Semiconductor forcast.................................
electronicnews.com
It’s Still in the Chips
By Grant Johnson, Cahners In-Stat Group
Scottsdale, Ariz.—In looking forward to semiconductor industry trends in 2001, Cahners In-Stat Group, based here, remains optimistic about the tremendous opportunities that lie ahead. The big three sectors of the Internet era will continue to be chips, storage and digital pipes. The potential of these three sectors is limitless and we will never have enough until we are physically being beamed around the planet. In-Stat Group will look at recent signs of a cooling economy and slowing end-use markets as well as other factors in presenting an outlook for the semiconductor industry in 2001.
Economy to Set Semiconductor Stage
Grant Johnson is an industry analyst in the Semiconductor Services Group at Cahners In-Stat Group in Scottsdale, Ariz. Johnson specializes in product services.
We must watch the global economy closely, as industry downturns, with the exception of the mid-1990s, have almost always coincided with a slowing economy. The U.S. economy is the true global linchpin and growth has slowed down to 2.4 percent in the third quarter compared to the "heated" rate of 5.6 percent in the second quarter. The pace of economic growth is slowing significantly with some economists beginning to discuss the possibility of recession in 2001.
Higher interest rates are inhibiting consumer and corporate spending and raising the costs of doing business. Businesses may be reaching the saturation point having invested so much in information technology over the past decade. The fear is that investors may flee the U.S. market if domestic growth begins to falter. A sliding stock market and pullout of investors in the United States would force the Fed to keep interest rates high as the dollar falls. High interest rates could make a slowing economy even weaker. However, most likely, the U.S. dollar will hold its value due to the lack of attractive investments on the international scene as worldwide economies are also beginning to sputter.
In the Americas market, macroeconomic factors could impact the semiconductor market as gross domestic product (GDP) growth rates are forecast to fall from 4.9 percent in 2000 to 3.2 percent in 2001. The European economy will experience a similar slowdown in 2001, with GDP sliding from 3.1 percent to 2.5 percent. Japan will post a meager 1.5 percent GDP growth rate in 2000 and should remain relatively flat in 2001. The Asia/Pacific market remains home to some of the fastest-growing markets including China and Malaysia and will post GDP growth of 6.8 percent in 2000.
However, East Asian economies are showing signs of slowing, following their "v-shaped recovery" of the past two years. Asian economies will slow down in 2001 by a couple of percentage points but a repeat of the financial crisis is not likely considering two important changes. First, Asian short-term debt levels are more manageable and, second, foreign-exchange reserves are much higher.
Slower Semiconductor Growth, but Still Very Solid
Happy days are here again, at least for the vast majority of companies playing in the semiconductor and electronics markets. In-Stat Group projects that worldwide semiconductor revenue will increase from $149 billion in 1999 to $317 billion by 2004. This translates to five-year compound annual growth rate (CAGR) of 16.2 percent. The lion's share of growth this year will be seen in the Asia/Pacific market, followed by a strong resurgence from the Japanese market. Market share from Asia/Pacific will continue upward throughout the forecast period, making the region the largest consumer by 2004. However, in addition to a rather continual flow of capacity coming online over the next several quarters, quarterly device revenue growth rates required to sustain 40-plus-percent growth in 2001 would have to be quite fantastic.
The semiconductor end-use consumption forecast will experience a shift as communications explode and consume more chips at a faster rate than other segments. The PC will continue to dominate the end-use arena, as computers will consume $142 billion worth of semiconductors by 2004. Computer end-use applications will account for 44.8 percent of all chips produced by 2004. Communications is the fastest-growing segment at a nine-year CAGR of 13.9 percent by 2004. Communications will have increased from a 14.8 percent share in 1995 to 21.9 percent or $69.1 billion of the total market by 2004.
Market factors driving the communications evolution are numerous, with the common underlying thread of being able to communicate with anyone at any time. The communications world is also rapidly migrating from one dominated by analog technology to digital, with a concurrent shift from wired to wireless access. Pushing the wireless migrations are factors including the prohibitively high cost of laying new lines in remote areas like Asia, South America and Africa, a highly mobile workforce and new features and functions being continually introduced at reasonable costs. Opportunities in the communications market are always accompanied by great risk; however, long-term success is becoming more likely in lieu of the fact that the majority of the world is not yet connected.
The networking equipment market is also booming with expected manufacturers' revenue growth of 34 percent to $49.8 billion this year. The installed base of broadband subscribers is forecast to reach almost 9 million by the end of 2000. This phenomenal growth is a direct result of increasing reliance on the Internet as an information, communications, business and entertainment tool. New broadband-intensive applications are being introduced to make the argument for broadband services more compelling. DSL will become the most prominent broadband access technology worldwide during the forecast period. Average DSL subscriber growth will rise at 67 percent worldwide by 2004. In-Stat Group also believes the opportunities for optical component manufacturers are numerous and lucrative and that optical networking, especially dense wave digital multiplexing, is truly technology that will enable the bandwidth we need moving forward.
