Opportunity Calls: Semiconductor Consumption in Key Communications Applications Prepared for: In-Stat’s 20th Anniversary Forum Report #EA0104MF Emerging Semiconductor Applications Service By: Cahners In-Stat Group May, 2001 Edited by: Christie Van Gaal In-Stat Phone: (480) 483-4446 E-Mail: cvangaal@instat.com Cahners In-Stat Group 1101 S. Winchester Blvd., Building N • San Jose, CA 95128 • Tel: 408.345.4438 • Fax: 408.345.4400 6909 E. Greenway Parkway, Suite 250 • Scottsdale, AZ 85254• Tel: 480.483.4440 275 Washington St. • Newton, MA 02458-1630 www.instat.com • info@instat.com Copyright Cahners In-Stat Group 2001. All rights reserved. Reproduction in whole or in part is prohibited without written permission from Cahners In-Stat Group. This report is the property of Cahners In-Stat Group and is made available to a restricted number of clients only upon these terms and conditions. The contents of this report represent the interpretation and analysis of statistics and information that is either generally available to the public or released by responsible agencies or individuals. The information contained in this report is believed to be reliable but is not guaranteed as to its accuracy or completeness. Cahners In-Stat Group reserves all rights herein. Reproduction or disclosure in whole or in part to parties other than the Cahners In-Stat Group client who is the original subscriber to this report is permitted only with the written and express consent of Cahners In-Stat Group. This report shall be treated at all times as a confidential and proprietary document for internal use only. Cahners In-Stat Group reserves the right to cancel your subscription or contract in full if its information is copied or distributed to other divisions of the subscribing company without the written approval of Cahners InStat Group. Authentic Copies of this Report Feature a Red Color Bar Table of Contents: The Strength Behind this Report ..................................................................... 1 Executive Summary............................................................................................. 2 Introduction.......................................................................................................... 4 Methodology ........................................................................................................ 5 Definitions.................................................................................................. 6 Cracks in the WW Economic Foundations Followed by Cooling Demand Quickly Landed us in an Oversupply Situation .......................... 9 Economic Picture Remains Iffy ...................................................................... 9 And Our Semiconductor Outlook Isn’t Pretty . . ...................................... 11 Semiconductor Consumption by Application Segment.................... 12 US End-Use Market Orders Allow Us a Glimpse of The Future ............ 14 Communications IC Market: Wired Versus Wireless ............................... 17 Market Drivers ........................................................................................ 17 Specific Opportunities: Semiconductor Consumption by Segment...... 20 Networking Equipment ......................................................................... 20 Cable vs. DSL........................................................................................... 22 DSLAM ............................................................................................................ 26 Cable Headends.............................................................................................. 27 Cellular Handsets/Terminals....................................................................... 31 Summary – Time to Rethink a Few Things.................................................. 37 © 2001 Cahners In-Stat Group – EA0104MF – Confidential i List of Tables: ii Table 1. Worldwide Electronics Consumption – Dollars In Millions ................ 3 Table 2. Worldwide Semiconductor Dollar Shipments (Dollars in Millions).. 11 Table 3. Worldwide Semiconductor Consumption By Major End-Use Category (Dollars in Millions).......................................................... 13 Table 4. Total Communications Semiconductor Shipments Per WSTS – Wired Versus Wireless (Dollars in Millions) ................................... 19 Table 5. Dollar Value for Key Comm Segments for Equipment and Semis – 2001 versus 2004 (Dollars in Millions) ............................................ 36 © 2001 Cahners In-Stat Group –EA0104MF – Confidential List of Figures: Figure 1. Total Semi Dollars by Traditional End-Use (% of Total) – 1995 vs. 2000 vs. 2005.................................................................................... 13 Figure 2. US Electronic Component Growth vs. WW Semi Sales Growth – 1991 through 2000 ............................................................................ 15 Figure 3. U.S. Booked to Billed Ratio for Electronic Components – 1995 to 2001 .................................................................................................. 15 Figure 4. U.S. Booked-to-Billed Ratio for Total Communications Equipment – 1995 to 2001 .................................................................................. 16 Figure 5. U.S. Booked-to-Billed Ratio for Office & Computing Equipment – 1995 to 2001 ..................................................................................... 16 Figure 6. Total Communications IC – Wired vs. Wireless 2004 ......................... 19 Figure 7. Networking Switches & Semis – 2001 vs. 2004 ................................ 21 Figure 8. Networking Hubs & Semis – 2001 vs. 2004 ...................................... 21 Figure 9. Routers & Semis – 2001 vs. 2004 ...................................................... 22 Figure 10. Worldwide Cable vs. DSL Subscriber Forecast ............................... 23 Figure 11. XDSL Equipment & Semis – 2001 vs. 2004 .................................... 25 Figure 12. Cable Equipment & Semis – 2001 vs. 2004 ..................................... 25 Figure 13. DSLAM and Semis – 2001 vs. 2004 ................................................ 28 Figure 14. Cable Headend & Semis – 2001 vs. 2004 ........................................ 28 Figure 15. Optical Networking Equipment & Semis – 2001 vs. 2004............... 30 Figure 16. Cellular Handsets & Semis – 2001 vs. 2004 .................................... 32 Figure 17. Cellular Base Station Semiconductors.............................................. 34 © 2001 Cahners In-Stat Group – EA0104MF – Confidential iii iv © 2001 Cahners In-Stat Group –EA0104MF – Confidential The Strength Behind this Report The technology industry of 2001 is not the one of 1995 – It’s not even similar to 2000. Now, on the downside of an extremely successful year, we’re forced to rely on solid business fundamentals and ingenuity in lieu of simply taking orders. This report represents a compilation of forecasts and insights from around Cahner’s InStat Group. The strength of the message herein lies with the combined experience of the wireless, data & voice, converging technologies & semiconductor research groups within In-Stat. And, that experience is considerable. This study examines the total end equipment revenue and semiconductor content for key communications categories in 2001 and 2004. This is represented at a macro level – total dollar values for each segment. Likewise, detailed semiconductor content is not addressed here – total semiconductor dollar values for each market segment are presented. However, total semiconductor consumption and even total worldwide electronics consumption are forecast through 2005. © 2001 Cahners In-Stat Group – EA0104MF – Confidential 1 Executive Summary The rapidity with which the locomotive steamed over the cliff caught everyone by surprise, even the most seasoned industry veterans. But when multiple ordering combined with an undetected cooling in demand began to take a toll along about last September, an eerily familiar dread took hold in the business world of high technology. And, as dot.coms, largely responsible for inventory miscalculations, began to fizzle and the US economy weakened, things really began to get ugly . . . By nearly every measure, the communications market had become the growth driver for the