GS US Semi Equipment Q4 Preview -
Companies likely too optimistic on cycle
We expect CQ4 orders to be down 10% q-o-q and for guidance to call for another 10%
q-o-q decline in orders in 1Q. We believe Street expectations call for an approximately
5% q-o-q decline in 1Q orders, with some outliers calling for flat or slightly higher q-o-q
order growth. We believe managements will likely be too optimistic regarding the cycle
by predicting a bottoming in fundamentals in 1H'05 on upcoming earnings calls. This is
consistent with history, as managements have tended to "jump the gun" in anticipating a
trough in fundamentals. We believe continued excess capacity absorption will drive
deteriorating fundamentals for another several quarters with a fall-off in memory orders
likely to drive the next meaningful leg down in fundamentals. We believe that the
significant excess capacity currently being added to the memory segment is likely to drive
declining DRAM ASPs and, in turn, a significant fall-off in memory orders beginning in
2Q'05. We continue to recommend that investors maintain underweight positions in the
group.
Please contact us for a full copy of our fourth-quarter preview and industry update report.
Companies likely too optimistic:
Although we expect calendar fourth-quarter orders to be down approximately 10%
sequentially and for guidance to call for another 10% sequential decline in orders in the
calendar first quarter, we believe that the Semi Equipment companies on upcoming
earnings calls are likely to prematurely call a bottoming in fundamentals in the first half
of 2005. This is consistent with history, as managements have tended to be too
anticipatory of a cycle bottom.
We believe investors should remain underweight the stocks, as we expect continued
fundamental deterioration over the next several quarters, with a decline in memory orders
likely to drive the next meaningful leg down in fundamentals. Our view that the current
downturn will persist into 2005 is in contrast with the Street view that the downturn will
be short and shallow with a bottom likely in first-half 2005.
4Q ORDERS LIKELY DOWN 10% Q-O-Q, WITH 1Q GUIDANCE LIKELY
ANOTHER 10% DECLINE. We expect the Semi Equipment companies to generally
report revenue and earnings for the December quarter that are in line with expectations,
with revenues on average for the large front-end companies likely to decline
approximately 15% sequentially. We expect orders to be down approximately 10%
sequentially in the December quarter, as the industry begins to correct for the excess
capacity that was added during the upturn. We believe that the large orders making up
Semi Equipment order books in the December quarter were likely driven by Intel, Hynix,
and Toshiba, among a few others. We expect the Semi Equipment companies to guide
orders down another 10% sequentially in 1Q2005, driven by continued capacity
absorption across customers.
MANAGEMENTS LIKELY TO PREMATURELY CALL A BOTTOM IN
FUNDAMENTALS IN 1H2005. We expect the Semi Equipment companies to "jump the gun" in
calling a bottoming in the cycle in the first half of 2005 on upcoming earnings calls. To that end, a
look back at previous downturns highlights that Semi Equipment managements have historically
tended to prematurely call a bottom in fundamentals.
At the beginning of the last downturn, Applied Materials, for example, guided for a 10% sequential
decline in calendar 4Q2000 orders on its 3Q2000 earnings call, but actual fourth-quarter 2000
orders ended up coming in down 33% sequentially. Management chose not to issue order guidance
for calendar 1Q2001 on its 4Q2000 earnings call. In 1Q2001, orders came in down 44%
sequentially and the company indicated that it believed orders had bottomed and guided for flat
2Q2001 orders. Second-quarter 2001 orders ended up coming in down 10% sequentially and
management guided for a modest recovery in orders in 3Q2001. Orders ended up coming in down
another 9% sequentially in 3Q2001. Following the fundamental head fake in 2002, orders once
again fell short of expectations in 2Q2002, up 5% sequentially versus guidance for a 10% to 15%
sequential increase. Orders declined 12% sequentially in 3Q2002 versus guidance for a decline of
5% to 15% sequentially. Orders then declined another 35% sequentially in 4Q2002 versus
guidance for a 20% sequential decline before eventually bottoming in the 1Q2003 time frame.
A similar pattern can be seen by examining Lam Research's orders and guidance during the last
downturn. During the beginning of the 2001-2002 downturn, the company guided its 1Q2001
orders down 35% sequentially. Actual 1Q2001 orders came in down 60% sequentially and the
company then guided for a flat to slight increase in orders sequentially (+5% to 10%) in 2Q2001.
Actual 2Q2001 orders ended up coming in down 14% sequentially, with guidance calling for flat to
down 10% sequential order growth in 3Q2001. Actual 3Q2001 orders came in down 25%
sequentially and guidance called for orders to be flat to up slightly in 4Q2001. Actual 4Q2001
orders ended up coming in down 15% sequentially.
