GS US SEMI EQUIP WEEKLY: LET FUNDAMENTALS, NOT VALAUTION,
DICTATE STOCK THESIS SUMMARY: (1) Leading indicators including weak back-end data points and forward
looking foundry capacity utilization rates continue to point to a fundamental downturn
that we don't believe is priced into the stocks. Recall that foundry capacity utilization is
one of the best leading indicators for the stocks, (2) Some arguing that even if the upturn
is over, the next downturn will be short and shallow and therefore that the yet-to-be seen
fundamental weakness is priced into the stocks. We suggest letting fundamentals, not
valuation, dictate the stock thesis. If equipment shipments remain high over the next few
quarters, the downturn will likely be more severe than expected. If shipments slow
meaningfully over the coming quarters, the downturn will likely be less severe, (3) Trough
price/tangible BV multiples can be misleading, the Rule of Three will likely determine a
better entry point for the stocks, (4) Short interest data ticked up again in July but is still
close to 4- yr lows, and (5) News, Events and Price Performance.
LEADING INDICATORS CONTINUE TO POINT TO A FUNDAMENTAL
DOWNTURN THAT WE DON'T BELIEVE IS PRICED INTO THE STOCKS. It's
always tough to "separate the forest from the trees" after a busy week of earnings like the
one we had last week. That said, when we take a step back and try to focus on what the
leading indicators for the semi equipment cycle are telling us, we believe that
increasingly weak data points in the back-end of the semi equipment sector and declining
capacity utilization rates at the foundries make it even more likely that the industry is now
beginning its next cyclical downturn. We expect that the length and severity of the
downturn will be a topic of great debate over the coming 6-9 months (more on this a bit
later) but for those who are still not convinced that the downturn is at hand, we find the
commentary from the packaging and test houses and leading foundries most interesting.
CONTINUED WEAK DATA POINTS IN THE BACK-END A LEADING
INDICATOR OF THE CYCLE PEAKING. Last week there were a number of weak data
points in the back-end including cautious commentary from packaging and test houses,
Amkor and ASAT Holdings, regarding high customer inventory levels and guidance from
leading test supplier Advantest for a decline in orders in the September- quarter. Amkor
reported its June-quarter results last week, indicating that it saw softness in demand for its
advanced packaging products due to the absorption of IC inventories during the
June-quarter and that it expected softness to continue into the September-quarter due to
inventory build. As a result of the higher inventory levels at customers, Amkor guided for
flat sales in September in what is typically a seasonally stronger quarter. Regarding
capacity, management indicated that the company currently has sufficient capacity and
will begin to add capacity more selectively, with 2005 capital spending expected to
decline to $100 million from $400 million in 2004. ASAT Holdings offered similar
commentary regarding inventories and capacity when it negatively preannounced its
July-quarter results on Friday, indicating that sales would decline 14% sequentially as
opposed to the company's original guidance for flat to +4% sales growth. ASAT
indicated that an inventory correction in the communications and consumer sectors, a
shortage in small geometry wafers, and an accelerated capacity expansion at the test and
packaging houses that was leading to pricing erosion were drivers of the revenue
shortfall. Finally, Advantest reported its earnings last week with June-quarter orders up 14%
sequentially and guidance for a 20% decline in September-quarter orders. We continue to
highlight that we believe the weakness at the packaging and test houses (as well as weak order
guidance from Advantest) is a leading indicator for the entire semi equipment cycle, as the
back-end tends to lead the industry into and out of downturns. We believe that Tokyo Electron
underscored the weakening fundamental outlook for the front-end semi equipment suppliers with
its view that September-quarter orders (including both SPE and flat panel) will decline 15-20%
sequentially with SPE orders expected to decline 5% to 10% sequentially in the September-quarter
with FPD orders expected to decline as much as 60% sequentially.