new hi-tech economy. As the world's digital communications infrastructure continued to build out at a fairly frenetic pace, the technology industry's focus shifted from “pure computing” to the diverse array of devices required to access voice, data and video --, and to the technologies and equipment required to deliver an increasingly complex menu of applications and content. Regardless of form factor, whether cellular handsets, set top boxes, gateways, PDAs, MP3 players, notebooks, or desktops, the world had become focused upon accessing content as needed, when and wherever that need had arisen. While the Internet had become the growth driver in the wired communications market - primarily supporting data transfer -- voice communications had driven the initial growth for the massive wireless push. However, as convergence continued to run rampant, the clear-cut delineation once again had vanished as voice became enabled over IP networks, and data and Internet access became enabled over the cellular/PCS networks. The technology market of 2000 was clearly different from the semiconductor and electronics arenas of 1995 that were so intent on providing computing power for computer applications. But the shift had become more complex from a quantitative perspective, than simply throwing chips from the computing bucket into the communications bucket. A growing portion of computer OEM’s business began to be comprised of boxes that support the communications infrastructure – computing devices built to support networks (including the Internet), telecom infrastructure, cellular base stations and cable head-ends, for example. Not only had chips become application-specific, the boxes themselves had become optimized for certain applications, and a goodly portion of those devices were also functionally computers, being used in communications, running software that is optimized for communications applications. But, regardless of the manner you used to categorize chips, equipment and services, 2000 was the year to be involved in any way. Growing by leaps and bounds, the technology industry, driven by communications and access, was growing and profiting 2 © 2001 Cahners In-Stat Group – EA0104MF – Confidential at an amazing rate. And then it stopped. The orders stopped coming and those that were in hand began to be pushed out or cancelled. The technology/electronics industry, in a way, found itself in a situation that as a whole it had not experienced before. The world’s electronics industry is experiencing the down side of a classic semiconductor industry condition known as boom/bust. Communications gear companies, from networking equipment purveyors to handset manufacturers have not really encountered this wholesale, sobering situation prior to this year – build out for these companies was still too new and too much of a whirlwind. While semiconductor manufacturers (at least those who have been around for a while) have this drill down, the guys who consume the semiconductors may not be quite as accustomed to this scenario. Communications companies didn’t even feel the chip companies’ pain during the 1996 to 1998 time frame. But, even the most successful of the communications giants will remember the recent past and the year 2001. Great guns growth leads to ordering like there’s no tomorrow. Fearful that they could miss an opportunity, companies go overboard to ensure that product is available. However, in the heat of the market, no one really seems to notice quickly enough when demand cools a bit. All of a sudden (or so it seems) players from all levels of the value chain find themselves scratching their heads, with too much inventory in a market that has fallen off of a cliff. Couple that with US, and hence worldwide, economic difficulties, and no good can come of this. At least not for a few quarters. We at In-Stat feel that this situation should begin to stabilize a bit during fourth quarter of this year, barring anything catastrophic on the economic scene. Second quarter isn’t likely to get much better, but third quarter will flatten out, followed by a moderate rise in fortunes during fourth quarter. But, calendar year 2001 won’t look very pretty on paper. Figure 1 shows worldwide electronics consumption for the forecast period. Table 1. Americas % Change Europe % Change Japan % Change Asia Pac % Change Total Consumption Percent Change Worldwide Electronics Consumption – Dollars In Millions 1999 $464,245 $298,705 $143,422 $160,667 $1,067,039 2000 $501,187 8.0% $323,256 8.2% $153,394 7.0% $190,303 18.4% $1,168,140 9.5% 2001 $511,219 2.0% $330,234 2.2% $155,647 1.5% $195,397 2.7% $1,192,497 2.1% 2002 $545,321 6.7% $354,281 7.3% $165,410 6.3% $221,000 13.1% $1,286,012 7.8% 2003 $583,050 6.9% $378,256 6.8% $178,600 8.0% $254,901 15.3% $1,394,807 8.5% 2004 $619,068 6.2% $401,276 6.1% $192,480 7.8% $287,162 12.7% $1,499,986 7.5% 2005 $662,857 7.1% $432,415 7.8% $212,365 10.3% $313,265 9.1% $1,620,902 8.1% Source: In-Stat, Reed Electronics © 2001 Cahners In-Stat Group – EA0104MF – Confidential 3 Introduction The technology market has undergone a dramatic shift that has taken it from being focused largely upon computing technology to being an integrated partner in a digitally communicating realm, where the number of applications (and technologies required to implement those applications) are wide and varied. Where chips destined for computers once comprised over 50% of all semiconductors worldwide, the market is now, and will continue to become more and more diverse in terms of semiconductor applications. And, with this shift comes more complex inventory issues than what existed for computing – parts were pretty much parts in the land of computers. The digital communications infrastructure brings with it a wide variety of customers and customer needs, from many camps. This puts a whole different angle on the inventory scene, especially if you’re a chip guy. The intent of this study is to clarify the impacts that this dramatic shift will cause in terms of semiconductor demand and to simultaneously provide a good picture of key communications market segment performance for 2001 and for 2004. While semiconductor technology as a whole remains the enabling building material for any sort of electronics revolution, it is clear that there will be specific areas that will, in dollar terms, represent strategic contribution over the next five years. Infrastructure, services, access devices and technology will represent tremendous opportunity for all involved in the communications value chain over the forecast period. While services don’t impact semiconductor consumption directly, it is important to note that current trends would indicate that services represent a recurring revenue stream that will be crucial to tap into, whether you are a technology provider, a hardware vendor or rolling out infrastructure. Specific communications applications are segmented, forecasted and discussed in terms of market dynamics, In-Stat’s assumptions and vendor opportunities in the ensuing sections. 4 © 2001 Cahners In-Stat Group – EA0104MF – Confidential Methodology In-Stat utilized a concurrent top-down, bottoms-up methodology to arrive at the forecasts presented in this report. Analysts from each pertinent In-Stat market and component technology group began by compiling equipment unit forecasts based on market dynamics, projected penetration rates, infrastructure roll-outs, competitive pressures from market and technology perspectives and standards issues, to name a few of the major factors. Semiconductor content for each market segment was then assessed predicated upon technology requirements, integration trends, projected capacity, pricing and competitive pressures, etc. In-Stat’s position and views were then discussed in depth with key players in both the wired and wireless communications markets, in our best effort to define next-generation system and product requirements and features. Vendors contacted included: • Alcatel • Cisco Systems • Lucent Technologies • Ericsson • Mitel • Motorola • NEC • Nokia • Nortel Networks • PMC-Sierra • Toshiba Finally, the top-down assumptions are reconciled with the bottoms-up projections. This is an arduous process, but is in reality the truest depiction of what is likely to take place in these market segments. © 2001 Cahners In-Stat Group – EA0104MF – Confidential 5 Definitions For the purposes of this report, In-Stat’s total communications semiconductor forecast is first divided into major splits between wired and wireless, comprised of the following: • Wired Communications - Networking Equipment. Includes Hub Ports, Switch Ports, NIC, LOM, ATM WAN, ATM LAN, Router Ports, Remote Access Servers and WAN IP Telephony. - Analog Modem - XDSL Modem - Cable Modems - Cable Infrastructure - DSLAM - Telecom Equipment - Facsimile - Telephone Handsets, Analog (Corded and Cordless) and Digital Cordless - Other • Wireless Communications - Cellular Handsets/Terminals - Cellular Basestations - Wireless LANs - LMDS/MMDS - Other Broad strokes at wired and wireless were then segmented as follows: • Networking ICs consumed by Hubs, Switches and Routers • Broadband ICs destined for Cable, DSL and Other • Wireless ICs consumed by Handsets and Infrastructure • Wired ICs in consumed by Telecom Switches and Optics It is important to note that this study focuses only upon chips consumed in the communications sector. That being said, this 6 © 2001 Cahners In-Stat Group – EA0104MF – Confidential report does not size the total wireless chip market – it will only account for wireless communications chips. As wireless technology permeates every facet of our lives, we must realize that wireless chips are being shipped into each of the classic end-use segments. So, the wireless devices being worn at the drive-through at McDonalds, the wireless inventory mechanisms and the wireless technology that enables bedside patient check-in at the hospital would all be counted in the traditional industrial category. The wireless functionality in PDAs and notebooks would be counted in the computing segment, and the GPS “On-Star” chips in cars would, of course, be counted in automotive. We feel that functionality and final application is the most effective manner in which to account for these shipments, whether for chips or the end equipment that consumes them. Even then, as we all know, this process is difficult at best. But, In-Stat has painstakingly accounted for a large portion of these market segments and we feel that there is a probability that the total chip market is being underestimated, and some of the adjustments that we have to make in order to adhere to reported numbers are less than satisfactory. By every measure, whether it be geographic, product-based or by vertical industry, equipment shipments as reported by industry associations coupled with input we receive from vendors tells us that reported chip numbers are too low. Considering the fact that bottoms up estimates are almost always larger than the total, we feel there has to be a solid and plausible middle ground between what is being reported for the semiconductor market and what is being reported by the equipment manufacturers whose products consume the chips. Handsets (which are estimated with TIA data and vendor input) alone accounted for more than 90% of the wireless communications chip number in 2000. In-Stat has a long history and in-depth understanding of the movements of the semiconductor industry as reported by the WSTS. If the reporting methodology is left unchanged, there is no way to grow the communications IC portion of the total market to a level that would account for what we see coming down the pike in communications equipment. Therefore, please note that the In-Stat forecasts for total semiconductor by geography and total semiconductor by traditional end-use segment are based upon the market as currently reported by the WSTS. Some of the individual pieces we have chosen to highlight are not based upon the current semiconductor © 2001 Cahners In-Stat Group – EA0104MF – Confidential 7 industry reporting structure. As we feel that our forecasts tend to be conservative to begin with, based on solid logic, we are more than comfortable with the revenue figures that appear herein for each of the segments. 8 © 2001 Cahners In-Stat Group – EA0104MF – Confidential Cracks in the WW Economic Foundations Followed by Cooling Demand Quickly Landed us in an Oversupply Situation That which goes up nearly always comes down, and the worldwide semiconductor market is definitely the poster child for that old adage. But beginning with fourth quarter 2000, global economic conditions began to also contribute to a slow-down in end-equipment markets. We started to witness an erosion in order growth an backlog at the equipment level along about October – but, overheated demand and build-out for the 15 months prior wasn’t easy to put the brakes on. Momentum in growth for everything Hi-Tech, particularly broadband communications, landed us in an all too familiar position: massive inventory build-up that had catapulted past demand, which unfortunately, can stop on a dime. Economic Picture Remains Iffy We don’t expect that things will pick up appreciably on any front during 2001. Economic conditions continue to be iffy at best – difficulties in the US in particular ripple soundly throughout the world. After a very healthy 5.0% GDP growth rate in 2000, 1.7 is the consensus for 2001. Easier fiscal and monetary policy should allow the U.S. economy to recover modestly rising to 3.1 percent in 2002. A sustained interest rate cutting campaign by the Federal Reserve combined with expected tax cuts should help to offset a slowdown in growth prompted by a deterioration of the business environment. Many in the economic community feel that the brunt of the slowdown should be felt only during the first half of this year. However, we would caution that a recovery could be pushed out beyond the second half of 2001 if consumers and businesses continue to pull in the reigns on spending in the face of unemployment rises and investment declines. Meanwhile, unemployment is expected to rise, helping to keep a lid on inflation, with the proportion of the labor force out of work forecast to rise to 4.6 percent in 2001 and 5.0 percent in 2002, compared with 4.0 percent last year. © 2001 Cahners In-Stat Group – EA0104MF – Confidential 9 Inflation, as measured by the OECD's GDP deflator, is forecast to come in at 2.3 percent in 2001, up from 2.0 percent in 2000, but will moderate to 1.9 percent in 2002. The OECD inflation forecast includes investment prices as well as export and import prices, not just consumer prices. And, while there are bits and pieces of good news, the overarching theme remains – a much softer economy than what has been experienced in recent years. Europe is a bit of a mixed bag as the larger economies there were hesitant to lower forecasts in spite of economic woes in the US. But lower them they have – Germany, the region’s lead economy, has had its growth forecasts slashed from 2.75 percent to about 2.0 for 2001. Fearing that slower growth will present risks to the German government’s budget on both the spending and revenue sides, leaders there have been urged to forge ahead with structural reforms. The outlook for the British economy remains generally status quo, but downside risks from the US slowdown may justify another precautionary interest rate cut. British economy is currently being forecast at 2.5% for 2001. Italy, Europe’s third largest economy, is expected to eke out a 2.25% increase as weaker exports cut into last year’s 2.9% increase. Meanwhile, Japan may be at risk for a considerable downward spiral unless it bails out some of the country’s ailing banks to ensure the stability of the financial system. GDP is expected to grow by only 1% this year and 1.1% in 2002. The OECD acknowledges that purging bad loans could lead to rising bankruptcies and unemployment in an economy that is already in deep trouble. However, this being said, the Japanese market continues to consume semiconductors at a reasonable rate in comparison to the other major geographic markets -- likely a result of moving focus from EDP to communications and consumer products which are being produced domestically. Likewise, the larger economies of the Asia Pacific are cutting forecasts as they too are impacted by lower trade prospects. For Taiwan as an example, worsening US economy caused exports to the US (Taiwan’s main export destination) to fall 13.5% in April alone. 