GS US Semi & SPE: GS Tech Conference
Day 1 - Product cycle stories stand out
Semi and SPE takeaways from Day 1 at the GS Tech Conference:
1) Semi companies are
very bullish about business trends but are clearly less confident when the conversations
shift to inventory issues, which only fuels our fears about a repeat of the 2004 inventory
driven downturn, 2) Semi equipment companies are also very bullish about continued
order growth into CQ2?06 and, while admitting visibility is limited for H2?06, SPE
managements are hopeful about continued order growth throughout the year, 3) Our own
take is that our meetings on Day 1 of the conference lead us to believe that the product
cycle driven stories in both semis and SPE (i.e. MRVL, BRCM and FORM) are now,
more than ever, the most attractive options for investors over the more cyclically driven
stocks as the semi cycle seems to be on borrowed time from both an inventory and
capacity perspective.
Semiconductor and Semiconductor Equipment (SPE) presentations from the first day of
the Tech Symposium were generally upbeat, as near-term business conditions appear solid
and CQ1 results appear to be tracking at least in-line with expectations. The only cautious
issue we noted was that the companies seemed notably less bullish when the
conversations turned to inventory issues, which doesn't help alleviate our concerns around
rising inventories in the channel with IC units the furthest above their trendline since the
2000 bubble and inventories accumulating in certain market segments (i.e. Taiwanese
fabless companies, Intel, Texas Instruments, Marvell, EMS companies, etc.). SPE
companies were also bullish, highlighting continued solid order momentum into CQ2.
While SPE companies do not have clear visibility into H2'06, managements are hopeful
that business will remain solid throughout 2006 driven primarily by memory and logic
spending. We remain concerned that order momentum will slow as 2006 progresses given
the aforementioned inventory concerns potentially leading to a decline in utilization rates.
Relative to the stocks, we come away continuing to believe that investors should focus on
product cycle driven stories including Marvell, Broadcom (assuming investors can get
comfortable with the options expense issue), and Form Factor, where we expect to
continue to see robust growth, and continue to avoid the more cyclically driven semi and
SPE companies where it seems that we are at/close enough to the peak of the fundamental
cycle that investors are rightly beginning to discount the peak in the stocks.
FORM (OP/C) - PRESENTATION UNDERSCORES THAT FORM REMAINS THE
BEST GROWTH STORY IN SPE. FormFactor hosted a bullish presentation during
which the company focused on product cycle growth drivers, supporting our view that
FormFactor remains the best growth story in SPE. Management also touched upon the
company's manufacturing strategy and current IP litigation with Phicom.
(1) Long-term growth drivers - FormFactor's growth continues to be driven by: A)
Memory - new memory probe card sales are being driven by robust DRAM and NAND
bit growth, 300mm transitions, linewidth shrinks, DDR2/DDR3, and mobile RAM, as
each technology transition or new device design drives the need for a new custom probe
card. B) Move of final test to the wafer level - FormFactor expects the move of final test
to the wafer level through Known Good Die (KGD) and wafer level burn-in applications to drive
the most meaningful growth for the company over the next 12-18 months. KGD refers to
conducting high speed test at the wafer level, with growth in this segment being driven by the
proliferation of consumer electronic devices and the use of multi-chip packages. Management
expects KGD applications to become a more important growth driver than NAND flash in 2007.
Wafer level burn-in refers to testing devices in high temperature conditions at the wafer level,
which is also expected to drive meaningful growth in the intermediate-term. C) Logic - Lastly, the
company continues to drive growth through greater penetration in the logic market.
(2) Manufacturing strategy - The demand environment has been stronger than expected, which
management reiterated during its presentation. As a result of stronger than expected demand,
FormFactor announced on its Q4 earnings call that it has pulled forward its capacity expansion
plans at its new manufacturing facility. Near-term capacity levels are targeted at $100-125M/quarter
and long-term capacity levels are targeted at $150-200M/quarter, compared to previous capacity
target levels of a little over $80M/quarter. The company has kept its old facility open in order to
meet more customer demand, which has resulted in an ~2% drag on gross margins. Management
expects to regain most of the 2% loss in gross margins by the end of 2006, but was not explicit
about how long the old facility is expected to remain open.
(3) IP protection - FormFactor continues to be focused on defending its IP position and is currently
involved in several lawsuits with a Korean competitor, Phicom. Recall that FormFactor is suing
Phicom for patent infringement in both Korea and the US. We understand that Phicom is
challenging the validity of Formfactor's patents in the Korean courts, with the courts having
invalidated some of the patents, which FormFactor is suing Phicom for infringing. It strikes us that
the company's legal dispute with Phicom has recently not gone in FormFactor's favor, but
management appears confident in its ability to compete against Phicom even if it is not successful in
its legal disputes with the company.
