GS US SEMI EQUIP WEEKLY: WILL
SECULAR CONCERNS ARISE IN THE
DOWNTURN?
Summary: (1) Short interest up m-o-m in September; stock rallies likely to occasionally
occur as short interest continues to bounce from May lows, but absolute short interest
levels are still significantly lower than the peak they reached in the previous cycle. Our
analysis shows the tight inverse correlation between SPE short interest and the SOX index
which implies downside risk for the stocks, (2) Adjusting Novellus model to reflect lower
sharecount as the company continues to buyback significant shares, (3) What might turn
investor focus from hope of a short and shallow downturn to the many secular challenges
facing the industry and what kind of impact might that have on the stocks? (4) Entegris
reporting FQ4 on Thursday; management tone likely more cautious, and (5) News, Events
and Price Performance.
SHORT INTEREST UP M-O-M IN SEPTEMBER; STOCK RALLIES LIKELY TO
OCCASIONALLY OCCUR AS SHORT INTEREST CONTINUES TO BOUNCE
FROM MAY LOWS, BUT ABSOLUTE SHORT INTEREST LEVELS ARE STILL
SIGNIFICANTLY LOWER THAN THE PEAK THEY REACHED IN THE
PREVIOUS CYCLE. OUR ANALYSIS SHOWS THE TIGHT INVERSE
CORRELATION BETWEEN SPE SHORT INTEREST AND THE SOX INDEX
WHICH IMPLIES DOWNSIDE RISK FOR THE STOCKS. Data released Friday after
the close showed that aggregate short interest for the companies in our universe ticked up
month-over-month in September (recall that the data is taken from mid- August to
mid-September), with the average short interest for the stocks in our coverage universe up
8% month-over-month. Some increases were more pronounced than others in September,
for instance, short interest in KLAC ticked up 20% month-over-month, short interest in
NVLS increased 30% month- over-month, short interest in AMAT increased 12%
month-over-month, and short interest in ATMI increased 32% month-over-month. Other
stocks registered month-over-month declines in short interest, for instance, short interest
in MKSI declined 46% month-over-month, short interest in AEIS declined 12%
month-over-month, and short interest in LRCX declined 5% month- over-month. We note
that short interest is still significantly lower than the peak levels reached during the last
cycle. For instance, AMAT reached peak short interest of 75 million shares (adjusted for
split) in 2001 vs. current short interest of 41 million shares and KLAC reached peak short
interest of 26 million shares vs. current short interest of 13 million shares.
While short interest is now well off of the 4-year lows reached back in May thus opening
the door for trading rallies along the way to the bottom, it is still significantly below
levels reached in the last cycle implying that if/when short interest continues to move
higher, the SOX Index is likely to move lower. This concept ties with our thoughts below
that we believe it is likely the Street will reach a point in this cycle in which negative
secular concerns begin to dominate short-term trading patterns in the stocks, which would
likely lead to higher short interest levels and a lower price for the SOX index.
ADJUSTING NOVELLUS MODEL TO REFLECT LOWER SHARECOUNT AS THE
COMPANY CONTINUES TO BUYBACK SIGNIFICANT SHARES. Last week,
Novellus Systems issued a press release that highlighted that sharecount as of September
15th, 2004 was approximately 139.9 million. The company ended last quarter with over 151
million shares and we had been modeling 145 million shares for the September quarter, as we were
expecting some share buyback but not one of this magnitude. As a result of the significantly lower
sharecount, we are raising our CQ3'04 EPS estimate to $0.39 from $0.38 and our CQ4'04 EPS
estimate to $0.44 from $0.41 as we now expect the share buyback to continue over the next several
quarters. These increases in EPS estimates are solely driven by the lower sharecount as we made no
changes to our operating assumptions for H2'04.