10 © 2001 Cahners In-Stat Group – EA0104MF – Confidential And Our Semiconductor Outlook Isn’t Pretty . . . After a 36.8% increase for 2000, the semiconductor industry is caught in the midst of a classic “down” cycle predicated by overheated order rates, inventory build-up at the end equipment level – caused in part by the crash of the dot.com craze – and economic conditions that just plain stink. While the end markets that are fed by the semiconductor industry will likely still eke out some unit growth in 2001, much of this demand will be satisfied out of inventory. Therefore, units, ASPs (Average Selling Prices) and hence dollars for the chip industry will decline in 2001. And, although there were major additions to semiconductor capacity throughout 2000, the current situation had essentially put the brakes on capital spending – the likes of which we’ve not seen before. This will also halt the current overcapacity situation, which is likely to help us return to strong growth rates for 2002 and 2003. In-Stat feels that we’re likely to see this market pop back up next year – again barring unforeseen economic catastrophy. Table 2. Total Semiconductor Americas % Change Market Share Europe % Change Market Share Japan % Change Market share Asia Pacific % Change Market Share Total World Wide % Change Dollars in Millions Worldwide Semiconductor Dollar Shipments (Dollars in Millions) 1997 1998 1999 2000 2001* 2002* 2003* 2004* 2005* CAGR 99-05* $45,852 7.4% 33.4% $41,431 -9.6% 33.0% $47,478 14.6% 31.8% $64,071 34.9% 31.3% $52,841 -17.5% 30.7% $60,312 14.1% 30.6% $70,852 17.5% 30.5% $79,909 12.8% 30.5% $89,640 12.2% 30.3% 11.2% $29,089 5.5% 21.2% $29,405 1.1% 23.4% $31,881 8.4% 21.3% $42,309 32.7% 20.7% $35,231 -16.7% 20.5% $39,794 13.0% 20.2% $45,768 15.0% 19.7% $50,811 11.0% 19.4% $57,446 13.1% 19.4% 10.3% $32,079 -6.1% 23.4% $25,921 -19.2% 20.6% $32,834 26.7% 22.0% $46,749 42.4% 22.9% $40,430 -13.5% 23.5% $46,692 15.5% 23.7% $55,160 18.1% 23.8% $62,923 14.1% 24.0% $70,532 12.1% 23.8% 13.6% $30,184 9.6% 22.0% $28,853 -4.4% 23.0% $37,184 28.9% 24.9% $51,264 37.9% 25.1% $43,651 -14.9% 25.4% $50,357 15.4% 25.5% $60,469 20.1% 26.0% $68,292 12.9% 26.1% $78,248 14.6% 26.4% 13.2% $137,204 $125,610 $149,377 $204,393 $172,153 $197,155 $232,249 $261,935 $295,866 12.1% 4.0% -8.4% 18.9% 36.8% -15.8% 14.5% 17.8% 12.8% 13.0% Source: In-Stat Group © 2001 Cahners In-Stat Group – EA0104MF – Confidential 11 Semiconductor Consumption by Application Segment The semiconductor end-use consumption forecast will experience a shift as communications applications contnue to explode and consume more chips at a faster growth rate than other segments. But, the PC will continue to dominate the end-use arena, as computer applications will still consume the lion’s share of chips at $122.2 billion worth of semiconductors, or 41.3% of all chips, by 2005. Communications is the fastest growing segment at a nine year CAGR of 17% through 2005. Communications will have increased from a 14.8% share in 1995 to 25.6%, or $75.9 billion of the total market by 2005. So, when you gauge chip demand by its final destination, the end result is a dramatic shift in the classic end-use categories – the likes of which had not been witnessed before by this industry. Computing, even though it will retain the largest portion of chip consumption during the forecast period, dramatizes these shifts. In 1995, a full 53% of all chips were manufactured for use in computing applications, which included PCs (desktop and mobile), workstations, etc. By the end of 2000, that number had fallen to 47.5% and by 2005, further erosion will have taken place, landing computer chips (by the purest form of the definition) at 41.3% of the total. Prior to recent history, none of those end-use percentages shifted wildly from year to year – for all intents and purposes, it often took several years to identify a shift, and those shifts were mostly nominal. Traditional computing applications drove the semiconductor industry. Period. That has all changed. Figure 1 epitomizes the shift from pure computing to chips headed for communications, consumer and automotive applications. Table 3 presents In-Stat’s semi end-use consumption forecast. 12 © 2001 Cahners In-Stat Group – EA0104MF – Confidential Figure 1. Total Semi Dollars by Traditional End-Use (% of Total) – 1995 vs. 2000 vs. 2005 60 50 Percent of Total 40 1995 2000 30 2005 20 10 0 Consumer Automotive Computer Industrial Communications Military Source: inSearch Research, Cahners In-Stat Group Table 3. Total Semiconductor Consumer % Change Market Share Automotive % Change Market Share Computer % Change Market Share Communications % Change Market Share Other (Industrial, Military) % Change Market share Total World Wide % Change Dollars in Millions Worldwide Semiconductor Consumption By Major End-Use Category (Dollars in Millions) 1997 1998 1999 2000* 2001* 2002* 2003* 2004* 2005* CAGR 99-04* $22,708 0.0% 16.6% $20,550 -9.5% 16.4% $25,973 26.4% 17.4% $37,940 46.1% 18.6% $31,362 -17.3% 18.2% $36,722 17.1% 18.6% $43,284 17.9% 18.6% $50,973 17.8% 19.5% $58,734 15.2% 19.9% 14.6% $7,354 1.0% 5.4% $7,439 1.2% 5.9% $9,512 27.9% 6.4% $12,025 26.4% 5.9% $9,590 -20.2% 5.6% $11,210 16.9% 5.7% $13,716 22.4% 5.9% $16,648 21.4% 6.4% $18,730 12.5% 6.3% 12.0% $68,598 3.8% 50.0% $62,246 -9.3% 49.6% $71,000 14.1% 47.5% $93,128 31.2% 45.6% $80,043 -14.1% 46.5% $85,962 7.4% 43.6% $101,456 $111,324 18.0% 9.7% 43.7% 42.5% $122,248 9.8% 41.3% 9.5% $24,057 14.9% 17.5% $22,281 -7.4% 17.7% $28,980 30.1% 19.4% $44,823 54.7% 21.9% $38,069 -15.1% 22.1% $47,272 24.2% 24.0% $56,115 18.7% 24.2% $64,099 14.2% 24.5% $75,856 18.3% 25.6% 17.2% $14,487 -3.3% 10.6% $13,094 -9.6% 10.4% $13,912 6.2% 9.3% $16,477 18.4% 8.1% $13,089 -20.6% 7.6% $15,989 22.2% 8.1% $17,678 10.6% 7.6% $18,891 6.9% 7.2% $20,298 7.4% 6.9% 6.3% $137,204 $125,610 $149,377 $204,393 $172,153 $197,155 $232,249 $261,935 4.0% -8.5% 18.9% 36.8% -15.8% 14.5% 17.8% 12.8% © 2001 Cahners In-Stat Group – EA0104MF – Confidential $295,866 11.9% 13.0% Source: In-Stat Group 13 US End-Use Market Orders Allow Us a Glimpse of The Future In developing our forecasts for semiconductors and end-use electronics segments, we are allotted one luxury – we still get order rates, inventories and backlog data in addition to the sales numbers for Total Communications, Office & Computing and Electronic Components in the US market. While there is an awful lot of “stuff” in these numbers, and they are confined to the US market only, we have over the years been able to draw some definite correlations to the worldwide semiconductor and electronics sectors. As an example – the following graph contrasts worldwide semiconductor sales growth year to year against US Electronic Component sales growth. The Electronic Components category is where semiconductor sales are reported to the Department of Commerce for US consumption. Depending on the year, semis account for 40 to 50 percent of this category. You can see clearly that although Electronic Components doesn’t tend to go negative in terms of growth, there is a distinct correlation in the growth trends. Because we still get order rates and other pertinent information for the components category, we are able to come up with very sound short-term forecasts for electronic components AND it allows us substantial insight into what’s going on with the semiconductor market in total – it’s another piece of the pie that gets factored into the overall forecast.. Following Figure 2, please note Figures 3 through 5 that depict InStat’s Booked-to-Billed ratios for Electronic Components, Total Communications and Office & Computing equipment, which contains, among other things, PCs shipped in the US. With all of the information that the government provides us with for these categories, we are able to solidly predict performance 12 months out. Additionally, with the amount of product that these three categories encompass, we are able to get a very solid gauge on the US market in general. 