SIGNIFICANT INCREASES IN WAFER START CAPACITY IN H2'04 AT THE LEADING
FOUNDRIES PORTENDS LOWER CAPACITY UTILIZATION RATES. RECALL THAT
FOUNDRY CAPACITY UTILIZATION IS ONE OF THE BEST LEADING INDICATORS FOR
SEMI EQUIPMENT STOCKS. As we have been highlighting for the last several months, we
believe it will be excess supply, not a lack of demand, which will ultimately lead the semi
equipment industry into its next cyclical downturn. We got some very concrete signs of the
significant amount of capacity that is set to come online in H2'04 from the Taiwan foundries that
reported last week. First, UMC highlighted that its wafer start capacity is expected to increase 19%
sequentially in CQ3 (including the acquisition of a fab from SiS) and another 8% sequentially in
CQ4. TSMC commented on its earnings call on Thursday morning that its wafer start capacity is
expected to increase 10% sequentially in CQ3 and another 5% sequentially in CQ4.
What is perhaps the most troubling aspect of the utilization issue is that we don't expect we will see
the worst of the foundry customers' inventory drawdown, and in turn the harshest decline of
foundry capacity utilization rates, until H1'05. The problem is that the foundries (or at least two
out of the three major foundries) will be shipping well above trendline in CQ3. SMIC guided to a
23-27% increase in CQ3 shipments and UMC guided to a 15-16% increase in CQ3 shipments.
TSMC had more muted CQ3 shipment guidance of 4-6%. The main takeaway here is that UMC
and SMIC's customers are likely to continue to build inventory in H2'04 as the end markets aren't
growing at those accelerated rates. As a result, we believe that foundry customers are likely to be
decreasing orders to UMC, SMIC, etc. at the same time that significant incremental wafer start
capacity will be coming online. This dynamic is likely to create a meaningful decrease in foundry
capacity utilization rates and lead to a significant decline in orders to the equipment suppliers. We
continue to believe that foundry utilization rates are one of the, if not the best leading indicators for
the semi equipment stocks.
SOME ARGUING THAT EVEN IF THE CYCLE IS OVER, THE NEXT DOWNTURN WILL
BE SHORT AND SHALLOW AND THEREFORE THAT THE STILL-TO-COME
FUNDAMENTAL WEAKNESS IS PRICED INTO THE STOCKS. WE BELIEVE THAT
FUNDAMENTALS, NOT VALUATION, WILL TELL US WHEN TO BUY THE STOCKS. One
of the most interesting aspects of the debate raging on Wall Street right now about the semi
equipment stocks is that everyone seems to have already made up their minds about what the next
downturn is going to look like. The funny thing about that is that no one can know the answer to
that question or even begin to make an intelligent argument until we see how the leaders in the
semi equipment industry react over the coming quarters.
What we mean by this is that the severity of the downturn is going to be a function of how much
more wafer start capacity gets shipped over the coming quarters (i.e. how much worse the already
existing excess capacity situation gets). While the semi industry is already headed fro an excess
capacity situation, in our minds, what will determine whether the excess capacity gets as bad as the
last cycle or whether excess capacity can get soaked up more quickly (thus leading to a
shorter/shallower downturn) is how quickly the semi equipment shipments slow down over the
next few quarters. For example, if Applied Materials ships $2.5 billion in equipment for the next
several quarters (well above trendline) then the current excess capacity situation will get
exacerbated as that high level of shipments creates even more capacity that needs to be absorbed
over the next 12-18 months. Under this scenario, the downturn will last longer and be more severe
than anyone is currently contemplating which would lead to the stocks having 40% more downside.
However, if Applied reacts more quickly to this downturn (and resists the temptation to stretch to
bring in orders/ make shipments at the end of a quarter just to make Wall Street targets), then this
downturn could indeed be less severe and the stocks will have only 20% or so more downside this
cycle. Remember that during the last downturn, the supply/demand situation in the semi industry
didn't truly begin to fix itself until Applied Materials negatively preannounced a 30% order
shortfall in January of 2003. It's not a coincidence that the semi industry was back in a tight
supply/demand situation 9 months after Applied stopped shipping more capacity to customers than
they reall
y needed. Pricing aggressively at the end of a quarter (which we believe that Applied did) only
serves to prolong a downturn by giving customers more capacity than they need. Wall Street needs
to wait to see how this dynamic plays out over the coming quarters before being overly aggressive
on the semi equipment stocks.
