Changewave Research: Semiconductor and Semi Equipment Trends

ChangeWave Research: Semiconductor and Semi Equipment Trends

ChangeWave Research Report:

Semiconductor and Semi Equipment Trends –

A Modest Rebound for Chips

Abstract

Back in June, our Alliance 2Q/3Q Quarterly Sales survey showed an uptick for the semiconductor industry, raising hopes that a meaningful rebound was underway.

To follow-up, we surveyed our semiconductor industry members on emerging market trends, along with the segments and companies most likely to win or lose in the chip and equipment marketplaces. The survey was conducted during the week of July 30 – August 4, and a total of 50 members participated.

Key Findings

A Modest Rebound for Chips. The rebound in the semiconductor industry is a modest if somewhat shallow one, creating a narrow range of winners but leaving significant parts of the industry still lagging behind.

· Two-thirds (66%) of respondents see little or no possibility that the chip sector will experience double-digit growth during the second half of 2003.

· But three-in-five (60%) respondents expect double-digit growth for chips in 2004.

· Graphics/Video Chips (net +16) and Specialty Chips (net +14) are considered the best performers over the next six months. Memory Chips (net -20) and Microprocessor Chips (net –16) are seen as the weakest performers.

· Xilinx (net +33) is seen as the chip company best positioned to pick up market share over the next six months, followed by Texas Instruments (net +27), SanDisk (net +25), Intel (net +21) and Zoran (net +21).

· AMD (net –34) and National Semiconductor (net –22) were seen as the chip companies most likely to lose market share over the next six months.

· By a 44% to 28% margin, respondents say chip sales continue to be a viable leading indicator of corporate IT sales and capital spending.

· But changes in IT company inventory management are having a great impact on corporate visibility – and 52% of respondents say chip companies can no longer provide accurate forecasts beyond one quarter.

· On the semiconductor equipment side, Lithography (net +11) and Wafer Processing (net +9) are seen as the top performing sub-segments for the next two quarters

Bottom Line: These survey results point to a modest rebound for the overall semi industry, coupled with a mixed outlook for certain sub-segments. For example, the trend is upbeat for graphics/video chips for consumer devices, but the semi equipment sector lags and will continue to struggle to return to historical growth rates.

Helping You Profit From A Rapidly Changing World ™

www.ChangeWaveResearch.com

Table of Contents

Summary of Key Findings 3

The Findings 4

Summary and Analysis 14

ChangeWave Research Methodology 16

About ChangeWave Research 17

I. Summary of Key Findings

Back in June, our Alliance 2Q/3Q Quarterly Sales survey showed an uptick for the semiconductor industry, raising hopes that a meaningful rebound was underway.

These survey results focus on four areas:

· Semiconductor Industry Trends

· Semiconductor Companies

· Semiconductor Indicators

· Semiconductor Equipment Sector

II. The Findings

A. Semiconductor Industry Trends

(1) Question Asked: Do you believe there will be double-digit growth for the semiconductor sector in either the 3rd or 4th Quarters of 2003?

Current Survey
Aug ‘03 / Previous Survey Jul ‘02*
Yes - Double-Digit Growth in 3Q 2003 / 2% / 4%
Yes - Double-Digit Growth in 4Q 2003 / 22% / 9%
Yes - Double-Digit Growth in Both 3Q and 4Q 2003 / 4% / 2%
Little Possibility of Double-Digit Growth in 3Q or 4Q 2003 / 52% / 51%
Absolutely No Possibility of Double-Digit Growth in 3Q or 4Q 2003 / 14% / 31%
Don't Know / 6% / 2%

*Note: In the July 2002 survey, this question was asking about the 3rd and 4th Quarters of 2002, whereas in this survey we are asking about the 3rd and 4th Quarters of 2003.

Mild Optimism Over Near-Term. While 28% of respondents expect double-digit growth for the second half of 2003, 66% see little or no possibility for this level of growth during the second half. Nevertheless, these numbers are more upbeat than the results from our July 2002 semiconductor survey.

(2) Question Asked: Do you believe there will be double-digit growth for the semiconductor sector in FY2004?

Current Survey
Aug ‘03 / Previous Survey Nov ‘02 / Previous Survey Jul ‘02
Yes - Double-Digit Growth in FY2004 / 60% / 29% / 53%
Little Possibility of Double-Digit Growth in FY2004 / 30% / 42% / 27%
Absolutely No Possibility of Double-Digit Growth in FY2004 / 2% / 27% / 9%
Don't Know / 8% / 2% / 11%

*Note: In the November and July 2002 surveys, this question was asking about FY2003, whereas in this survey we are asking about FY2004.