14 © 2001 Cahners In-Stat Group – EA0104MF – Confidential Figure 2. US Electronic Component Growth vs. WW Semi Sales Growth – 1991 through 2000 Annual % Growth 50.0% 40.0% 30.0% 20.0% 10.0% 0.0% 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 -10.0% -20.0% Electronic Components WW Semiconductor Growth Source: Cahners In-Stat Group Figure 3. U.S. Booked to Billed Ratio for Electronic Components – 1995 to 2001 22000 2.00 1.90 20000 US ELECTRONIC COMPONENTS Bookings/ Billings (Three Month Average) 1.80 18000 16000 Book to Bill Ratio 1.60 14000 1.50 BOOK TO BILL RATIO SALES 1.40 1.30 1.20 12000 10000 8000 1.10 6000 1.00 4000 0.90 01 Ja n- -0 0 Ju l 00 Ja n- -9 9 Ju l 99 Ja n- -9 8 Ju l 98 Ja n- -9 7 Ju l 97 Ja n- -9 6 Ju l 96 Ja n- -9 5 2000 Ju l 95 0.80 Ja n- BOOKINGS AND BILLINGS (Millions) 1.70 Source: Cahners In-Stat Group/ Dept. of Commerce Source: Department of Commerce, In-Stat © 2001 Cahners In-Stat Group – EA0104MF – Confidential 15 Figure 4. U.S. Booked-to-Billed Ratio for Total Communications Equipment – 1995 to 2001 2.00 12500 1.90 US- TOTAL COMMUNICATIONS Bookings/ Billings (Three Month Average) 1.80 BOOK TO BILL RATIO SALES ORDERS 1.60 1.50 9500 8500 7500 1.40 6500 1.30 1.20 5500 1.10 4500 1.00 3500 0.90 2500 0.80 1500 Ja n A -95 pr Ju 95 l-9 O 5 ct Ja - 95 n A -96 pr Ju 96 lO 96 ct Ja - 96 nA 97 pr Ju 97 l-9 O 7 ct Ja - 97 nA 98 pr Ju 98 l-9 O 8 ct Ja - 98 n A -99 pr Ju 99 l-9 O 9 ct Ja - 99 n A -00 pr Ju 00 lO 00 ct Ja - 00 n01 Book to Bill Ratio 10500 BOOKINGS AND BILLINGS (Millions) 1.70 11500 Source: Cahners In-Stat Group/ Dept. of Commerce Source: Department of Commerce, In-Stat Figure 5. U.S. Booked-to-Billed Ratio for Office & Computing Equipment – 1995 to 2001 2.00 US- OFFICE AND COMPUTING MACHINES Bookings/Billings (Three Month Average) 1.90 1.80 14500 12500 Book to Bill Ratio 1.60 1.50 10500 1.40 1.30 BOOK TO BILL RATIO 8500 SALES 1.20 ORDERS 1.10 6500 BOOKINGS AND BILLINGS (Millions) 1.70 1.00 1 n0 Ja 0 l-0 0 Ju n0 Ja 9 l-9 9 Ju 8 l-9 8 n9 Ja Ju n9 Ja 7 l-9 7 Ju n9 Ja 6 l-9 6 Ju n9 Ja l-9 5 4500 Ju Ja n9 5 0.90 Source: Cahners In-Stat Group/ Dept. of Commerce Source: Department of Commerce, In-Stat 16 © 2001 Cahners In-Stat Group – EA0104MF – Confidential Communications IC Market: Wired Versus Wireless This section will forecast end-use semiconductor consumption for major sub-segments of the communications market, for both the wired and wireless categories. Values presented in the various figures presented in this report relate to total semiconductor consumption in the communications markets, and not to the value of the individual sub-segments contained therein. Market Drivers Factors driving the communications evolution are numerous, with one common, underlying thread: the ability to communicate with anyone at anytime, in any locale has become the prevalent expectation. While this need to communicate is nothing new, having been around since the invention of the telephone and telegraph, it has reached new heights in importance. Today, this need to communicate has gone from voice-only, to voice and data, and the requirements are accelerating at an unbelievable rate. This is further coupled with the fact that the communications world is rapidly migrating from one dominated by analog technology to digital, with a concurrent shift from wired to wireless access. Pushing the migration from a wired world to a wireless one are factors that include: • The prohibitively high cost of laying new lines, especially an issue in remote regions, such as in parts of Asia, South America and Africa • The fact that the global workforce has become very mobile and expectations with regard to communications are extremely high. • New features and functions being introduced on a continual basis at reasonable cost are making an extremely attractive case for increasing wireless access to information, goods and services. While the market drivers are relatively straightforward, the associated issues and complexities are another matter altogether. These problems range from legal and regulatory to a plethora of standards and protocols that must be met, and are further complicated by rapid system evolution and deployment. While in © 2001 Cahners In-Stat Group – EA0104MF – Confidential 17 the short term many of these market segments remain rather chaotic in terms of technology AND economy, solutions will be developed over time and emerging markets will settle and solidify. In fact, many solutions to standards and protocol complexities will be met via the use of reprogrammability and/or reconfigurability of software and/or hardware. What all this translates to is almost unlimited opportunities for the companies that participate in this market over the longer term. However, opportunity is also associated, in many cases, with risk, but for the majority of companies, long-term success is assured. After all, the majority of the world is not yet connected. Even in the face of worldwide economic turmoil, you can’t put a damper on the big picture view of communications markets. While the wired segment will continue to dominate the communications market in terms of dollar value through the forecast period, it will continually lose ground to its’ wireless brother. Total wired communications semiconductor shipments are forecast to increase from $19,610 million in 1999 to $43,383 million by 2004, translating to a compound annual growth rate of 17.2%. However, over this same period, wired communications will see its’ market share of total communications semiconductor consumption decline from 71.9% in 1999 to 62.8% by the end of the forecast period, 2004. And, in spite of good growth for total wired communications chips, not all of the defined sub-segments will prosper, primarily as a result of the migration from analog- to digital-based communications, and the integration of certain functions, such as facsimile, into higher-level products. The factors previously discussed, will drive the growth in the wireless communications market. In product terms, this growth will be manifested primarily by the cellular terminal and infrastructure segments. To give you an idea as to the growth in these two sub-segments, cellular terminals are forecast to increase by a factor of 3X over the forecast period, while cellular base stations will almost double over the same period. Basestations are being produced with increased capacity – a lower number of basestations will serve a larger number of customers. Wireless communication semiconductor consumption will increase its’ overall share of the communications market, going from 28.1% 18 © 2001 Cahners In-Stat Group – EA0104MF – Confidential in 1999 to 37.2% by the end of the forecast period, 2004. This translates into a compound annual growth rate of 27.4%. Table 4. 1999 Total Wireless 2000 $8,172.00 % Change % of Total Total Wired Total Comm. 2001 2002 2003 2004 CAGR $14,657.00 $12,792.00 $16,403.00 $20,482.00 $23,845.00 79.4% -12.7% 28.2% 24.9% 16.4% 28.10% 32.7% 33.6% 34.7% 36.5% 37.2% $20,808.00 $30,166.00 $25,277.00 $30,869.00 $35,633.00 $40,254.00 45.0% -16.2% 22.1% 15.4% 13.0% % Change % of Total Total Communications Semiconductor Shipments Per WSTS – Wired Versus Wireless (Dollars in Millions) 71.80% 67.3% 66.4% 65.3% 63.5% 62.8% $28,980.00 $44,823.00 $38,069.00 $47,272.00 $56,115.00 $64,099.00 54.7% -15.1% 24.2% 18.7% 14.2% % Change 21.4% 14.6% 17.2% Figure 6. Total Communications IC – Wired vs. Wireless 2004 Wireless Wired 37% 63% By 2004 the Comm IC Market will Have Grown to $64 Billion © 2001 Cahners In-Stat Group – EA0104MF – Confidential 19 Specific Opportunities: Semiconductor Consumption by Segment Networking Equipment The networking equipment market was booming, with manufacturer's revenue growth at 34 percent in 2000 to $49.8 billion. Cahners In-Stat Group believes the following are among the key growth segments in networking equipment. • • Gigabit Switch Routers/Terabit Routers Gigabit Ethernet Products • No-new-wires Home Networking products • • • LAN telephony VoIP PBXs and gateways Storage In 1999, the top four networking markets in ranked order were LAN Packet Switches (predominantly comprising Ethernet and Fast Ethernet Switches, along with the rapidly growing Gigabit Ethernet segment), Routers, Access Concentrators and ATM WAN Switches, with respective shares of 26 percent, 17 percent, 8 percent and 8 percent. 