Finally, the same trend can be seen by looking at Novellus' guidance as compared with actual order
results during the last downturn. Novellus issued guidance calling for a 30% to 35% sequential
decline in 1Q2001 orders. In 1Q2001, actual orders ended up coming in down 58% sequentially
and management guided for a 5% sequential increase in orders in 2Q2001. In 2Q2001, orders came
in slightly light of guidance but were essentially flat sequentially and guidance called for another
10% sequential decline in orders in 3Q2001. Third- quarter 2001 actual orders came in down 41%
sequentially and guidance called for flat to up about 20% sequential order growth in 4Q2001. The
company lowered its order guidance to down 15% sequentially to flat on its mid-quarter update
call. Actual 4Q2001 orders ended up coming in down 11% sequentially. Once the head-fake upturn
of 2002 ended, guidance called for a sequential decline of 9% in 3Q2002 orders. Actual orders
came in lower than guidance in 3Q2002, down another approximately 25% sequentially, before
eventually improving in 4Q2002.
We believe that managements will be equally anticipatory of calling a bottom in the cycle this
downturn as they have been in previous downturns. We believe that the trend of prematurely
calling a cycle bottom has already begun. Novellus management, for example, highlighted on its
mid-quarter update call in late November that business was likely to stay flat around fourth-quarter
levels until mid-2005 before another leg of growth would ensue. KLA-Tencor also offered a
similar view at a competitor conference in early November. If/when Semi Equipment managements
guide orders down 10% sequentially for the March- quarter, we would expect commentary similar
to the last downturn, suggesting that the first half of 2005 could mark the fundamental bottom.
We continue to believe the above history underscores that investors should monitor the industry
supply/demand equation rather than follow company commentary to forecast cycle bottoms. As we
highlight below, we believe supply/demand dynamics remain out of balance, which will likely lead
to another several quarters of correction.
INDUSTRY SUPPLY/DEMAND DYNAMICS REMAIN OUT OF BALANCE. We continue to
believe that the current downturn is being driven by excess supply, as the industry shipped excess
manufacturing capacity during the upturn that needs to be absorbed during the downturn before
fundamentals can improve again. Below we review our analyses that lead us to believe that the
current downturn is supply- driven.
WORLDWIDE FRONT-END SHIPMENTS REMAIN ABOVE THEIR NORMALIZED
LEVELS. Our analysis of worldwide front-end Semi Equipment shipments continues to show that
shipments remain above their normalized levels. We plot three-month rolling average worldwide
front-end Semi Equipment shipments (as reported by Semiconductor Equipment and Materials
International, SEMI) on a log scale and then fit a trendline to the data. While shipments have
declined from third- quarter levels, they remain above their normalized trendline and therefore need
to decline to below the trendline (and remain there for a couple of quarters) in order to correct for
the excess capacity (or the area above the curve) that was added during the recent upturn.
Further, while shipment levels have been trending downward, the capacity that has been shipped in
2004 will continue to come online and yield at the chipmakers well into 2005. This means that
even as near-term shipment levels continue to decline, there is still capacity that is continuing to
come online as previous shipments begin to yield.
SUPPLY/DEMAND MODEL ALSO HIGHLIGHTS EXCESS SUPPLY IN 2005. Another method
that helps us gauge the amount of capacity that has been added to the industry is our
supply/demand model, which shows that if semiconductor revenues are flattish in 2005 and capex
is down 7.1% year
over year in 2005 (which is what our bottom-up capex model currently predicts although we expect
capex cuts to come throughout the year to drive the number to down 10%-15% year over year),
there is still excess capacity coming online that will cause capacity utilization rates to decline to the
low- to mid-80% level in 2005 from 93% in 2004. Please contact us if you would like a copy of our
detailed supply-demand model.
SPREAD BETWEEN SEMI REVENUE AND CAPEX GROWTH POINTS TO EXCESS
SUPPLY. A final measure of excess capacity is shown in a chart (please contact us for the chart),
which plots the spread between year-over-year semiconductor revenue growth and capital spending
growth against the yearly percentage change in semiconductor average selling prices. We use the
spread between revenue growth and capex growth as a proxy for the amount of excess capacity that
is added to the industry during an upturn. The higher the negative spread (i.e. the more the
semiconductor industry is spending on capex relative to revenue growth), the more excess capacity
the industry is adding. Our bottom-up capex model indicates 58% year-over-year capex growth in
2004 and the Goldman Sachs semiconductor team is estimating about 27% year-over-year semi
revenue growth in 2004, implying a spread between semiconductor revenue growth and capex
growth of approximately negative 30%.
Only once before (in 2000) did the semiconductor industry see such a high spread between revenue
growth and capex growth. This spread has been a strong leading indicator of how much excess
capacity is being added to the industry in the following year. One effect of a large negative spread
(and hence excess capacity) is that semiconductor ASPs tend to fall the following year. Note that
after a negative 33% spread in 2000, semiconductor ASPs declined 14% in 2001. We therefore
believe semiconductor ASPs are unlikely to find significant support in 2005, which is important
because lower ASPs drive lower cash flows with which to fund capex.