While our sharecount is now also significantly lower for CY2005, we have made some downward
adjustments to our revenue and margin assumptions for 2005 based on takeaways from our recent
Asia trip, which offset the lower sharecount. As a result, there is no change to our CY2005 EPS
estimate of $1.60. Street consensus for 2005 is $1.50 and the fact that we are modeling above
consensus EPS for 2005 while maintaining a bearish view on fundamentals for 2005 has led to
investor controversy for having intentionally overly aggressive estimates. As we have now
highlighted on many occasions, we believe that the semi equipment companies that enjoy the
buffer of SAB 101 deferred revenues will be able to print higher than expected revenue and EPS in
the early stages of the downturn. For historical precedent, we point to 2001 when Novellus had CY
revenues of approximately $1.34 billion, but shipments of only approximately $1.0 billion. In
other words, without SAB 101, revenues would have been approximately 25% lower. We expect a
repeat of this dynamic in the first year of this semi equipment downturn and that is why we have
above consensus EPS for the front-end companies. For companies who will not enjoy the SAB
101 buffer, we are modeling significantly below consensus revenue for CY2005, reflecting our
view that underlying fundamentals will be below current Street expectations.
WHAT DOES NOVELLUS' AGGRESSIVE SHARE BUYBACK PROGRAM IMPLY ABOUT
ITS EXPECTATIONS FOR THE MARKET? We applaud Novellus management for being the
first major semi equipment company to take the important step of drawing cash balances down
significantly to buy back shares. Novellus management has been more open than any other major
SPE company about the secular challenges facing the semi equipment industry and the
ramifications of slowing long-term growth. As the Street continues to come to grips with the
slowing long-term industry growth rates, we believe that valuation multiples will ultimately be
dictated by return driven metrics (i.e. ROA). With most of the semi equipment companies having
significantly over capitalized balance sheets and therefore not earning any kind of meaningful
return on a big portion of their assets (cash), Returns of Assets (ROAs) are lower than in previous
cycles. The long-term margin pressures which Novellus has been very open about are only
contributing to this problem. We therefore applaud Novellus for taking the difficult step of dealing
with the slowing growth rates head on drawing down excess cash to buy back stock and, in turn,
increasing ROA. While other semi equipment companies have started to buy back some stock
given the weakness in share prices over the last 6 months, most still maintain significant excess
cash, which is dragging down their ROA.
IF/WHEN/WHAT WILL TURN INVESTOR ATTENTION FROM THE LIKELIHOOD OF A
SHORT AND SHALLOW DOWNTURN TO THE MANY SECULAR CHALLENGES FACING
THE INDUSTRY. CAN WE GET THROUGH THIS DOWNTURN WITHOUT THE STREET
ONCE AGAIN BECOMING OBSESSED WITH THE NEGATIVE SECULAR ISSUES
FACING THE SEMI EQUIPMENT INDUSTRY? COULD THIS BE WHAT ULTIMATELY
DRIVES SPE STOCKS TO A TROUGH VALUATION COMMENSURATE WITH OTHER
PURE CYCLICAL INDUSTRIES?
Over the last several weeks as semi and semi equipment companies have shrugged off significant
amounts of bad news, many of the investors who acknowledge that we are in a downturn have
chosen to focus on the belief that the downturn will be short and shallow and thus many feel a need
to be early in buying the stocks.
The question that we keep asking ourselves is whether or not the industry can go through a
downturn without the Street becoming obsessed with the many negative secular issues that are
facing the semi equipment industry. When we think back to the bottom of the last downturn, all
that investors wanted to talk about was how bad the secular outlook for the semi equipment
industry was while we were trying to tell a bullish cyclical story within the context of our negative
secular view. The question that strikes us most at this point in the cycle is whether or not investors
will once again become focused on the negative secular issues facing the industry a few quarters
from now when we are in the heart of the downturn? Said another way, can the semi equipment
stocks really find a bottom until those secular issues are dominating investor mindsets again? One
thing we know for sure, regardless of how this stock cycle plays out, is that we would feel much
more comfortable waiting to recommend the stocks until more of our incoming calls are focused on
the negative secular trends as opposed to how likely it is that the downturn will be brief and
relatively painless.
Trying to predict if/when investors will become preoccupied by the secular issues is difficult.
However, if it is to happen, we be
lieve that it will ultimately be some subset of the issues that we discuss below that will once again
arouse secular fears and drive stocks down to similar trough valuations from the last cycle. While
bullish investors will likely be quick to point out that this downturn isn't likely to be as severe as
the last, we continue to highlight two points: 1) even "normal" downturns (which our work
indicates this will be) are much more severe than Street expectations for the current cycle, and 2) as
the Street continues to recognize that long-term growth rates are slowing, there is likely to come a
point in this downturn where the Street concludes that there need not be a great sense of urgency to
own stocks during a downturn in an industry that is growing at no more than nominal GDP
long-term. This would be the scenario where the Street "re-rates" the semi equipment stocks during
the downturn similar to how the Street "re-rated" the industry during the upturn.