1999 demonstrated the trend toward network value moving into the WAN. With the exception of high-end LAN switches, LAN markets experienced rapid price erosion. By 2000, Fast Ethernet switches, the most popular commoditized segment, dominated the market (70%), so the IC content remained a pretty low portion of the total costs of the system. However, highend systems (GigE, 10GE and L4-7) will pick up during the forecast period, which will cause the IC content to inch up. The move will be incremental due the inertia caused by the sheer size of the Fast Ethernet Market. Routers are currently seeing two areas of growth – SOHO (extreme low end) and high-end (Gigabit and Terabit). IC content in the high-end is a huge portion of costs in these systems where a port can cost tens of thousands of dollars, but this segment is only 40% of the total market. SOHO is making huge strides in volume, but prices are dropping. These divergent trends create a fairly flat line for the overall IC trend as a percent of system costs. 20 © 2001 Cahners In-Stat Group – EA0104MF – Confidential Figure 7. Networking Switches & Semis – 2001 vs. 2004 $32.9 Billion $21.8 Billion Not overly rich in semiconductor value as a percent of total equipment, semis hold pretty steady at about 13% of total. $4.6 Billion $2.7 Billion 2001 2004 Switches Switch ICs Source: In-Stat Figure 8. Networking Hubs & Semis – 2001 vs. 2004 $661 Million Clearly not the opportunity of the decade, hubs and the semis that are consumed by them will continue to erode over the forecast period. $51 Million $33 Million $2.5 Million 2001 2004 Hubs Hub Semis © 2001 Cahners In-Stat Group – EA0104MF – Confidential Source: In-Stat 21 Figure 9. Routers & Semis – 2001 vs. 2004 $26.7 Billion $14.4 Billion Not expected to grow by leaps and bounds by any means, the semiconductor opportunity in routers maintains about a 16% share of total value. $3.6 Billion $1.9 Billion 2001 2004 Routers Router ICs Source: In-Stat Broadband Cable vs. DSL The installed base of Broadband subscribers will reach almost 21 million worldwide by the end of 2001, and is expected to increase to almost 69 million by the end of 2004. This phenomenal growth is a direct result of increasing reliance on the Internet as an information, communications, business and entertainment tool. At the same time, new bandwidth intensive applications are being introduced that make the argument for broadband services very compelling. In many of the earlier markets where broadband services were being deployed, cable and DSL were competing head to head for the same customers. Although In-Stat believes that there will be plenty of customers to go around for both types of broadband access, our research demonstrates that DSL will be the most prominent broadband access technology worldwide during the forecast period. 22 © 2001 Cahners In-Stat Group – EA0104MF – Confidential Figure 10. Worldwide Cable vs. DSL Subscriber Forecast 35000 30000 In Thousands 25000 20000 Cable DSL 15000 10000 5000 0 1999 2000 2001 2002 2003 2004 Source: In-Stat Both services are facing hurdles to deployment, many of them are technical issues. However, the deep pockets of the telco’s and the comparatively lower infrastructure investment required to deliver DSL services, will play a prominent role in who wins the battle for the broadband subscriber. In addition, our comparison includes residential and business subscribers, and copper penetration to the business enables DSL to approach a broader total available market. DSL modems and cable modems are the key customer premise equipment enabling deployment of broadband services. Cahners In-Stat Group expects DSL to enjoy strong growth over the next few years. Broadband DSL services grew substantially on a worldwide basis during 2000 compared to 1999 particularly in Asia with approximately 1,800,000 new DSL subscribers. There were many contributing factors that produced the big broadband surge over the last eighteen months. Cahners In-Stat examines not only the exponential growth of DSL service, but reviews the activity driving the deployment of DSL chips shipped for broadband modems, Integrated Access Devices (IADs) and Digital Subscriber Line © 2001 Cahners In-Stat Group – EA0104MF – Confidential 23 Access Multiplexer) DSLAMS around the world. The correlation between DSL equipment manufactured, both Central Office and CPE, chips shipped for DSL equipment, and the final year-end installed DSL subscriber numbers are all taken into consideration. Our analysis provides the reader with a comparison of the ground covered by DSL implementation to date and a sense of where DSL is going from here. The market for DSLAM units is expected to grow considerably over the forecast period, but not as briskly as subscribers. In the US, a good amount of infrastructure has already been rolled out, waiting for subscribers – so as subscribers continue to come on line, units already in the field will merely be used up. Additionally, service providers have called upon DSLAM manufacturers to increase density on a per unit basis. Port densities will increase, naturally allowing for more subscribers per box, yielding higher growth rates for subscribers vs. DSLAM units. Additionally, Cahners In-Stat Group reviewed the data reflecting the number of chips shipped worldwide during 2000 to DSL equipment manufacturers. The total worldwide chip production of 47 million was slightly more than twice the 21.6 million combined ports of DSLAM, CPE Modem, and IAD equipment shipped. To look at it another way, there were a total of 3.45 million new DSL subscribers worldwide in 2000 compared to 7.2 million CPE ports shipped, making the total CPE ports twice the amount of new subscribers installed during the year. This being the case, there is in all probability a bit of inventory still out there that will cause new orders, and hence shipments, to continue to be soft for at least another quarter. The following graphs depict equipment and semiconductor consumed for XDSL and DSLAM segments. 24 © 2001 Cahners In-Stat Group – EA0104MF – Confidential Figure 11. XDSL Equipment & Semis – 2001 vs. 2004 $3,500 $3.24 Billion $3,000 Semis as a % of xDSL equipment diminish over time as ASPs continually erode In Millions $2,500 $2,000 XDSL Units XDSL ICs $1,500 $1,000 $1.15 Billion $1.10 Billion $673 Million $500 $0 2001 2004 Source: In-Stat Figure 12. Cable Equipment & Semis – 2001 vs. 2004 $1.2 Billion Comprising a very high percentage of the total value, semis are worth about 79% of the value in this segment $1.6 Billion $1.3 Billion $952 Million 2001 Source: In-Stat 2004 Cable Units Cable ICs © 2001 Cahners In-Stat Group – EA0104MF – Confidential 25 Factors impacting the growth of the worldwide cable modem market include: Slow but steady worldwide growth of cable TV subscribers – at an average of about 5% per year through 2004. Most of this worldwide growth will come from China. The high cost of cable broadband data service will also continue to be a factor. As an example, Internet usage in Europe is charged by the hour. Unlike in North America, local calls are not free. However, the cable operator may subsidize some of these costs in the future. Local and National telecommunications preference for telephone line-based broadband service also impacts this market. In some countries, like Italy and Portugal, most of the cable infrastructure is owned by the incumbent telephone company that is a state monopoly. For the most part, these telcos are more interested in pushing other broadband services like DSL. And finally, fragmentation makes an impact. In many countries, especially in Asia and Latin America, cable service is divided between many “mom and pop” cable operators. This fragmentation denies many areas the economies of scale needed to introduce broadband cable data services. This is especially true in Southeast Asia and parts of Latin America. Even so, worldwide cable modem subscriptions will increase from 6.7 million at the end of 2000 to over 26 million by the end of 2004. During the same period, worldwide revenues from cable modem services will increase from $3.1 billion in 2000 to over $10.8 billion by 2004. DSLAM DSLAMs (Digital Subscriber Line Access Multiplexers) are the central office link in the DSL chain. They aggregate a multitude of DSL lines into a multiplexing unit and send the traffic on through the network via either ATM or Frame Relay. DSLAM growth will shadow the growth of DSL modems, as service providers provision DSL services. Several changes will affect the DSLAM market. First, the market is shifting to allow for multiple types of DSL in a single unit. This 26 © 2001 Cahners In-Stat Group – EA0104MF – Confidential means that as providers vary their service offerings, they will not need to purchase separate units. Likewise, these units can be stacked and daisy-chained