IC UNITS LIKELY TO DECLINE FOR ANOTHER FEW QUARTERS. Finally, we consider
closely what semiconductor demand will look like in 2005. As part of our analysis, we look at
quarterly IC units against trendline IC unit growth from 1993 through 2004 using a 5% sequential
decline in IC units for our December- quarter estimate. The analysis shows that for five consecutive
quarters IC units have been above their trendline unit growth, which equates to chip inventories
being built in the supply chain. IC units have begun to decline back toward the trendline, but must
remain below the trendline for a few quarters (i.e., filling in the area below the curve) until the
industry absorbs the excess inventories. We would therefore expect another couple of quarters of
declines in IC units. The Semi Equipment stocks have not historically bottomed until IC units have
reached a bottom. Given our belief that IC units are likely to continue to decline, we do not expect
the stocks to find a trough in the near term.
THE NEXT LEG DOWN IN FUNDAMENTALS WILL LIKELY BE DRIVEN BY MEMORY
CUSTOMERS. We expect fundamentals to be weak over the next couple of quarters, with memory
likely to drive the next meaningful leg down in fundamentals. We believe the continued
overinvestment in DRAM will drive declining DRAM ASPs, which, in turn, is likely to drive
significant declines in memory orders beginning in 2Q2005.
MEMORY ORDERS MAKE UP A SIGNIFICANT PERCENTAGE OF EQUIPMENT ORDER
BOOKS. Memory orders are expected to represent a significant percentage of Semi Equipment
order books in the fourth quarter of 2004. Memory capital spending (mostly DRAM spending)
increased 23% sequentially in the third quarter and is expected to increase another 27%
sequentially in the fourth quarter, which we believe underscores that memory will again be a
significant driver of orders.
Further, several equipment companies, including KLA-Tencor and Tokyo Electron, have
commented that memory orders are expected to represent approximately 50% of order books in the
fourth quarter. Our checks with equipment vendors lead us to believe that memory spending is
expected to continue to be strong into the first quarter of 2005. We believe that the high percentage
of memory orders is a reflection of customers building excess capacity that will need to be
absorbed, which we expect to begin in 2Q2005. DRAM ASPS LIKELY TO DECLINE IN 2005.
While one need not believe that DRAM ASPs must decline in order for DRAM makers to slow
their exuberant investment in new capacity, as digestion of new capacity will also likely slow
orders, we expect a decline in DRAM ASPs in 2005 to exacerbate a fall- off in memory orders.
Spot prices were a bit stronger than expected in the second half of December, which our team in
Asia believes has been driven by a mismatch between supply and demand for DDR1 and DDR2
chips. Goldman Sachs believes that DRAM makers overestimated the demand for DDR2 chips and
hence allocated too much capacity to DDR2, leaving tighter-than-expected supply of more heavily
demanded DDR1 chips. Our Taiwan Tech team believes that DDR2 penetration in the first quarter
of 2005 should be at best 20% to 30%, according to PC and notebook makers. This compares with
DRAM makers' previous expectation of 30% to 40% penetration of DDR2.
The DDR2 transition appears to be a bit rockier than anticipated, which explains the
lower-than-expected demand. Yields on DDR2 are still approximately 5% lower than yields on
DDR1 and testing/packaging costs are still significantly higher than on DDR1. The higher DDR2
costs are in turn leading to a higher price premium over DDR1, which explains why customers
continue to be reluctant to make the transition to DDR2.
Our team in Asia believes that DRAM makers will likely adjust their capacity allocation according
to market demand back toward DDR1, which should alleviate the current tight supply situation
within two and a half months. They further believe that the recent relative strength in spot prices
has been driven partly by inventory stocking heading into the Chinese Lunar New Year, which
should be finished before the capacity re- allocation to DDR1 we highlighted above is completed.
We believe that the confluence of the significant capacity that is coming online from memory
orders placed in the third and fourth quarters of 2004, the likely capacity re-allocation toward
DDR1, and the anticipated end to inventory stocking prior to the Chinese Lunar New Year, should
all lead to a significant fall-off in DRAM prices around the March time frame. Those second- and
third- tier DRAM makers that rely on strong ASPs to fund capital spending will likely be forced to
cut back on their capital investments if ASPs decline as we expect.
FOUNDRY ORDERS ARE UNLIKELY TO OFFSET WEAKNESS IN MEMORY ORDERS. We
believe that foundry orders are unlikely to offset the expected fall-off in memory orders in the
first-half of 2005. While the foundries ordered at significantly lower levels in the third quarter and
are expected to be at low levels again in the fourth quarter, we believe that continued low
utilization rates are likely to prevent them from making significant capacity expansions in the first
half of 2005.