But a Considerably Better Picture for 2004. Three-in-five (60%) respondents expect double-digit growth for the semiconductor sector in 2004, double the 30% who see little chance for double-digit growth next year. Only 2% completely rule out the possibility for double-digit growth in 2004.

(3) Question Asked: Which one of the following semiconductor sub-segments do you think will perform the best over the next six months? (Please Choose Only One)

Current Survey
Aug ‘03 / Previous Survey Nov ‘02 / Previous Survey Mar ‘02
Communications Chips / 28% / 9% / 5%
Graphics/Video Chips / 20% / 22% / 22%
Logic/Programmable Chips / 14% / 16% / 32%
Memory Chips / 4% / 4% / 19%
Microprocessor Chips / 4% / 4% / 5%
Specialty Chips / 28% / 42% / 11%

(4) Question Asked: And which one of the following semiconductor sub-segments do you think will perform the worst over the next six months? (Please Choose Only One)

Current Survey
Aug ‘03 / Previous Survey Mar ‘02
Memory Chips / 24% / 11%
Communications Chips / 22% / 51%
Microprocessor Chips / 20% / 8%
Specialty Chips / 14% / 8%
Logic/Programmable Chips / 10% / 0%
Graphics/Video Chips / 4% / 5%

Net Difference Score Chart

Perform Best / Perform Worst / Net Difference Score
Graphics/Video Chips / 20% / 4% / +16
Specialty Chips / 28% / 14% / +14
Communications Chips / 28% / 22% / +6
Logic/Programmable Chips / 14% / 10% / +4
Microprocessor Chips / 4% / 20% / -16
Memory Chips / 4% / 24% / -20

Best Performing Segment – Graphics/Video Chips. Looking at the next six months, Graphics/Video Chips and Specialty Chips topped the list of best performers with scores of net +16 and net +14 respectively. Also, Communications Chips (net +6) have made impressive gains over previous surveys.

Worst Performing Segment – Memory Chips. The weakest performers will be Memory Chips (net -20) and Microprocessor Chips (net -16), according to industry respondents.

(5) Question Asked: Capital spending by chip companies can be characterized as falling into one of four categories:

· 300mm fab construction;

· shrinking to .18 micron or less;

· shifting to copper; and

· fab improvements to support increased shipment volume.

Which one of the following four categories do you see as the number one area for capital spending by chip companies over the next 12 months?

Current Survey
Aug ‘03 / Previous Survey Jan ‘03 / Previous Survey Nov ‘02 / Previous Survey May ‘02
Fab Upgrades for Shrinking to .18 Micron or Less / 58% / 37% / 42% / 50%
New 300mm Fab Construction / 12% / 32% / 33% / 22%
Fab Improvements to Support Increased Shipment Volume / 8% / 8% / 9% / 13%
The Shift to Copper / 2% / 8% / 9% / 9%

Majority of Cap-Ex Dedicated to Upgrading to Smaller Sizes. According to 58% of respondents, fab upgrades to .18 micron or below will be the prime beneficiary of capital spending by chip companies. Expenditures on 300mm fab construction (12%) and fab upgrades to support increased shipment volume (8%) lagged far behind, while the shift to copper is now garnering only 2%.

(6) Question Asked: We hear rumors of chip overcapacity in some areas of the industry (e.g., memory) and undercapacity in other areas (e.g., sub .18 micron foundry lines). Which areas do you see as having overcapacity, and which as having undercapacity before year's end?

We wanted to find out what our members thought about capacity issues within the semiconductor industry. Interestingly, the majority of respondents supported our premises for overcapacity among memory manufacturers and undercapacity at the sub .18 micron foundry lines.

Sample of Alliance Members Responses

a. Overcapacity

· KJO9758 writes, "Customers are having trouble getting wafers because of cutbacks and closed lines. Once visibility beyond 1Q can be trusted, TSMC, CSM and the like will begin to open new lines. ATE has overcapacity but there is no desire to purchase to forecast at 0 margins."

· INV9492 writes, "Overcapacity 90nm, under 0.18um and maybe 0.13um next year if we get double digit growth."

· ASC8035 writes, "Memory."

· CHI8142 writes, "Overcapacity, I believe is graphics chips. The buzz is that very few people require the super high end chips which leaves oversupply in the middle-of-the-road graphics chips.”

· KIM0900 writes, "The over-capacity in memory will continue.”

· VIR2071 writes, "Over capacity in greater than .18 micron wafer production capacity."

· HAM5839 writes, "Many of the analog type devices seems to have a very short lead time (i.e. possible overcapacity) while I don't see in my area any undercapacity in any of the products. Lead time have generally been less than 4 to 6 weeks for most products. "

· KDP6740 writes, "Overcapacity in low-end process technology (above 0.18 micron CMOS)."