with existing equipment, meaning an add-on is all that is necessary. The DSLAM market will also benefit from changes in the way a provider’s business is conducted. Rather than simply co-locating in the incumbent’s central office, providers are beginning to locate DSLAMs in basements of hotels, apartment buildings and office spaces. As the unit itself becomes more weather-hardened, it may not be unusual to find a DSLAM on a telephone pole. Expanding the addressable market in this manner means increased subscribers for service providers and good times ahead for manufacturers of equipment. Cable Headends An Entire Cable TV System includes all of the headends that are operated by the Cable TV operating organization. For very large systems, there may be up to a dozen or more headends that are all connected to one Primary Headend. Why does a single Cable TV system have multiple headends? As the industry consolidates, major Cable TV MSOs strive to create clusters of systems that serve major metropolitan areas. As larger Cable TV MSOs buy up the assets of smaller local Cable systems, the main “cluster” system ends up having more than one headend. The Primary Headend is the one that has been designated to be the most important one in the multiple headend system. Usually, the Primary Headend houses the most modern Cable TV plant equipment, along with the Cable TV system's business office. The billing system and subscriber authorization computers are usually also part of the Primary Headend. As can be seen in the data table, Cable headends consume semiconductors at a fairly stable rate. Semiconductor content for equipment in headends (above what has already been accounted for in the form of monitors and receiving boxes) will account for about $300 million by 2004. © 2001 Cahners In-Stat Group – EA0104MF – Confidential 27 Figure 13. DSLAM and Semis – 2001 vs. 2004 $1,400 $1,200 While the market for DSLAM Equipment declines in dollar terms, ASPs per box also decrease even though unit densities increase. Dollar growth for both the equipment and the semis are flat over the forecast period. $1.203 Billion In Millions $1,000 $1.041 Billion $800 DSLAM Equip. DSLAM Semi $600 $532 Million $537 Million $400 $200 Source: In- Stat $0 2001 2004 Figure 14. Cable Headend & Semis – 2001 vs. 2004 $1.176 Billion $1.023 Billion Cable Headend Semis hang in there at about 25% of the total value for both 2001 and 2004 $300 Million $250 Million 2001 2004 Cable Headend 28 CH ICs © 2001 Cahners In-Stat Group – EA0104MF – Confidential Source: In-Stat Telecom Equipment – Switches & Optical While In-Stat has not traditionally had a specific focus upon Central Office Switches in the traditional telecom equipment market, we were able to arrive at a solid estimate of the IC value for switches based upon the number of lines that are currently deployed, and those expected for 2004. In the world of optical networking, we are well ensconced. The Internet Service Provider (ISP)-driven, e-commerce economy, with its hunger for speed and data, has given rise to new network architecture. Gone are the days when a business, be it big or small, will require a phone network for voice applications alone. Optical networks are the solution for the New World business strategy, and Dense Wave Division Multiplexing (DWDM) systems are the required, capacity-enabling applications. DWDM systems transmit multiple light wavelengths simultaneously over the same fiber thread, drastically increasing the number of transmissions that can be carried over the fiber thread. DWDM systems are also protocol-independent and have the ability to carry IP, ATM, SONET, etc., while still offering Quality of Service (QoS). The optical network, and DWDM systems in particular, has found market penetration within the carrier backbone infrastructure and ring architecture environments. The optical network consists of much peripheral and complementary technology. Just like the copper wire based systems of old, optical networks require repeaters/regenerators to strengthen a light signal periodically. A limiting factor in the all-optical network is that of the opticalelectrical interface and the developing photonic switch technology advances. Currently, all network traffic is routed or switched by network processors that read the packet or cell headers in electrical form; therefore, all optical traffic must be converted to its electrical equivalent for the decision process. Photonic switching technology will eventually allow the decision process to take place in light form increasing the speed and efficiency of the network. © 2001 Cahners In-Stat Group – EA0104MF – Confidential 29 At any rate, In-Stat feels that the opportunities for optical component manufacturers are numerous and lucrative and that optical networking, especially DWDM -- since DWDM systems are strong capacity-multipliers, in effect turning each optical fiber into many data channels -- are truly technologies that will enable the bandwidth that we need, moving forward. For 2001, Cahners In-Stat Group expects moderate growth in optical networking systems, with better growth numbers for 2002 through 2004, as long-haul markets begin recovering in the third quarter of 2001. Figure 15. Optical Networking Equipment & Semis – 2001 vs. 2004 $60,000 $50,000 $52 Billion In Milions $40,000 $30,000 $29.8 Billion Semis inch up from 25% of total Equipment value in 2001 to 27% by 2004. Growth won't be extremely overheated, but will be very solid Optical Equip Optical Semis $20,000 $13 Biilion $10,000 $7.45 Billion $0 2001 2004 Source: In-Stat 30 © 2001 Cahners In-Stat Group – EA0104MF – Confidential Cellular Handsets/Terminals There have been a number of factors that have led to the reduction in handset sales levels for 2001. While the current economic situations are certainly impacting many of the world’s electronic sectors, this particular segment may also be suffering from a lack of advanced devices and features (i.e. PDA-functionality, GPRS and UMTS). Much-awaited features have not been delivered in a timely manner, and a growing number of consumers are becoming indifferent to more "me-too" products. Currently, manufacturers appear to be unable (or unwilling) to produce state-of-the-art multifunction, multi-band, data-enabled handsets at the required price and quantity. While growth will continue on in the larger, industrialized nations at a solid, steady pace, growth in the emerging nations, especially Brazil, China and Eastern Europe, is accelerating very rapidly, as significant cellular deployments are made. Cahners In-Stat Group does not expect to see a dramatic shift from 2G to 3G within this forecast period. Today’s 2G based (including 2.5G+) service will actually peak in early 2005, with many 2G systems remaining in use beyond 2010. While the growth rates may not appear sustainable to some, they have been predicated on the ability of wireless to supplant wire-line as the preferred voice media. However, we do not believe that there will be sufficient incentive in this forecast period for wireless to make significant progress in displacing the more traditional data schemes. Having said that, however, we do believe that 2.5G+ service will continue to drive subscriber rates, and in fact, will make the wireless handset the most pervasive method of accessing the Internet within the next two years. The only negative factor that is possibly stalking the industry is the ability to obtain a steady supply of components for advanced handsets. Component manufacturers, especially the semiconductor and display manufacturers, will be hard-pressed to keep up with demand, especially as dual and triple-mode handsets with PDAfunctionality come into vogue. © 2001 Cahners In-Stat Group – EA0104MF – Confidential 31 For component manufacturers, all of these trends will point to more growth, more products, and more opportunities for innovation. Overall semiconductor revenue from handsets will grow at a compounded annual growth rate of 27% through 2004. This translates into a total revenue of nearly $32 billion. But the above trends also indicate that the future will entail some challenges. While past handsets have generally contained roughly the same type of components, future handset content will be much more diverse, opening the way for new players in the market, and making the handset