The cellular handset market is forecast to reach $69.2 billion in 2000. Cellular/ PCS handsets represent one of the fastest-growing segments in the entire electronics industry. The market for ICs in handsets will be tremendous as 1 billion handsets will be shipped annually by 2003. Some trends in the handset market are that units are getting smaller and lighter with increasing functionality. Cellular handsets will require more innovative system-on-a-chip, new packaging, increasing memory, greater processing power and greater connectivity options like Bluetooth. Future semiconductor growth also depends on other markets continuing to take off including home networks, data encryption ICs and biometrics.
Fab Industry Very Active Despite Slow Q3
While fab activity continues to be solid in 2000, the third quarter saw both fab announcements and groundbreakings decline from the previous quarter. The level of fab projects that started construction and equipping was the highest ever in the industry's history in the first half of 2000. Therefore, a slowing in the third quarter was not as shocking. The value of new greenfield announcements in the third quarter was 33 percent lower than the second quarter but still at historically high levels. In fact, the cumulative value of fab announcements to the end of September was almost three times higher than a year earlier. Year-to-date, the industry has started construction on $50 billion worth of new fabs. By year-end, the value of new fab starts will total $60 billion.
Between third quarter 2000 and third quarter 2001, In-Stat Group expects 38 new fab projects to get underway, valued at more than $27 billion. During the same time frame, we expect 36 new fabs to open their doors. We can expect the industry to add capacity at a more moderate rate in 2001 but fab activity will continue to bustle.
Taming the Boom-to-Bust Beast
The boom-to-bust cycle will continue but its volatility will be tamed. Chip vendors will always chase profits during boom times and overbuilding is inevitable unless major industry transformations occur. We have taken five years to return to the high growth rates of 1995, but due to excess capacity coming online from astronomical capital expenditures in 2000, the reality of 300mm wafers and slowing end-use demand, the market euphoria will lose some luster in 2001.
Some people still insist we have outgrown the boom to bust again with the burgeoning Internet- and communications-related applications providing so many IC sockets. In-Stat Group has heard the same story before. In fact, we also recall the story about the $300 billion industry and 300mm-dominated wafer market by 2000. The reality is that analysts have missed many critical calls, including the Asian financial crisis, but one call that In-Stat and others do agree on is that the semiconductor cycle is here to stay. Fortunately, the new version of the beast will be tamer.
Following the 1993-1995 boom, there was a major bust, and the good news from this boom is that the bust may not be as drastic as past ones. In-Stat Group thinks that there are several factors that will begin to turn the historic peaks and valleys of the semiconductor market into rolling hills. While some of the trends are not new, they do seem to finally be taking shape during this current upswing cycle.
Consolidation —As with the consolidation in the DRAM market, which has seen a shift from 15 major players in 1995 to four major players in 2000, the rest of the industry will see more consolidation and joint venturing to help absorb the enormous costs of adding additional fab capacity. Fewer companies will be able to afford to build a multibillion-fab facility and this will once again limit the number of independent players adding capacity. Most chipmakers will either seek partners or look to outsource manufacturing. New start-ups will be sold more quickly to larger companies needing to extend the competitiveness of their offerings.
E-supply chain management —Another exciting inhibitor to chip volatility will be e-supply chain management. Outsourcing component management to an e-supply third-party vendor will become more prevalent, although some integrated device manufacturers (IDMs) and OEMs will continue to develop their own in-house systems. Either way, putting the supply chain on the Internet will bring semiconductor materials vendors, IC designers, manufacturers, distributors and OEMs together in real-time, providing better information and drastically reducing costs at every link of the semiconductor supply chain. Historically, communications breakdowns in the semiconductor supply chain disrupted production, leading to inventory build ups or shortages and missed profit opportunities. E-supply management puts the different rungs on the same ladder and provides better tools for gauging demand. Customers are also demanding fewer suppliers to provide more products and services globally.
Diversified chip demand —Geographic opportunities and new products will provide more target markets for chipmakers and help insulate the market from end-use slowdowns like we are experiencing in the PC market. The next wave of chip consumption will shift more toward the global rollout of an advanced communications infrastructure. We are no longer riding solely on the wings of the PC industry. Semiconductors will also become increasingly more pervasive in emerging markets like China and Latin America.