No change to our OP/C rating, as FORM remains our favorite long-term growth and product-cycle
driven story in SPE. We continue to expect FormFactor's revenues to grow at a rate of 15-20% over
the next few years while we expect the rest of the SPE market to continue to grow at a rate of about
3-5%. We would use any pullbacks to add to positions as we believe the premium valuation is more
than justified by what we believe to be FORM's industry best secular growth rate. The shares are
expensive at 37x our FY2006E EPS of $1.00 (including ESOs) and 31x our FY2006E EPS of $1.20
(excluding ESOs). On a relative valuation basis, we would note that the shares look similarly valued
to the other SPE names, with the front-end SPE stocks trading at about 30-35x our normalized EPS
estimates.
SNDK (IL/N) - BULLISH PRESENTATION FOCUSED ON GROWTH OPPORTUNITIES AND
ROYALTY INCOME. SanDisk's presentation was upbeat, with the company highlighting
long-term growth drivers in the NAND flash market and much Q&A centered around
IP/litigation-related issues (see our analyst meeting note from 2/23 for additional details).
(1) Long-term growth - SanDisk continues to expect robust bit growth for NAND flash in the next
several years, driven by growth in handsets (removable flash storage), computing (i.e. instant-on
applications), and other consumer applications (i.e. gaming, MP3 players, etc.). Specifically,
SanDisk expects NAND bit growth in the range of 130-170% and a revenue CAGR of 25-40% over
the next 3 years. In 2006, management expects to cut prices by 50-55% Y/Y, with price reductions
in Q1 of about 25-30% Q/Q. Also of note, the company believes the one-time programmable 3D
technology it purchased when it acquired Matrix in 2005 may be a potentially disruptive technology
if the company is able to make it read-writable.
(2) Elasticity - Bit growth in the NAND flash market has historically been driven by continued
reductions in prices driving increased demand and new end markets. SanDisk cut prices by about
40% in early February in the US retail market, with price cuts in Europe and Asia in January
(consistent with guidance for a Q1 price decline of about 25-30% Q/Q). Management indicated that
it takes between 60-90 days before these price cuts would drive elasticity, so if the price cuts
stimulate significant bit growth demand, this should be visible in late Q1/early Q2.
(3) Capacity plans - Management reiterated its expectation that Fab 3 (it's 300mm JV fab with
Toshiba) will ramp in capacity to 30k wafer starts per month (WSPM) in March 2006 and then to
70k WSPM in March 2007 and finally to about 100k WSPM by the end of 2007 (which is the
facility's full capacity). The company will decide if it will invest in a second 300mm fab with
Toshiba, Fab 4, in approximately the 2007 timeframe, with the final decision to move forward with
another facility dependent on the demand environment. In 2006, the majority of production will be
on 70-nm MLC, with 55-nm MLC production beginning at the end of 2006.
(4) IP strategy - SanDisk intends to seek more license/royalty contracts from competitors that it
believes are utilizing its technology. Recall that Samsung accounts for the majority of
royalty/license revenues, but SanDisk also believes that STMicro, Hynix, and Micron are uti
lizing its technology without paying a royalty/license fee (note that Intel is not violating SanDisk's
patents due to a cross-licensing agreement and would therefore not be responsible for paying any
license/royalty fees for the output it sells from its IM Flash JV with Micron). SanDisk recently
brought suits before the ITC against STM for 3 patents that it believes STM is violating with its
NOR and NAND products. The company expects to hear an outcome from the ITC in regards to the
recent suits in 2007.
There is no change to our IL/N rating on the stock. We view the issue of royalty/license revenue
sustainability as one of the most important potential drivers of the stock, given that royalty revenues
account for a significant percentage of operating income (we estimate about 60% in 2006) and
SanDisk's contract with Samsung expires in August, 2009. We are also concerned about potential
excess supply in the NAND flash market given significant capital investments being made in
2006/2007. That said, we would wait to take a more aggressive stance on the shares until we better
understand how much demand is being stimulated by recent price cuts and whether this demand will
be enough to soak up all of the supply set to come on-line.