component market a very interesting one, to say the least. Figure 16. Cellular Handsets & Semis – 2001 vs. 2004 $140,000 $120,000 $123 Million In Millions $100,000 But never fear. Long term growth is a given - by 2004 handset semis will comprise 25.8% of equipment value $80,000 $60,000 $67 Million Handset Rev Handset Semi All good things will slow this year as semi $s in handsets decline $40,000 $31.7 Million $20,000 $13.3 Million $0 2001 2004 Source: In-Stat Cellular Base Stations Deployment of Cellular Base Stations will see a nice, steady increase during the forecast period of 2001 through 2004. The infrastructure market is entering a very pivotal time in its history. In one respect, infrastructure expansion is slowing down a bit in the US, Europe and Japan since most major markets have most of the coverage they need. On the other hand, the number of worldwide subscribers continues to grow at a fast clip, infrastructure continues to build in ROW, and the industry continues to anticipate the migration to 3G. 32 © 2001 Cahners In-Stat Group – EA0104MF – Confidential Carriers face tough choices that will have a large impact on their future infrastructure needs. In Europe, GPRS is emerging as the step technology of choice bridging the gap between GSM and WCDMA. In Asia, networks are reaching capacity and the move to 3G is a necessity more then just an option. In the United States, with its three different air interfaces, carriers must not only decide if they will go to 3G, but pick from a variety of migration paths as well. Frequency allocation as well as technology questions remain. Despite some of these uncertainties, the infrastructure market continues to grow, and with it, the base transceiver station (BTS) semiconductor market follows. All types of base transceiver station semiconductor revenue are showing growth over the forecast period, with the exception being PDC. Overall, the CAGR for all types of BTS semiconductors is forecast to be 8.2%. Costs for BTSs range greatly, from under $30K to well over $250K, depending on how many RF channels are supported and what air interface technology or technologies are supported. This price range does not include a base station controller, towers/antennas, shelter, land, and site work. Base transceiver station semiconductor revenue is a multi-billion dollar industry. In 2001, BTS semiconductor revenue was over $4.5 billion. In 2004, this figure will rise to almost $6 billion. Of this, the power amplifier section accounts for the largest revenue, with radio front-end, modulation, demodulation, and signal processing following behind. © 2001 Cahners In-Stat Group – EA0104MF – Confidential 33 Figure 17. Cellular Base Station Semiconductors $5.7 Billion $4.5 Billion Not able to clearly determine an "average" cost of Cellular Basestations because of the wide variance in cost, we based our semi estimates on the number of channels. 2001 2004 BTS Semis Source: In-Stat 34 © 2001 Cahners In-Stat Group – EA0104MF – Confidential And, Finally, the Numbers The following table lays out In-Stat’s estimates and forecasts for key communications market segments. Please note that semiconductor content as a percentage of equipment revenues is expected to increase from a current level of about 20% to over 23% by 2004. This is a substantial increase as, overall, this type of percentage analysis has remained fairly static over a number of years. The percentage accounted for by semiconductors in equipment for 2001 will be smaller than would normally be the case due to rather substantial price attrition in the current climate. Please note that the IC values for Cellular Base Stations and CO Switches were backed out of the total IC value vs. Equipment value comparison because we were unable to arrive at solid total equipment dollar values for those two segments. The costs for these infrastructure pieces can vary greatly, and we felt that adding an estimate would skew the semi $ content percentage. We arrived at the semi $ value for these two segments based on number of lines deployed at the CO level, and number of channels deployed for base stations. The total switch number that appears in the table is comprised of ATM WAN, ATM LAN, LAN packet, and Frame Relay switch types. © 2001 Cahners In-Stat Group – EA0104MF – Confidential 35 Table 5. Dollar Value for Key Comm Segments for Equipment and Semis – 2001 versus 2004 (Dollars in Millions) Cellular 2001 2004 CAGR '01-'04 Handset Revenue $67,150 $122,714 22.3% Handset Semi $ $13,256 $31,678 27.3% Semi % of Equip 19.7% 25.8% BTS Semis $4,507 $5,705 8.2% Broadband 2001 2004 XDSL Equip $1,083 $2,245 CAGR '01-'04 27.5% XDSL Semis $673 $1,154 19.7% Semi % of Equip 62.1% 51.4% Cable Modems $1,205 $1,610 10.1% $952 $1,295 10.8% 79.0% 80.4% Cable Semis Semi % of Equip Other BB 2001 2004 DSLAM Equip $1,203 $1,041 -4.7% DSLAM Semis $532 $537 0.3% Semi % of Equip 44.2% 51.6% Cable Headend $1,023 $1,176 4.8% CH Semis Semi % of Equip $250 24.4% $300 25.5% 6.3% Telecom CAGR '01-'04 2001 2004 Switch Semis $3,904 $4,906 7.9% Optical Equip $29,800 $52,376 20.70% $7,450 25.0% $13,125 25.1% 20.80% Networking 2001 2004 Hubs $661 $51 -57.4% $33 $3 55.0% 5.0% 4.9% $21,889 $32,957 14.6% $2,689 $4,605 19.6% Optical Semis Semi % of Equip Hub Semis Semi % of Equip Switches Switch Semis Semi % of Equip CAGR '01-'04 CAGR '01-'04 12.3% 14.0% $14,400 $26,700 22.9% Router Semis $1,900 $3,600 23.7% Semi % of Equip 13.2% 13.5% Total Networking Equip $36,950 $59,708 17.3% Total Networking Semis Semi % of Equip $4,622 12.5% $8,208 13.7% 21.1% 2001 2004 $138,414 $240,870 20.3% $27,735 $56,297 26.6% 20.0% 23.4% Routers Total Equipment Total Key Segment Equipment Sub-Total Semi Value Semi % of Equip No Equipment Numbers 2001 2004 Telecom Switch Semis $3,904 $4,906 Base Station Semis $4,507 $5,705 2001 2004 $36,146 $66,908 Total Semi Value CAGR '01-'04 CAGR '01-'04 7.9% 8.2% CAGR '01-'04 22.8% Source: Cahners In-Stat Group 36 © 2001 Cahners In-Stat Group – EA0104MF – Confidential Summary – Time to Rethink a Few Things In summary, the opportunities for communications semiconductor suppliers are almost limitless, providing you have the right product, at the right time. Semiconductors in future communications systems or products, will take two forms and in large part, these forms will be very dependent upon their final destination. The infrastructure, in many cases, will continue to require highly integrated, high performance (bandwidth), relatively high-cost solutions. Mass market access devices, whether they be handsets, PDAs or set top boxes will continue to require lower cost solutions. This opens the door to virtually any supplier of semiconductor products, regardless of their position in the technology supply chain. And just when we thought we were hitting up on our physical bandwidth limitations, along comes the commercial viability of optical networking. The vast amount of information that is zipping around today’s world, will find a quicker and more efficient route on the optical pathways being built-out as this is being written. This will become a crucial piece of any semiconductor reporting structure that will effectively allow us to know and understand the worldwide communications infrastructure. It is In-Stat’s belief that, as members of the worldwide semiconductor industry, we have reached a point where we must re-evaluate the manner in which we present data. We will continue, on a macro level, to adhere to the findings of the SIA/WSTS. This is necessary in order to be comparing apples to apples with the rest of the industry when we issue our semiconductor forecasts. We do, however, find it increasingly difficult to present plausible segment forecasts for semiconductors when we are fairly certain that there is a substantial variance between what is being reported at the chip level and what is being reported by end equipment associations. Further exacerbating the problem is the fact that the variance is becoming larger over time. Putting together a “total” communications forecast has indeed been a very insightful exercise for all involved here at In-Stat. We hope that the data contained here is enlightening and helpful in the increasingly complex task of accounting for and forecasting the world’s electronic components. As always, we welcome your input. © 2001 Cahners In-Stat Group – EA0104MF – Confidential 37