We believe that utilization rates at the large Taiwanese foundries will be in the low- to mid-70%
range in the first quarter of 2005. With capacity that was added in early to mid-2004 beginning to
yield in early 2005, coupled with weak unit demand, we believe that it will take several quarters of
unit improvement at the foundries to drive utilization rates to high-enough levels that the foundries
would feel compelled to begin placing significant rounds of orders again.
THE SEMI EQUIPMENT STOCKS ARE TRADING AT RICH VALUATIONS. While most of
the large Semi Equipment stocks lagged the broader market during the fourth- quarter 2004 ral
ly (AMAT +4%, NVLS +5%, KLAC+12%, versus the Nasdaq Composite Index (COMP) +15%),
they still enjoyed an upward move in late 2004, driving valuations even further above the fair
values suggested by our normalized earnings and normalized free cash flow valuation analyses.
THE STOCKS ARE PRICING IN VERY HIGH LONG-TERM FREE CASH FLOW GROWTH
RATES. First, our normalized free cash flow-based valuation methodology allows investors to
understand what free cash flow growth rate the market is discounting in the current stock prices. As
we highlight below, the market is currently pricing in extremely aggressive free cash flow growth
rates, especially considering that the growth rate utilized in this analysis is in perpetuity. In other
words, if the Semi Equipment companies were actually able to grow their free cash flow at these
rates (i.e., 10%) in perpetuity, they would eventually take over the world economy. We believe that
this analysis highlights that stock prices need to come down in order to better reflect the "true" fair
values of the stocks.
Ticker Implied FCF Growth rate AMAT 9.6% KLAC 10.0% LRCX 10.6% NVLS 9.9% TER
13.1% Source: Goldman Sachs Research estimates.
STOCKS ARE TRADING AT EXTREMELY HIGH MULTIPLES ON NORMALIZED
EARNINGS. Second, the stocks in our coverage universe are trading at very high multiples on
normalized earnings. Applied Materials, for example, is trading at 32X our estimate of current full
cycle (2003-2006) normalized earnings. This is approximately 2X the market multiple for a
company in an industry that has grown at a 1% CAGR over the last two cycles and that has
experienced approximately 500 basis points of margin erosion from the 2000 cycle peak to the
recent cycle peak in 2004. Similarly, Lam Research is trading at 26X our estimate of normalized
earnings, Novellus Systems is trading at 34X our estimate of normalized earnings, and
KLA-Tencor is trading at 31X our estimate of normalized earnings. We believe this analysis
highlights that the stocks are likely to move lower if/when the Street realizes that the downturn is
not over.
WE RECOMMEND UNDERWEIGHT STOCK POSITIONS IN THE GROUP. We continue to
recommend that investors maintain underweight stock positions in the Semi Equipment sector. Our
belief that fundamentals will continue to deteriorate for another several quarters is in contrast to the
Street view that the current downturn is likely to bottom in the first half of 2005. We expect the
stocks to move lower as the market is likely to be disappointed by the worse-than-expected
fundamental environment over the coming quarters.
WE RECOMMEND BUILDING INTERMEDIATE- AND LONG-TERM POSITIONS IN FORM
ON WEAKNESS. Despite our general bearish view on the sector, we continue to believe that
FormFactor represents a core long-term holding in Semi Equipment and the stock remains our top
intermediate- and longer-term pick in the group. That said, the stock has been weak near-term
driven by the company's 4Q2004 revenue miss (FormFactor negatively preannounced its 4Q2004
results in early January) and the choppy demand environment that we expect will persist in
1H2005. We would use near-term weakness to build intermediate- and long-term positions in
FormFactor, as we expect the stock to outperform in full-year 2005 and beyond.
We first indicated that fundamentals at FormFactor were likely to be weaker than expected in
1H2005 in our weekly note out on December 19, 2004. Before then we had expected the company
to be able to report flat revenues in each quarter of 2005 despite a tough overall market
environment. While we no longer expect the company to avoid sequential revenue declines in
2005, we believe the magnitude of the declines FormFactor is likely to report will be less severe
than those that we expect other Semi Equipment companies to report.
Up until the company's negative preannouncement in early January, we would also have classified
management's execution as flawless. However, the 4Q2004 revenue miss was driven primarily by
contamination in the company's production line, which drove lower yields and in turn lower
shipments to customers. We would therefore no longer classify execution as flawless although we
still believe that the management is solid and capable of driving revenue growth.
We continue to expect FormFactor to generate significant revenue growth of about 15% to 20%
annually over the next five years versus our expectation that the broader Semi Equipment industry
will continue to grow only in line with the broader economy. We expect strong growth to be driven
by increased penetration in existing markets as well as by new applications and technology
transitions within those markets. Over the longer term, we believe that FormFactor has the
opportunity to penetrate new markets, namely parts of the front-end manufacturing process
including metrology and design verification.