TROUGH PRICE/TANGIBLE BOOK VALUE MULTIPLES CAN BE MISLEADING, USE THE
RULE OF THREE TO DETERMINE A GOOD ENTRY POINT FOR THE STOCKS. We got a
surprisingly high number of calls last week from clients (who were dusting off trough valuation
analyses from 2002) that argued that we were close enough to trough valuation to buy the semi
equipment stocks. The argument goes that if we take an average of the trough price/tangible book
value multiples from the last 2 or 3 cycles, then the semi equipment stocks are close enough to
trough to start establishing positions. In our view, the problem with this analysis is that is doesn't
account for the large swing in trough multiples over the last several cycles. For example, NVLS
traded as low as 1.7x its tangible book value in the 2002 downturn However, in the 1998
downturn, it only traded down to 2.6x. Investors are trying to simplify the trough analysis by taking
an average of the last few cycles. This exercise would indeed lead one to the conclusion that some
stocks (or at least NVLS) are cheap enough. Taking an average of the last two cycles' multiple,
2.1x, times an approximate tangible book value exiting 2004 of about $12 would lead one to buy
the stock in the mid-20s today. However, using only the 2002 trough multiple implies a trough
value of $20 and would lead one to conclude that it is too early to buy the stock. This then leads
into the debate we referenced above about whether or not this downturn is going to be as severe as
the 2002 downturn. With + multiple point making the difference between buy or sell, we find the
trough multiple analysis very imprecise. This type of tail chasing exercise is what led to so many
headfake rallies or false bottoms in the SOX from late 2000 to late 2002.
Rather than argue about + multiple points on tangible book value, we intend to stick with the Rule
of Three framework that we first introduced in March 2002. Recall the three rules: 1) Have cash
flow margins bottomed out, 2) Has management capitulated, and 3) Is there a glimmer of hope that
business is poised to improve. Please see any of our pure cyclical reports in order to get the full
detail behind these rules but the main idea is that we can't hit a bottom in the stocks until
managements avoid giving away capacity such that the supply/demand equation remains out of
balance and the stocks can't begin to work (for anything more than a trade) until there are at least a
few tangible signs of hope that business is improving. We plan to carefully follow these rules
during the upcoming downturn to avoid getting caught in headfake rallies over the coming
quarters.
SHORT INTEREST DATA TICKED UP AGAIN IN JULY BUT IS STILL CLOSE TO FOUR
YEAR LOWS. Short interest data for the month of July came out last week (recall that the data is
taken from June 15th to July 15th). Short interest ticked up about 9% month-over-month for the
stocks in our coverage universe after increasing another 9% month-over-month in June, but is still
close to four- year lows. Significant increases in short interest in July were notable in FORM (up
190% month-over-month), KLAC (up 17% month-over-month), LRCX (up 34%
month-over-month), TER (up 44% month-over-month), NVLS (up 8% month- over-month), and
CMOS (up 10% month-over-month). The increases in LRCX and KLAC ahead of their respective
quarterly reports were likely the reasons behind the strong short-covering rallies post their reports.
Goldman Sachs Global Investment Research 3
Analyst Comment August 1, 2004
Significant declines in short interest in July were visible in AEIS (down 9% month-over-month)
and MKSI (-30% month-over-month). While short interest was clearly up in July, we would note
that the data is still close to four- year lows likely as a result of the hedge fund community taking a
more balanced approach on the stocks as he head into the typically seasonally stronger H2'04.
News, Events and Price Performance
Last week
Monday 26 July (1) Veeco Instruments (VECO-$22.75; NC) reported $0.15; Street $0.13. (2)
Advanced Energy was found guilty of infringing on patents underlying MKS Systems' ASTRON
reactive gas generators. (3) Semitool licensed its seed layer enhancement patents to Ebara
Corporation. (4) Asyst Technologies announced that it's CFO, David While, has resigned to accept
a CFO position at a Fortune 200 company. Asyst also announced that it will miss its June quarter
sales guidance of $145mn-$155mn. The company now estimates June quarter sales of $141mn. (5)
Robotic Vision raised approximately $1.9 million through a private placement. (6) Rudolph
Technologies (RTEC-$16.00; NC) reported $0.09; Street $0.08.