During the last downturn, the most prevalent issues which dominated the negative mindset at the
bottom of the cycle generally fell into three categories:
1) The capital efficiency of the 300mm transition is likely going to hurt the long-term growth rate
of the semi equipment industry,
2) The DRAM industry is in a state of secular decline and therefore, in aggregate, won't ever be
able to spend money in an upturn like it has in the past, and
3) Intel's capex budget is likely in secular decline due to both a slowdown in its long-term growth
rate and the fact that that Intel is further along than the rest of the industry in its 300mm transition
thus decreasing spending going forward.
Well, a funny thing happened during the upturn. It turns out that all of those issues are true. This
upturn was the first in history where capex spending didn't grow peak-to-peak. Moreover, the
CAGR of capex from the peak of the 1996 cycle to the peak of the 2004 cycle is 1%. We believe a
large part of the reason why capex growth has been nonexistent over the last cycle is exactly what
the Street was concerned about during the last downturn. So the obvious question is, "why won't
the Street get concerned about these same issues yet again several quarters from now when we are
in the heart of the downturn?" We believe the answer is, the Street will! It's easy for bullish
investors to highlight how much less severe this downturn is going to be before any of the bad
news starts to come out. However, a few more announcements from companies like the one we saw
from Teradyne last week, in which EPS gets cut in half in one quarter, and we believe it's going to
be a lot harder for the Street to focus on the positive aspects of the downturn (if there is such a
thing).
Moreover, once the Street gets through the annual excitement over the seasonally stronger CQ4
throughout the electronics supply chain, we expect that attention will turn to several additional
issues which are likely to dominate negative sentiment throughout 2005.
1) Due to the significant amount of capex currently being spent by the DRAM industry, our
DRAM team in Taiwan believes that DRAM prices could be as low as $2.50 by CQ4'05. How
excited is the Street going to be about the SPE stocks with DRAM prices 20% below cash costs?
2) As foundry utilization rates dip back into the low 80% range (which we expect to happen in
CQ1'05), we would expect that the Street will once again become concerned about the lack of
pricing power at the foundries (even during the upturn) due to the hyper-competitive environment
in that space. At this point, foundry managements are likely to be reminding the Street of their
commitment to ROE levels over the course of the cycle, which implies a secular decline in capex
levels,
3) While the top tier companies in the space are not likely to lose money in this downturn, the
second and third-tier companies likely will. As this begins to happen and the Street realizes the
consequences of a "normal" semi equipment downturn (as opposed to the bloodbath of 2001 and
2002), secular concerns about the lack of profitability of the industry over the course of cycles are
likely to raise their head again. This could become especially damaging for stocks given that some
of the companies have balance sheets that are not prepared for the downturn given that the stock
upturn didn't last as long as expected and companies never got the opportunity to do equity deals to
"clean-up" the converts that were done during the downturn as has been the case in previous
upturns, and
4) Finally, we expect that several quarters into the downturn, the Street will look back at the
peak-to-peak margin declines for most of the companies in the universe and once again become
secularly disenchanted with an industry that promises margin improvement over the course of
cycles (even on lower revenue levels), but rarely delivers on this promise. The Street might give
some companies the benefit of the doubt by assuming that margins won't deteriorate by as much
during the downturn, but if/when they do deteriorate by more than expected, the Street is likely to
look upon this development with disdain.
Lest we get carried away, let us be clear that we do believe there will be a time when we want to
own the semi equipment stocks again. There are times in every cyclical business (even in secularly
challenged ones) that you have to be long the stocks. We just don't think now is the time. With too
many investors focusing on the likelihood of a shallow cyclical downturn as opposed to the secular
issues which could eventually drive multiples lower during the downturn, we believe investors who
are buying the stocks today could wind up losing a lot of money. We would likely be much more
bullish on the industry once these secular fears become more prevalent but until then, we continue
to recommend that investors avoid overweight positions in the stocks.