Regarding valuation, we use P/E to growth (PEG) as our valuation methodology for FormFactor,
given our view that the company has significant potential growth opportunities. We estimate that
FormFactor can generate approximately $0.70 in earnings per share in 2005 and that the company
can grow at 15%-20% over the next five years. The stock is therefore trading at a PEG of 2.2,
assuming a growth rate of 15%.
We assume that the Semi Equipment industry is growing at 5% to estimate PEG ratios for the
major front-end Semi Equipment companies: Applied Materials, KLA- Tencor, Lam Research, and
Novellus. We also use current full-cycle (2003-2006) normalized EPS estimates for the large
front-end names to smooth out the volatility inherent in their cyclical earnings. As the table below
shows, the major front-end stocks are trading at significantly higher PEG multiples than
FormFactor. We therefore believe that FormFactor is reasonably valued relative to its peers when
its long-term growth potential is taken into account.
PEG FORM 2.2* AMAT 6.4 KLAC 6.0 LRCX 5.1 NVLS 6.9 *FORM's EPS is based on our
published 2005 estimate. All other EPS estimates are based on our estimate of current full-cycle
(2003-2006) normalized earnings. Source: Goldman Sachs Research estimates.
INDIVIDUAL EARNINGS EXPECTATIONS:
TERADYNE (TER; U/C) earnings expectations. Teradyne is reporting December- quarter earnings
on January 18th in the evening with a conference call held the following morning at 10:00am
Eastern. We forecast fourth-quarter revenues of $350 million (down 24% sequentially) with
breakeven earnings per share, versus the Street consensus earnings per share estimate of $0.02.
While Teradyne doesn't provide explicit order guidance, we forecast December-quarter gross
orders of $285 million, down 3% sequentially, which we would note includes services bookings
that we estimate will be approximately $30 million. We forecast March-quarter bookings of
approximately $260 million, down 9% sequentially.
We expect business conditions at Teradyne to remain weak in the near term, driven by weakness in
the Semi Test and TCS segments. While we expect the Semi Test segment to remain under
pressure in Q4, we believe that the segment's performance should not differ significantly from
previous expectations. On the other hand, we believe that the TCS business was weaker than
expected during the quarter, primarily due to weakness at some of the company's largest TCS
customers. We would note that the TCS segment has a less variable cost structure relative the
company's other segments, which we believe is likely to drive worse than expected profitability.
In early December we met with Teradyne management during our Boston Bus Tour. Management
indicated that the pricing environment in the back end remains difficult and we believe that the
SOC test companies continue to be focused on gaining market share at the expense of margins at
certain key cus
tomers. To mitigate some of the impact of continued pricing pressure, Teradyne is currently
working on lowering its quarterly revenue breakeven run rate to approximately $330 million from
the current $355 million level. The lower breakeven level is not expected to be achieved until
mid-to-late 2005.
Given the company's current quarterly breakeven level, 3Q net orders of $285 million (and as we
highlighted earlier, we expect no better than flat Q4 orders despite the fact that Q4 bookings are
likely to be boosted by service bookings), and the company's significantly smaller backlog exiting
2004 as compared to the early stages of the previous downturn, we find it hard to envision how
Teradyne won't lose money in the current cycle. Street expectations have Teradyne losing only
$0.01 in FQ1 2005 and remaining profitable for the rest of the year with a full year 2005 EPS
estimate of $0.23 vs. our LPS estimate of -$0.55. We believe that the stock is likely to be
negatively impacted by what we expect will be significantly worse than Street consensus
profitability.
LAM RESEARCH (LRCX; IL/C) earnings expectations. Lam Research will report
December-quarter (second fiscal quarter 2005) earnings on January 20th after the market close. We
model revenues of $390 million (down 6% sequentially) and earnings per share of $0.55, versus the
Street consensus earnings per share estimate of $0.53. We forecast December-quarter orders of
$390 million (down 9% sequentially), approximately at the mid-point of management's guidance
for a 5% to 15% sequential order decline. We expect the company to guide for an approximate
10% decline in March-quarter bookings to $350 million.
We would expect incremental weakness at Lam to continue into 2005 as we expect business to be
negatively impacted by what we believe will be a fall off in memory orders. Recall that 57% of
Lam's orders in CQ3 were driven by memory customers and memory is expected to continue to
represent a significant percentage of the order book in CQ4, which we believe leaves the company
vulnerable to a meaningful decline in orders from the memory makers in 2005.