Tuesday 27 July (1) Asyst Technologies' joint venture, Asyst Shinko, was selected to install its
Automated Material Handling System in a Generation 6 flat panel display factory in Taiwan. (2)
Tokyo Electron placed an order for Therma-Wave's Integra metrology tools to be used with Tokyo
Electron's Clean Track Act and Clean Track Lithius coater developers in 90nm and 65nm
production. (3) Therma-Wave (TWAV-$16.49; NC) reported -$0.03; Street - $0.04. (4) Tegal
Corporation announced that a wireless component supplier placed a multiple system order for its
Advanced Control System i901 series diode plasma etch systems.
Wednesday 28 July (1) Varian shipped single wafer VIISta 80HP high current and VIISta 810EHP
medium current ion implanters to a 300mm fab in China. (2) FEI Company (FEIC-$20.11; NC)
reported $0.13; Street $0.13. (3) Axcelis Technologies (ACLS-$9.33; IL/N) reported $0.30; GS
$0.30; Street $0.27. Please see our 7/29 note for additional details. (4) Varian shipped a single
wafer VIISta 80HP high current ion implanter to Nanya Technology. (5) MEMC Electronic
Materials (WFR-$9.09; NC) reported $0.20; Street $0.18. (6) Veeco received an order for its
GEN2000 molecular beam epitaxy system from a wireless device manufacturer.
Thursday 29 July (1) Brooks Automation (BRKS-$14.41; NC) reported $0.31; GS $0.29; Street
$0.30. Please see our 7/29 note for additional details. (2) Entegris introduced a new clean-in-place
method for hygienic cleaning of filter housings using Entegris' precision spray devices. (3)
Mykrolis Corporation (MYK-$10.00; NC) reported $0.18; Street $0.22. (4) Varian Semiconductor
(VSEA-$29.87; NC) reported $0.52; Street $0.51. (5) KLA- Tencor (KLAC-$41.23; NC) reported
$0.48; GS $0.46; Street $0.45. Please see our 7/30 note for additional details. (6) Veeco received a
$10 million order to supply GaNzilla MOCVD (metal organic chemical vapor deposition)
production systems to Lumileds.
This week's calendar:
Friday 6 August: (1) Varian Semiconductor Equipment Associates annual analyst meeting in
Gloucester, MA.
GS Universe Price Performance 7/30/04 Price performance
Ticker Company Name Rtg Close Week MTD QTD YTD Y-Y
Semiconductor Capital Equipment
AEIS Advanced Energy IL/N 10 12% -37% -37% -62% -47%
AMAT Applied Materials IL/N 17 4% -14% -14% -24% -12%
ATMI ATMI Inc. IL/N 20 5% -25% -25% -12% -19%
ACLS Axcelis Technologies IL/N 9 7% -25% -25% -9% 26%
BRKS Brooks Automation IL/N 14 -2% -28% -28% -39% -23%
CMOS Credence Systems U/N 9 -3% -35% -35% -32% -4%
ENTG Entegris IL/N 9 3% -23% -23% -31% -38%
4 Goldman Sachs Global Investment Research
August 1, 2004 Analyst Comment
FORM FormFactor OP/N 20 4% -11% -11% 1% 8%
KLAC KLA-Tencor OP/N 41 6% -17% -17% -30% -20%
LRCX Lam Research IL/N 24 5% -11% -11% -26% 10%
MKSI MKS Instruments IL/N 15 13% -35% -35% -49% -34%
NVLS Novellus Systems IL/N 27 3% -14% -14% -36% -22%
TER Teradyne Inc. U/N 17 2% -25% -25% -33% 1%
Mean -- -- 5% -23% -23% -29% -13%
Median -- -- 4% -25% -25% -31% -19%
Source: Factset.
I, Jim Covello, hereby certify.. that all of the views expressed in this report accurately reflect my
personal views about the subject company or companies and its o |