Our reluctance to take a more cautious
view on the stock in the near-term is driven by the fact the company's long-term growth prospects
have driven tremendous outperformance in the stock over the past 12 months (SNDK increased
152% Y/Y in 2005), despite more recent underperformance (SNDK is down 3% YTD vs. our
Semiconductor Devices coverage universe median performance of up 8% YTD). The stock is also
expensive at 31x our 2006E EPS of $1.95 (including ESOs) and 27x our 2006E EPS (excluding
ESOs) of $2.26.
KLAC (OP/C) - NEAR-TERM BUSINESS TRENDS APPEAR SOLID WITH ORDERS LIKELY
UP Q/Q IN Q2. KLA hosted a bullish presentation during which management indicated that
near-term business trends appear solid with orders likely up again in Q2 Q/Q. The company also
highlighted its recent acquisition of ADE and the competitive environment.
(1) Current business environment - the current business environment continues to be robust. Recall
that KLA guided CQ1 orders +15% Q/Q, +/- 10%, and management indicated that it expects CQ2
orders to increase again Q/Q. While KLA recognized that it is too early to be certain on the business
outlook in H2'06, it is optimistic about business prospects, with increases in orders in H2'06 likely
to come from memory and logic customers. Management noted that NAND-related spending
appears flat Q/Q in CQ1 but looks likely to increase Q/Q in CQ2. With respect to 2006 capex
budgets, KLA expects memory spending to be in the range of flat to +10% Y/Y, with foundry
spending likely to increase about +10-20% Y/Y, and overall capex likely up +10-15% Y/Y (we
forecast Y/Y capex growth of about 8-10%). Management also noted that the foundries no longer
appear to be adding capacity based upon a certain "trigger" utilization rate level, and are instead
adding capacity more strategically. Foundry spending is picking up in CQ1 and is expected to
increase again Q/Q in CQ2, according to the company. Management is also cautiously optimistic
regarding foundry spending in H2'06.
(2) ADE acquisition - late last week KLA announced its intention to acquire ADE for ~$490M in a
stock-for stock transaction, which is expected to close in June/July. Management indicated that it
expects the deal to be accretive to earnings in the first year after it is completed, although the
company did not quantify the extent to which the deal will be accretive. We view the acquisition as
a positive, as we believe it will help the company extend its reach into the bare wafer market and it
increases KLA's scale. Importantly, we applaud management for taking a leading role in driving
consolidation in the SPE industry. With regards to additional M&A activity on the horizon, the
company indicated that it is continuously on the look-out for new opportunities that would be
complimentary to KLA's business model.
(3) Competition - KLA remains confident in its competitive position. The company does not
consider Applied Materials' UVision brightfield wafer inspection tool to be a meaningful
competitive threat. Recall that UVision was introduced to the market last year. Importantly, KLA
continues to appear unwilling to compromise on price in customer negotiations.
There is no change to our OP/C rating on KLAC, as we continue to believe that KLA has one of the
best business models in SPE with high margins, dominant market share, and defensive revenue
drivers, as the company's sales tend to be driven by technology transitions which hold up even
during softer periods for the industry. We expect SPE order momentum to slow as 2006 progresses
given our concerns about an inventory correction in semis leading to declining utilization rates
which would, in turn, drive lower orders to the SPE suppliers. Given KLA's more defensive
revenue-drivers, we therefore prefer the stock on a relative basis. Note that the shares have lagged
the group at +5% YTD vs. our SPE coverage universe median performance of +19% YTD. We
would not own the shares on an absolute basis, however, as the stock is trading at 29x our estimate
of normalized EPS of $1.80 (excluding ESOs) and 26x peak EPS (excluding ESOs).
LRCX (U/C): LAM MEETING ALSO HIGHLIGHTS SOLID NEAR-TERM BUSINESS
TRENDS. Lam's presentation was bullish, with management highlighting good near-term business
conditions, as well as recent share gains.
(1) Current business environment - Consistent with KLA, Lam believes that orders will be up again
sequentially in CQ2 and management is optimistic regarding H2'06 fundamentals, although
visibility is limited that far out. Lam continues to see a robust business environment in 2006 with
capex budgets being revised upward to +12-15% Y/Y from earlier expectations for capex to
increase +7-10% Y/Y in 2006 (we forecast capex growth of +8-10% Y/Y). Management indicated
that memory spending remains strong, with NAND spending expected to account for about $6-7B
of capex in 2006, out of total memory capital spending of about $17B. The company also noted that
the foundries tend to accelerate orders when leading edge utilization rates approach 95% or higher.
To that end, the company indicated that foundries have already begun to order and Lam expects to
see increased order activity from UMC and SMIC in 2006.