KLA-TENCOR (KLAC; OP/C) earnings expectations. KLA is reporting December- quarter
(second fiscal quarter 2005) earnings on January 20th after the market close. We model revenues of
$520 million (flat sequentially) with earnings per share of $0.59, in-line with the Street consensus
estimate. We model calendar 4Q2004 gross orders of $480 million (-10% sequentially), at the
mid-point of company's guidance for orders of flat to down 20% sequentially. We expect
management to guide for an approximate 10% sequential decline in orders in the March-quarter, or
about $440 million.
While we expect KLA's revenues to hold up better than its peers during the downturn, as they are
driven more by technology transitions than capacity additions, we expect the company to be
negatively impacted by a slowdown in the DRAM business beginning in Q2 2005. During its third
calendar quarter earnings call, management indicated that memory will comprise approximately
50% of 4CQ orders, which we believe leaves the company exposed to incremental downside risk
if, as we expect, memory orders begin to decline.
On the upcoming earnings call, we expect to get greater detail on the company's Blue29 initiative,
which is a wholly-owned joint-venture between KLA and Dainippon Screen. Blue29 is a Silicon
Valley-based start-up that has developed an electroless film deposition technology used for copper
interconnects. We would expect management to be asked about its intention to enter the deposition
business on the earnings call.
NOVELLUS SYSTEMS (NVLS; IL/C) earnings expectations. Novellus will report
September-quarter earnings on January 27th after the market close. We model 4Q2004 revenues of
$330 million (down 21% sequentially) and earnings per share of $0.26, versus the Street consensus
earnings per share estimate of $0.23. We also project 4Q net orders of $325 million, down 23%
sequentially. We expect the company to guide for 1Q net orders of approximately $300 million,
down 8% sequentially.
As we highlighted earlier in this report, Novellus indicated on its mid- quarter update conference
call held in late November that it believed business would stay flattish from fourth-quarter levels
through the first- half of 2005. However, we believe that management will guide March orders
down about 10% sequentially given the lack of visibility on on-DRAM orders in Q1 2005.
ATMI (ATMI; IL/C) earnings expectations. ATMI is reporting December-quarter earnings on
February 2nd. We model fourth-quarter revenues of $62 million (down 4% sequentially) with
earnings per share of $0.18, versus the Street consensus estimate of $0.17. We expect ATMI's
business to be negatively impacted by weakness in wafer starts in the December quarter.
Management indicated on its last quarterly earnings call that worldwide wafer starts are expected to
decline approximately 7% to 10% sequentially in the fourth-quarter. Our recent checks with
wafer-start driven companies suggest that global foundry wafer starts are likely to be down 18%
sequentially in the fourth-quarter. Our checks also indicated that all wafer starts (including DRAM
and foundry) in Taiwan were down 8.3% m-o-m in October. This was followed by another 3.1%
decline m-o-m in November and wafer starts are projected to have declined yet again in December.
In total, Taiwan wafer starts are expected to be down 11% q-o-q.
While we believe that ATMI may at some point during the downturn represent an interesting
defensive play for investors who must maintain some exposure to the group, we would wait for
more wafer start reductions among the device makers outside of Taiwan before becoming more
aggressive on the name.
AXCELIS TECHNOLOGIES (ACLS; IL/C) earnings expectations. Axcelis is reporting
December-quarter earnings on February 2nd. We model revenues of $100 million (down 22%
sequentially) with earnings per share of $0.06, in line with the Street consensus earnings per share
estimate. We model 4Q2004 gross orders of $100 million (down 21% sequentially), excluding the
Sumitomo Eaton Nova (SEN) joint venture. While Axcelis does not provide official order
guidance, we currently expect orders to be down about 10% sequentially in 1Q2005 to about $90
million.
We would note that over 50% of Axcelis' business in the third-quarter came from memory. Given
that it is our expectation that memory orders are likely to experience significant weakness
beginning in Q2 2005, we would expect Axcelis' business to continue to experience significant
weakness in 2005.
We will be paying careful attention to developments in the high current segment. Recall that the
level of energy used during implantation impacts how deep ions go into the transistor while the
dose used during implantation impacts the ion concentration in the transistor. High current
implantation uses low levels of energy but a high implant dose and is typically the longest implant
step. High current implantation is used to build the gate electrode, source, drain, and extension (or
ultra shallow junction) in a transistor.
High current implant is one of the slower implant steps so wafer throughput is lower during this
step. Customers therefore require more high current ion implant tools in a given fab in order to
maintain a higher overall throughput level, which in addition to the increasing number of high
current steps as devices shrink is driving the ion implant companies to focus their efforts in this
segment.
Axcelis has had a batch-wafer (meaning more than one wafer is exposed at once) high current tool
available for some time and introduced a single wafer high current tool in late 2004. We believe
that the market has increasingly acknowledged the need for single-wafer high current implant as
device geometries continue to
shrink. We therefore believe that it will be important for Axcelis to gain traction with its single
wafer high current tool given that both of its competitors in the ion implant market, Varian and
Applied Materials, have solid offerings in this segment.