ENTEGRIS REPORTING FQ4 ON THURSDAY; MANAGEMENT TONE LIKELY MORE
CAUTIOUS. Entegris is reporting earnings on Thursday before the market open. We model fourth
fiscal quarter (ending August) sales of $99 million, up 0.4% sequentially, with earnings per share
of $0.14, in-line with the Street. Recall that on August 16th, we had cut our estimates on a number
of companies in our sector, including Entegris (for the out-quarters), as we began to gather more
datapoints regarding pushouts and a slowdown in order rates for semi equipment companies.
Although ENTG is likely to fair better than competitors as we move further into the fundamental
downturn, due to the high portion of its revenue (about 65%) coming from unit driven businesses
(which are likely to hold up better over the course of the full downturn), we believe it is unlikely
that the company will be able to escape the negative impact of the current slowdown in wafer
starts, which we believe will continue through the first half of 2005.
Regarding guidance, we believe that the company will be conservative in its outlook, as many
chipmakers have recently cut their wafer start plans and several semi equipment OEMs have
lowered their own build plans. We believe this is evident in several of the recent preannouncements
in our space. Recall that in the last two weeks MKSI and TER downwardly revised their guidance
for CQ3, with Teradyne lowering its previous EPS guidance by approximately 50%. As a result of
4 Goldman Sachs Global Investment Research
September 27, 2004 Analyst Comment
these preannouncements and the overall more cautious sentiment within the industry, we believe
that Entegris is likely to take a conservative approach in its outlook. To that end, we are modeling
below consensus EPS for CQ4 (GS $0.14 vs. Street $0.16) and for FY2005 (GS $0.50 vs. Street
$0.69).
News, Events and Price Performance
Last week
Monday 20 September (1) Asyst Technologies announced that it will team with IBM Global
Services to provide a range of manufacturing automation solutions for
the worldwide semiconductor marketplace. Under the agreement, Asyst will supply its suite of
distributed equipment connectivity solutions, such as the NexEDA and EIB software products, and
IBM Global Services will provide worldwide systems integration services for semiconductor
manufacturers in the United States, Asia and Europe. (2) Novellus Systems announced the
appointment of Matthew Grech to the newly created position of vice president, investment and
investor relations. Grech will be responsible for the company's cash investment activities, and
replaces Robin Yim as the contact point for members of the investment community. (As vice
president and treasurer, Yim will continue to manage Novellus' treasury function). Grech's previous
experiences include positions at Fidelity Investments, Thomas Weisel Partners, Zweig-DiMenna
Associates and most recently heading his own investment management firm. (3) Novellus Systems
also announced that its board of directors has approved an increase to Novellus' previously
authorized stock repurchase program. The terms of the expanded program will permit Novellus to
repurchase up to an aggregate of approximately $1.1 billion of its outstanding common stock
through September 14, 2009. (4) ASAT Holdings Limited announced that it has received the ISO
9001:2000 certification for its new manufacturing facility in Dongguan, China. (5) Brooks
Software (division of Brooks Automation) announced that Powerchip has awarded a multi-million
dollar contract for the deployment of Brooks' fault detection solution at its new 300mm factories.
(6) Agilent Technologies introduced a modular semiconductor parameter analyzer (4157B) that
provides a customizable parametric analysis solution for extreme measurement challenges. (7)
LTX Corporation announced that Berkana Wireless is testing its BKW9000 family of single-chip,
quad-band CMOS transceivers for GSM/GPRS applications on LTX's Fusion CX. (8) Ultratech
reaffirmed its Q3 and full (fiscal) year 2004 EPS guidance of $0.10 to $0.15 and $0.20 to $0.30,
respectively. The company also stated that it expects revenue to be up 40% to 60% sequentially for
Q3 and up 10% to 20% for FY 2004. (9) Teradyne revised its Q3 2004 revenue guidance to $450
million to $470 million from $530 million to $560 million and its Q3 EPS guidance to $0.21 to
$0.26 from $0.39 to $0.46. Please see our 9/21 note for additional details.