(2) Market share gains and competition - Lam continues to be focused on gaining market share and
indicated that the company expects to exit the year with 43-44% market share in etch. With regards
to the Japanese market, management indicated that market share gains are fairly difficult to come by
and the company would be willing to sacrifice a few points in margins in order to gain incremental
share in Japan. That said, management indicated that in order to accelerate the company's overall
growth, Lam needs to grow outside of the etch market. The company currently has an offering of
strip and clean solutions and continues to look for other products/acquisition targets to compliment
its existing product portfolio.
(3) Balance sheet - Lam has cash of ~$980M and no long-term debt, and it seems as if the company
would be more willing to consider a dividend once its stock buyback is large enough to offset
options-related dilution on a run rate basis.
While the near-tem fundamental environment appears robust, there is no change to our longer-term
U/C rating on Lam as: (1) We expect SPE order momentum to slow as 2006 progresses given our
concerns about an inventory correction in semis leading to declining utilization rates which would,
in turn, drive lower orders to the SPE suppliers. (2) The company is likely approaching its peak
gross margin levels in CQ1'06 and we do not believe investors should own cyclical stocks that are
at or approaching previous peak gross margin levels, as they are no longe
r likely to get any incremental operating leverage by owning the shares when margins are nearing/at
their peak. (3) The company's significant exposure to memory spending (80% of last quarter's
orders) creates outsized risk given our view that NAND could be in excess supply in 2006. And (4)
the stock remains expensive at 26x our estimate of normalized full cycle earnings vs. the S&P 500
at 17x.
XLNX (IL/N) - PRESENTATION FOCUSED ON LONG-TERM OUTLOOK, MARGIN
PROFILE, AND MANUFACTURING STRATEGY. Our fireside chat with Xilinx focused on the
long-term growth outlook for Xilinx and the PLD industry, the company's margin profile relative to
Altera, and Xilinx's dual foundry manufacturing strategy. We rate Xilinx In-Line/Neutral as we
view valuation at 26x CY'06 P/E as reasonable given healthy growth prospects and returns profile
and would look to be more aggressive pending improved visibility regarding acceleration of
communications, which we believe will be necessary to re-establish outsized growth prospects for
Xilinx and PLDs.
(1) Xilinx maintains optimism regarding long-term PLD industry growth, which the company
expects will continue to outpace the broader semiconductor industry. The company pegs the
semiconductor industry growth rate at high-single digits to low-teens and expects PLDs will show a
couple points better growth. Xilinx's optimism regarding PLD industry prospects is driven by
end-market diversification, in particular related to the explosion in consumer-consumption of
semiconductors, but also by improved trends in communications infrastructure, which continues to
represent just under 50% of PLD industry sales.
(2) PLD market share continues to concentrate around the industry's two leaders: Xilinx and Altera,
which continue to capture share at the expense of smaller participants. Xilinx remains optimistic
regarding its position in the industry based on improved traction of newer products, which it
believes has outpaced Altera over the past few quarters as Xilinx gets past issues at 130nm and
90nm process transitions. In addition, Xilinx is hopeful regarding upcoming 65nm process
transition (likely to begin towards the end of CY'06) as it believes there will not be as many issues
at 65nm since tools are essentially the same (except for two modules) and the company has
implemented more rigorous preparatory measures.
(3) Xilinx is constructive on its dual foundry manufacturing strategy as it believes this approach
mitigates risk and provides some cost advantage as it is able to negotiate better pricing through
competition, which Xilinx believes offset the dual costs associated with bringing up wafers at two
foundries.
(4) Xilinx and Altera have different target gross margin models (Xilinx 62-63% vs. Altera 64-66%).
Xilinx believes there are many reasons for this disparity, including Xilinx having more product
SKUs, manufacturing efficiencies, and asset utilization. Xilinx continues to evaluate the difference
and will look to drive better efficiencies and margin opportunities, including continued driving of
cost optimization through the life of products, including when move to mainstream and mature
status.
LSI (U/N) - HIGHLIGHTING GROWTH DRIVERS IN CONSUMER AND STORAGE;
CONTINUE TO LOOK TO SHARPEN INVESTMENT FOCUS. Our fireside chat with LSI Logic
focused on growth opportunities and progress of restructuring/re-organization efforts, which are
on-going. We rate LSI Underperform/Neutral as we view valuation at 28x CY'06 P/E as somewhat
aggressive given our mixed view of the company's long-term sales opportunities due to conservative
outlook for ASICs. However, we continue to be encouraged by operational improvements, which
Goldman Sachs Global Investment Research 5
Analyst Comment February 28, 2006
could continue to be a driver for earnings and investor support over the intermediate term.