· HFC8959 writes, "Memory was overcapacity for long time."

· STE7087 writes, "Memory - over capacity."

· SHA6674 writes, "Memory still over capacity vision/video.”

· DSP7412 writes, "Memory."

· WTK2295 writes, "DRAM overcapacity.”

· BCO9216 writes, "SRAM and dram Memory is definitely over capacity ASIC applications are the most promising."

b. Undercapacity

· PHI1286 writes, "Undercapacity for techno 0.18 and below because of no more capacity built over the last 2 years."

· CHI8142 writes, "Undercapacity may be .13 micron parts."

· KEV8327 writes, "Under capacity in the memory areas, production is moving down to 130 and 110nm nodes."

· MAR4145 writes, "Newer technologies mostly have capacity shortages: sub 0.18. Older technologies of capacity surplus. Almost all new designs are being done on the newer technologies, because of the advantages of smaller die size and better performance. Semi equipment spending is/will continue to be driven by the newer fab technologies. This will increase as more of the new designs ramp into production. The older fab processes will likely never return to full capacity utilization. Or will only do so when much of this capacity has been shut down or converted sub 0.18 micron."

· KIM0900 writes, "The under-capacity for 0.15 micron and below will get even worse."

· JAY9297 writes, "Undercapacity: Sub .18 micron.”

· SHA6674 writes, "Vision/video under capacity specialty under capacity."

c. Other

· KRI9418 writes, "Prices on flash have been going up!! SanDisk and others should have a great quarter.

· SHU6149 writes, "The situation of overcapacity and undercapacity is not very clear right now. Things can change very quickly. With cutthroat competitions, I guess that a lot of player will be hurt. Only the nimblest ones can survive."

· JES5813 writes, "RF fabs are running no more than 50% in industry. There are not enough designs yet to fill the sub .18 foundries. One wafer boat on a 300mm fab is ~100,000 units of a 1cm square ide. What end applications have that demand other than wireless. If a semi does not have wireless in its portfolio it will have difficulty surviving in the market in 5 years when 300mm is mainstream."

B. Semiconductor Companies

(1a) Question Asked: Let's focus on the chip companies themselves. Based on your own knowledge and what you are seeing in the industry, which of the following chip vendors are most likely to pick up market share over the next six months? (Choose All That Apply)

Xilinx / 35%
Texas Instruments / 35%
Intel / 34%
SanDisk / 31%
Zoran / 21%
Nvidia / 17%
Broadcom / 15%
Maxim Integrated Circuits / 15%
ATI Technologies / 15%
AMD / 12%
Analog Devices / 10%
Altera / 8%
LSI Logic / 8%
Vitesse Semiconductor / 8%
National Semiconductor / 6%
Microchip / 6%
Intersil / 4%
Semtech / 4%
Standard Microsystems / 0%

(1b) Question Asked: And which of the following chip vendors are most likely to lose market share over the next six months? (Choose All That Apply)

AMD / 46%
National Semiconductor / 28%
LSI Logic / 16%
Nvidia / 14%
Vitesse Semiconductor / 14%
Intel / 14%
Broadcom / 12%
Standard Microsystems / 10%
Altera / 8%
Intersil / 8%
Semtech / 8%
Analog Devices / 8%
Texas Instruments / 8%
Maxim Integrated Circuits / 8%
SanDisk / 6%
Microchip / 6%
Xilinx / 2%
ATI Technologies / 2%
Zoran / 0%

Net Difference Score

Gain Market Share / Lose Market Share / Net Difference Score
Xilinx / 35% / 2% / +33
Texas Instruments / 35% / 8% / +27
SanDisk / 31% / 6% / +25
Intel / 34% / 14% / +21
Zoran / 21% / 0% / +21
ATI Technologies / 15% / 2% / +13
Maxim Integrated Circuits / 15% / 8% / +7
Broadcom / 15% / 12% / +3
Nvidia / 17% / 14% / +3
Analog Devices / 10% / 8% / +2
Altera / 8% / 8% / 0
Microchip / 6% / 6% / 0
Intersil / 4% / 8% / -4
Semtech / 4% / 8% / -4
Vitesse Semiconductor / 8% / 14% / -6
LSI Logic / 8% / 16% / -8
Standard Microsystems / 0% / 10% / -10
National Semiconductor / 6% / 28% / -22
AMD / 12% / 46% / -34

Market Share Gainers – Xilinx Tops the List. Looking at the next six months, Xilinx is the chip company best positioned to pick up market share, followed by Texas Instruments, SanDisk, Intel and Zoran.