwan's Top-Six Notebook PC Makers, January 2002-January (E) 2005 (000 units)
(k units) Quanta MoM Wistron MoM Inventec MoM Compal MoM Arima MoM Asustek MoM Top 5/6 MoM YoY
02-01 385 -9 120 -8 85 6 345 15 120 71 85 1,140 13 119
02-02 380 -1 135 13 85 0 280 -19 130 8 80 -6 1,090 -4 67
02-03 350 -8 165 22 110 29 320 14 190 46 83 4 1,218 12 39
02-04 360 3 170 3 85 -23 400 25 130 -32 50 -40 1,195 -2 52
02-05 360 0 220 29 95 12 300 -25 120 -8 60 20 1,155 -3 36
02-06 370 3 182 -17 110 16 300 0 160 33 57 -5 1,179 2 33
02-07 390 5 150 -18 80 -27 330 10 190 19 60 5 1,200 2 53
02-08 430 10 165 10 70 -13 345 5 130 -32 60 0 1,200 0 57
02-09 500 16 180 9 105 50 330 -4 135 4 69 15 1,319 10 61
02-10 600 20 200 11 90 -14 400 21 210 56 80 16 1,580 20 59
02-11 650 8 209 4 125 39 400 0 210 0 95 19 1,689 7 58
02-12 635 -2 170 -19 110 -12 360 -10 140 -33 87 -8 1,502 -11 49
03-01 620 -2 160 -6 110 0 350 -3 80 -43 110 26 1,430 -5 25
03-02 640 3 165 3 100 -9 300 -14 95 19 90 -18 1,390 -3 28
03-03 700 9 155 -6 130 30 350 17 100 5 80 -11 1,515 9 24
03-04 700 0 150 -3 135 4 400 14 110 10 70 -13 1,565 3 31
03-05 675 -4 145 -3 120 -11 350 -13 120 9 75 7 1,485 -5 29
03-06 790 17 145 0 115 -4 345 -1 100 -17 75 0 1,570 6 33
03-07 890 13 150 3 120 4 480 39 70 -30 180 140 1,890 20 58
03-08 810 -9 145 -3 120 0 480 0 30 -57 200 11 1,785 -6 49
03-09 960 19 220 52 140 17 540 13 70 133 220 10 2,150 20 63
03-10 940 -2 220 0 160 14 600 11 60 -14 240 9 2,220 3 41
03-11 950 1 200 -9 160 0 670 12 65 8 210 -13 2,255 2 34
03-12 855 -10 200 0 150 -6 710 6 52 -20 200 -5 2,167 -4 44
04-01 725 -15 170 -15 160 7 550 -23 52 0 180 -10 1,837 -15 28
04-02 730 1 200 18 180 13 460 -16 57 10 180 0 1,807 -2 30
04-03 730 0 230 15 210 17 520 13 60 5 200 11 1,950 8 29
04-04 810 11 250 9 190 -10 530 2 60 0 250 25 2,090 7 34
04-05 716 -12 200 -20 190 0 510 -4 60 0 250 0 1,926 -8 30
04-06 795 11 260 30 270 42 640 25 60 0 230 -8 2,255 17 44
04-07 880 11 270 4 300 11 720 13 80 33 200 -13 2,450 9 30
04-08 900 2 330 22 300 0 630 -13 70 -13 230 15 2,460 0 38
04-09 1,100 22 370 12 310 3 660 5 40 -43 300 30 2,780 13 29
04-10 1,375 25 380 3 300 -3 800 21 80 100 300 0 3,235 16 46
04-11 1,200 -13 350 -8 300 0 865 8 100 25 330 10 3,145 -3 39
04-12 1,169 -3 310 -11 300 0 760 -12 80 -20 297 -10 2,916 -7 35
05-01E 1,100 -6 280 -10 300 0 720 -5 80 0 267 -10 2,747 -6 50
e = Morgan Stanley Research estimates/ Source: Company Data, Morgan Stanley Research
Global Semiconductors – January 12, 2005
Please see analyst certification and other important disclosures starting on page 42.
Page 34
Electronics Supply Chain
Bernie Mahon (Bernie.Mahon@morganstanley.com)
Key Points
Initiated coverage of Computer Distributors: Ingram
Micro and Tech Data with Underweight ratings, and
CDW with an Equal-weight rating.
Monthly US printed circuit fabrication book-to-bill
fell to 0.96 in November from 1.05 in October
Industry view: In-Line. For the EMS companies,
consensus 2005 revenue estimates appear aggressive in
light of cyclical pressures. For the computer
distributors valuations appear stretched, particularly at
the two-tier distributors.
Secular Headwinds Mute Growth
We have initiated coverage of three leading IT distribution
stocks. We rate shares of direct marketer CDW Equalweight,
and shares of two-tier distributors Ingram Micro
and Tech Data Underweight. The three stocks have risen
sharply over the last few months, and we expect a pullback
given relatively lackluster growth prospects and valuations
near the high end of historical ranges, particularly at the
two-tier distributors.
We expect revenue growth at the two-tier distributors (IM
and TECD) to lag 1−2% behind growth of IT spending at
small and medium businesses (SMB), their primary market,
due to structural headwinds. Overall, we expect headwinds
to slow long-term revenue growth at Ingram Micro and
Tech Data to just 4% annually. We expect CDW to
continue to gain market share and to increase revenue by
roughly 10−12% annually.
Large, stable, addressable target market
We view the small and medium business (SMB) market as
the sweet spot for the one and two-tier distributors,
accounting for roughly 80% of CDW’s business — all
North America (NA) — and 50−70% of Tech Data’s and
Ingram Micro’s end customer demand, we estimate. The
SMB market represents approximately 40─45% of total IT
spending in both NA and Western Europe, or around $222
billion. There is still room for the distributors to grow
within this market, although we believe two-tier distributors
are losing share to Dell and the direct marketers.
Additionally, we view the SMB end market as less cyclical
than other segments, with a solid, diverse base of customers,
as evidenced in the 2001─2003 downturn.
IPC US Book-to-Bill Ratio Down in November
The IPC industry association released its November bookto-
bill statistics for the US printed circuit board (PCB)
fabrication industry. The three-month average ratio was
0.96, down from 1.05 in October.
The IPC estimates that it captures around 60% of the U.S.
market for rigid and flexible printed circuit boards. We
estimate that the US accounts for about 20−25% of the
world’s total PCB fabrication market, although this is likely
trending lower as production is being shifted to lower cost
locations such as Asia, including China.
Exhibit 16
US PCB
3-Month Book-to-Bill Ratio 1985−2004
0.80
0.85
0.90
0.95
1.00
1.05
1.10
1.15
1.20
12/85
6/86
12/86
6/87
12/87
6/88
12/88
6/89
12/89
6/90
12/90
6/91
12/91
6/92
12/92
6/93
12/93
6/94
12/94
6/95
12/95
6/96
12/96
6/97
12/97
6/98
12/98
6/99
12/99
6/00
12/00
6/01
12/01