(1) LSI highlighted growth opportunities, with particular optimism related to storage (components
and systems) and the consumer. The company expects storage opportunities to be keyed by
traditional strong positioning and upcoming technology transitions (4-gig fibre channel and SAS),
in which the company has strong market share and could drive unit and ASP improvements. LSI is
also optimistic about consumer opportunities with outlook for growth in DVD-recorder and
MP3/audio over the intermediate term and emerging opportunities related to digital TV and Zevio
applications processor platform solution, which LSI introduced at CES. Zevio is a platform solution
targeting a broad range of low-end portable devices including GPS, E-toys/edutainment, and
personal media players. Zevio combines ARM9 processor, ZSP DSP, and 3D graphics (from Koto
Laboratory, which is led by famed former Nintendo executive) and is slated to operate at
150-250mw; the solution targets end device price point under $140 with Zevio ASP expected to be
below $10 in volume.
(2) The company continues to look to sharpen overall business focus, evaluating opportunities on
scale and leverage potential. No update on actions or potential areas of focus, though we believe the
company is evaluating investment in RapidChip and traditional CoreWare standard cell ASICs. We
expect the company will continue to participate in custom silicon opportunities, but will look to be
active in more targeted areas, in which it is able to leverage building block technologies such as
DSP and high-speed SERDES.
NSM (IL/N) - UPBEAT PRESENTATION REFLECTING SOLID NEAR-TERM DEMAND
TRENDS AND CONTINUING BUSINESS MODEL IMPROVEMENTS. National
Semiconductor's presentation was upbeat, with management seeming very comfortable with the
current end demand environment and positive about ongoing improvements to its business model.
Note the company is reporting FQ3 (Feb) earnings results next Thursday, March 9.
(1) Current environment - management commented that it is seeing good business trends in the
handset end market in CQ1'06, with inventory appearing lean at customers, and no downward
revisions to orders placed in CQ4. For context, handsets represent about a third of National's total
sales, with the majority coming from power management; National's top handset customers are
Motorola, Nokia and Samsung. On the distributor side, which represents about half of National's
business, the company is also seeing steady demand, and perceives its inventory in the distribution
channel to be on the lean side. Lead times remain in the 6 week range for over 85% of all products,
consistent with the level of last quarter.
(2) Business model improvements - management noted that reaching their 60% gross margin target
is likely "mechanical" at this point, meaning that it is largely a function of transitioning out of the
low-margin foundry business and closing the Singapore assembly and test facility, rather than
underlying business mix improvement. Foundry services, which the company is performing to
support the new owners of its Super I/O and cordless handset businesses, represented 8-9% of FQ2
(Nov) sales but are expected at just 2-3% of sales in a year. Given their 20-25% gross margin,
mathematically this would equate to about 200 basis point improvement in gross margin. Similarly,
the shut-down of the Singapore assembly and test facility, expected to be completed by August
2006, will contribute around $5-7 million in cost savings, or around 100 basis points in gross
margins. Summing the two effects would result in a 60% gross margin, up from 57.2% in the
November quarter. Longer term, gross margin expansion will have to depend on growth in
higher-margin product areas such as operational amplifiers and data converters, where the primary
competitors are ADI, Linear and Maxim.
(3) Wireless infrastructure - National has observed a recent improvement in wireless infrastructure
demand from equipment makers such as Siemens, Alcatel and Huawei, partly attributing it to a
buildout in India. This presents a greenfield opportunity for National's data converter and amplifier
products. Wireless infrastructure represents abou
6 Goldman Sachs Global Investment Research
February 28, 2006 Analyst Comment
t 6-8% of company sales, and may see an additional uptick in 2H'06 as China starts upgrading its
next-gen infrastructure.
We continue to rate NSM In-Line in the context of our Neutral semiconductor coverage view with a
positive bias on company fundamentals, and recommend buying on any dips. On our CY'06
estimates, NSM is trading at 25x, essentially in-line with the analog median of 26x and slightly
below the semiconductor median of 28x. We view this as a fair multiple, as we expect National's
bottom-line growth rate will keep up with other stocks in our universe due to a combination of
leadership in the high-growth power management segment, and share gains in high-performance
analog, offset somewhat by de-emphasis of non-high performance analog products.
Each of the analysts named below hereby certifies that, with respect to each subject company and its
securities for which the analyst is responsible in this report, (1) all of the views expressed in this
report accurately reflect his or her personal views.. |