Tuesday 21 September (1) Teradyne and Acterna announced that they are working together to
provide broadband test solutions for leading telephone companies; shared technical specifications
relate to Teradyne's LDU 100 and Acterna's QT-200 and QMS Test OS & EMS. Teradyne also
announced that it is working with Spirent Communications towards DSL service assurance test
solutions. Their shared technical specifications also relate to Teradyne's LDU 100 and Spirent's
CopperMax test platforms. Prototypes will be shown at the upcoming Broadband World Forum
Europe 2004, trade show in Venice, Italy (September 20th - 23rd). (2) Brooks/Cimetrix's
automation solutions were used by SOLVision for its OEM automation hardware and connectivity
software needs; SOLVision was able to complete all of the automation requirements for its wafer
inspection tool in 10 weeks. (3) Akrion announced that its wholly owned subsidiary, Goldfinger
Technologies, has received an order for five Goldfinger Mach2HP single wafer systems from a
DRAM manufacturer in Asia. (4) Varian Semiconductor announced today that it has shipped its
250th single wafer VIISta platform ion implantation system. (5) Credence Systems announced the
appointment of Lori Holland to its Board of Directors. Holland operates an independent consulting
practice servicing Silicon Valley-based companies focused on the semiconductor and components
manufacturing industry. Her previous experience includes management positions at Bookham
Technology, Read-Rite, Zaffire, and NeoMagic. As a member of the Credence Board, Holland's
responsibilities will include participation on the Company's Audit Committee.
Wednesday 22 September (1) Asyst announced that it has named Robert J. Nikl to the position of
Senior Vice President, Finance and Chief Financial Officer. Nikl has over 25 years of experience in
financial management and public accounting, including management positions at Solectron
Corporation, Xerox Corporation and Xerox Engineering Services. (2) Rudolph Technologies
received multiple orders from a Taiwanese foundry for its MetaPULSE-II opaque film metrology
systems. (The foundry has purchased more than ten systems in the last twelve months). (3)
FormFactor announced the expansion of its global network of service centers with new locations in
Tokyo, Japan and Livermore, California. The centers will provide applications support, training
and repair services to customers.
Thursday 23 September (1) Novellus Systems announced the shipment of its 100th ALTUS PNL
module for tungsten deposition. The tool will be installed in a 300 mm production fab in Korea,
and will be used for advanced tungsten applications.
Friday 24 September (1) Credence Systems announced that ELMOS Semiconductor signed a
multiple-system purchase agreement for the Credence Piranha system. (2) Applied Materials
entered into an agreement with Novellus Systems to dismiss all pending lawsuits between the
companies involving allegations of patent infringement. The agreement provides for a payment to
Novellus of $8 million and waives a claim for unpaid royalties of approximately $3.5 million.
This week's calendar:
Thursday 30 September: (1) Entegris (ENTG-$8.00; IL/N) reporting earnings; GS and Street $0.14.
GS Universe Price Performance 9/24/04 Price performance
Ticker Company Name Rtg Close Week MTD QTD YTD Y-Y
Semiconductor Capital Equipment
AEIS Advanced Energy Industries IL/N 9 -8% -6% -41% -65% -55%
AMAT Applied Materials IL/N 17 -1% 5% -15% -26% -10%
ATMI ATMI Inc. IL/N 20 -5% 7% -26% -13% -21%
ACLS Axcelis Technologies IL/N 8 -5% 3% -35% -22% -9%
BRKS Brooks Automation IL/N 14 -5% 10% -32% -43% -38%
CMOS Credence Systems U/N 7 -2% 4% -50% -47% -45%
ENTG Entegris IL/N 8 -8% 1% -31% -38% -37%
FORM FormFactor OP/N 20 -2% 12% -13% -1% -13%
KLAC KLA-Tencor OP/N 40 0% 7% -19% -32% -24%
LRCX Lam Research IL/N 22 -4% 1% -19% -33% -6%
MKSI MKS Instruments IL/N 15 -6% 11% -34% -48% -36%
NVLS Novellus Systems IL/N 26 0% 6% -18% -38% -27%
TER Teradyne Inc. U/N 13 -3% 2% -42% -48% -36%
Mean -- -- -4% 5% -29% -35% -27%
Median -- -- -4% 5% -31% -38% -27%
Source: Factset.
I, Jim Covello, hereby certify that all of the views ex |