FORMFACTOR (FORM; OP/C) earnings expectations. FormFactor negatively preannounced its
4Q2004 results on January 6th and is reporting earnings on February 10th. In the negative
preannouncement, management indicated that 4Q 2004 revenues were approximately $46 million
vs. a previous range of $50 million to $53 million. The net impact of lower revenues and an
increase in costs associated with a faster ramp-up of the company's new manufacturing facility is
expected to have a $0.06 negative impact to the company's original 4Q2004 EPS guidance of $0.18
to $0.19. We are modeling 4Q2004 EPS of $0.11.
FormFactor also provided preliminary 4Q2004 orders of approximately $40.7 million, down 10%
sequentially. While the company had not provided order guidance during its 3Q2004 earnings call
as is its custom, we had expected 4Q2004 orders of $46.8 million, up 4% sequentially.
Management indicated on the call that it expects 1Q2005 revenues to be up approximately 2%
sequentially to $47 million, which is significantly below our original expectation for revenues of
$52 million.
The revenue shortfall during 4Q2004 was primarily attributable to contamination problems at
FormFactor's existing manufacturing facility (note that the company is currently ramping up
production at a new manufacturing facility which we would expect it to address in detail on the
upcoming earnings call). The contamination of a wafer production line caused a reduction in yields
and subsequent delays in shipments of customer orders. The company has said that it has identified
the causes of the problem and has taken action to improve yields.
That said, yields are expected to remain challenged heading into 1Q2005, which is limiting the
company's ability to generate more revenue (recall that Formfactor's revenue has a significant turns
component). The ramp-up of the company's new manufacturing facility is expected to improve
yields, so management remains focused on driving the ramp as quickly as possible.
In addition to the manufacturing execution problems, the company indicated that it saw a decline in
business momentum during the quarter driven by weakness in the flash segment while the DRAM
segment remained solid and logic was flat sequentially. This demand environment choppiness was
the driver behind the lower than expected orders in 4Q2004.
We believe that two things have changed about our bullish view on FormFactor. First, up until we
published our weekly on December 19th where we highlighted that fundamentals for the company
would likely be choppy in H1'05, we had expected FormFactor to be able to report flattish
sequential revenues in each quarter of 2005 despite a very tough overall market environment.
Clearly today that is no longer the case. Second, we would classify management's execution until
now as being flawless, which is certainly no longer the case after the yield problems at the factory
in 4Q2004.
Post the negative preannouncement, we believe there are three things about our bullish view on the
company that are unchanged. First, we continue to believe that FormFactor is the best growth story
in Semi Equipment given its ability to continue to penetrate the markets it already serves (DRAM,
flash, flip- chip/logic) while also entering new applications (i.e. wafer level burn-in and, eventually
front-end applications). We believe FormFactor has executed on its plan of moving more of the
back-end to the wafer probe stage and we believe that in the long-term the company may be able to
move more of the front-end processes to the wafer level, i.e. parts of the design verification and
metrology processes. Second, we believe that the company is likely to grow revenues at a 15% to
20% CAGR over the next five years while we expect the Semi Equipment industry to continue to
grow only in-line with the broader economy. Third, while the yield issues in 4Q2004 are a strike
against management's execution record, we still believe that FormFactor has a very strong
management team with a broad skill-set and the ability to drive continued, profitable growth.
We would use weak near-term stock performance driven by the 4Q2004 revenue miss and the
choppy fundamental environment in 1H2005 to build intermediate- and long-term positions in the
name. The stock is trading at 28x our recently reduced CY2005 EPS estimate of $0.70 vs. an
average of 31x current full cycle (2003-2006) normalized earnings for the four large front-end
stocks (AMAT, LRCX, KLAC, and NVLS), which we find to be a compelling valuation given
FormFactor's superior long-term growth prospects.
MKS INSTRUMENTS (MKSI; IL/C) earnings expectations. We model 4Q 2004 revenues of $125
million (down 11% sequentially) and earnings per share of $0.13, versus the Street consensus
earnings per share estimate of $0.11.
We believe that weakness in MKS' December-quarter will be driven primarily by deteriorating
business conditions at its semi OEM customers. Recall that MKS' revenues are directly tied to
shipments at semi OEMs (i.e. Applied, Novellus, and Lam, with Applied being MKS's largest
customer). During their earnings calls in October, Novellus and Lam both indicated that shipments
will be down significantly in the December quarter, with specific guidance calling for declines in
shipments of approximately 14% for Novellus and 16% for Lam. We expect similar declines in
shipments for other major Semi Equipment companies. Further, as we highlighted earlier in this
report, we expect shipments to continue to decline for another several quarters which is likely to
drive continued weakness at MKS.