Outsourcing —The dedicated fabs offer "fabless" companies and IDMs the state-of-the-art fab capacity, serving a broader range of customers and market segments. This flexibility will inevitably lead to a softening of the extreme semi market volatility due to the concentration of productive capacity decisions under fewer corporate boards. The ongoing disintegration of the traditional vertical semiconductor company will continue with pure-play foundries adapting advanced value-added services like IC design and packaging. Today, foundries produce roughly 15 percent of the total $160 billion semiconductor industry, with Taiwan's share of the foundry market surpassing 65 percent.
The Fabless Semiconductor Association (FSA) targets 2012 as the year when foundries will produce 50 percent of all ICs. With more than 500 fabless companies, the dedicated fab has eliminated the fab overhead for IC designers and unleashed unbridled technology. We can also forecast geographic diversity in foundry capacity as the Taiwan earthquake and Indonesian fires are testament to the supply vulnerability of relying so heavily on one country or region.
"Fab-Lite-like" models —More IDMs are also choosing to convert some trailing-edge fabs to foundry servicing rather than paying the high cost of upgrading. Some fabless companies are just deciding to open their own manufacturing facilities or joint venturing with IDMs to build manufacturing fabs. Despite these innovative strategies, most fabless chipmakers will continue to outsource due to costs and the risk of venturing into uncharted areas of expertise.
Another trend in outsourcing is the move to "fab-lite" models, representing those IDMs that are gradually winding down manufacturing operations and becoming fully committed to foundries. We can also expect fabless chipmakers and IDMs to commit more to the foundry relationship in the means of process development and technology transfer.
E-diagnostics in chip equipment —Service and support will also be enhanced as e-diagnostics on fab tools and machinery will reduce resource expenditures, shrink repair time, anticipate problems before they occur and increase output. In the perfect semiconductor world, the Internet will allow chip production to be effectively managed with the click of a mouse.
Real data on a regular basis will also permit chip and equipment makers to better forecast industry volatility and eliminate costly downtime. Ideally, fab managers will be able to tap into any of their worldwide fabs from a single, centralized site. Problems will be diagnosed earlier in the cycle with computer-aided design, yield and test-and-failure analysis all managed online.
Synchronizing manufacturing processes —The cost of ownership has divided the equipment and device makers as neither side wants to incur costs of implementing costly supply management software, at least until generational roadmaps on technology such as 300mm is genuinely in synch. Today, more and more equipment makers are crossing this divide. They are installing the necessary management software in advanced chip equipment to take advantage of the inevitable shift toward a service rather than technology-based equipment industry.
Equipment makers will strive to reduce the number of tools required by chipmakers. Next-generation equipment will have to combine functionalities while simultaneously raising throughput levels. Equipment suppliers will move beyond just technical support to provide an economic advantage to customers. The IC device makers and semiconductor equipment industry must become more collaborative to ensure future success.
300mm Making a Strong Play
Moving to 300mm wafers results in higher yields, higher volumes and higher quality. Capital equipment bellwether,Applied Materials Inc. reports the bulk of its sales were driven by 0.13-micron and 200mm tools, although a significant number of its sales are beginning to come from 300mm as well as copper and low-k materials sales. Unlike previous years, equipment purchases in 2001 will be both technology- and capacity-driven as the industry transitions to a larger wafer size and new materials and processes.
One 300mm fab that is already operational is Semiconductor 300 in Dresden, Germany, while 15 300mm fabs worth more than $23 billion had started construction by September of 2000. The 300mm fabs are spread pretty evenly between Asia/Pacific, North America and Europe.
By the end of 2001, we can expect an additional 11 300mm fabs to begin construction at a total value of $14.2 billion with equipment spending on 300mm fabs reaching $4 billion a quarter. Measured in terms of value, there is more new 300mm fab activity in 2000 than new 200mm fab activity. While the current value of 300mm equipment being installed is small, by the first quarter of 2001, the current crop of 300mm fabs will begin installing equipment.
Is a Soft Landing Inevitable?
The troubling signs are becoming more apparent with unspectacular earnings reports, slowing PC sales in Europe and now the rest of the world as well as slowing mobile phone sales. The stock market has stumbled and the global economy is beginning to teeter.
The unbridled euphoria in the first half of the year has given way to cautious concern in the semiconductor industry. A decline in the three-month moving averages for new fab announcements and new greenfield projects suggest to us that the coming year will not see the growth rates in fab building and capital spending we have enjoyed in 2000. Although capital spending will slow in 2001, most chipmakers will continue adding capacity. Intel continues to build and equip its new fabs while IBM recently announced a $2.5 billion 300mm fab project to begin immediately.
This raises concerns that too much capacity continues coming on-stream in spite of recent signs of trouble and this could end the current boom sooner than expected. As a result of recent developments, In-Stat Group is taking heed and revising its worldwide semiconductor forecast downward to 17.4 percent in 2001. Such growth is hardly indicative of bad times in the chip industry in 2001. |