ADVANCED ENERGY INDUSTRIES (AEIS; U/C) earnings expectations. We forecast 4Q 2004
revenues of $83 million (down 11% sequentially) with a loss per share of - $0.19 vs. the Street
consensus LPS estimate of -$0.18.
In terms of guidance, we would expect the company to call for another sequential decline in
revenues, as we expect shipments at its semi OEM customers to continue to decline in the coming
quarters. We are forecasting a 4% sequential decline in sales for the March quarter, with a LPS of -
$0.17, which we believe may have incremental downside risk.
We remain concerned about the company's balance sheet as we believe that AE needs to raise
approximately $150 million (possibly through a convertible note offering) in order to repay two
convertible notes (one in the amount of $121.5 million due in September 2006 at a conversion
price of $29.83 and the other in the amount of $66.2 million due in November 2006 at a conversion
price of $49.53), which would be extremely dilutive to a company with a market capitalization of
approximately $265 million. We expect management to provide an update on its balance sheet
situation on the call.
BROOKS AUTOMATION (BRKS; IL/C) earnings expectations. We model calendar fourthquarter
revenues of $125 million (down 24% sequentially) with earnings per share of $0.10, versus
the Street consensus earnings per share estimate of $0.09. We model fourth calendar-quarter gross
orders of $125 million (down 16% sequentially). Management does not provide official order
guidance, but we expect orders to decline about another 10% sequentially in the March-quarter to
10 Goldman Sachs Global Investment Research
January 10, 2005 Analyst Comment
$115 million.
We would expect management to be cautious on the industry outlook due to Brooks' meaningful
exposure to Semi Equipment shipments (50% of the company's business is driven by Semi
Equipment shipments), which we believe were down significantly sequentially in the fourth
calendar-quarter. As previously mentioned, Novellus and Lam have both guided shipments down
sequentially in the double digit range in the fourth calen
dar-quarter and we believe other tool vendors are experiencing similar shipment declines. We
would also expect further shipment declines into 2005, as our analysis of front-end semi
equipment shipments continues to show that shipments remain above their normalized levels,
implying that further declines are necessary before shipments can begin to improve.
APPLIED MATERIALS (AMAT; U/C) earnings expectations. Applied Materials is reporting
January-quarter (first fiscal quarter 2005) earnings on February 15th after the market close. We
model $1.75 billion in revenues (down 21% sequentially) with earnings per share of $0.16, vs. the
Street consensus estimate of $0.15. We estimate fourth calendar-quarter gross orders of $1.85
billion (down 29% sequentially), as compared with management guidance for a 35% sequential
order decline. We would note that we believe Applied recognized a large order at the end of its
October-quarter, which likely helped the company exceed CQ3 order expectations but also
exacerbated the magnitude of the CQ4 sequential order decline relative to its peers.
While we intend to provide a detailed quarterly preview for Applied after its January-quarter
closes, our preliminary thoughts are that the company's fundamentals are likely to continue to
deteriorate into 2005. We would expect the stock to be weak and to continue to underperform the
market (as we noted earlier, AMAT increased 4% in the fourth-quarter while the COMP increased
15% and the S&P increased 9%) driven by: 1) Deteriorating fundamentals: as we continue to
highlight, we expect fundamentals to be weak for another several quarters until the industry soaks
up the excess manufacturing capacity that it added during the recent upturn, and 2) Valuation:
Applied is trading at 34x our estimate of current full cycle (2003-2006) normalized earnings, which
is an approximate 2x premium to the market multiple for a stock in an industry that has grown at a
1% CAGR and for a company that has experienced approximately 500 basis points of peakto-
peak margin contraction.
CREDENCE SYSTEMS (CMOS; U/C) earnings expectations. We model January- quarter
revenues of $105 million (down 7% sequentially) with a loss per share of - $0.15, versus the Street
consensus loss per share estimate of - $0.16. We model January-quarter gross orders of $105
million, down 7% sequentially. While the company doesn't offer official order guidance, we
estimate April-quarter gross orders of $100 million, down approximately 5% sequentially.
We believe that back-end fundamentals will continue to deteriorate in 1CQ driven by excess
capacity that is still being absorbed and low utilization rates at the outsourced package and test
houses. We expect this continued deterioration in back end fundamentals to negatively impact
Credence's contracting margin profile (Credence achieved peak gross margins this cycle of 52%
vs. peak gross margins of 61% last upturn). We believe the pricing environment is having the
greatest adverse impact on margins, as the back end segment is plagued by severe price
competition with companies remaining unwilling to maintain pricing at the expense of losing
business to competitors.
We would also note that Credence generated only $0.26 in cumulative earnings over the course of
the entire recent upturn and will lose more than what it earned in the entire upturn in only the first
two quarters of the current downturn, thus generating significant negative earnings over the course
of a full cycle. As a result, we believe that the stock continues to have significant downside risk.
I, Jim Covello, he |
