But some product sectors are even looking good, real good. Microelectromechanical systems (MEMS), for example. And believe it or not, I've got market researchers predicting that cellphone sales will double in the next five years and that the market for next-generation SONET switches is ready to take off. The PDA market will keep on growing over the next couple of years---although not quite as fast as it has. And finally, it looks like the direct-conversion IC business is really going to boom over the next few years.

Flash market is looking more
like 2001--a really lousy year

Hey, this year is beginning to look like a repeat of 2001-at least for flash memory. iSuppli has just cut its earlier forecast of a 2% drop in global flash sales to a decline of "more than 20%" from the $7.6 billion in flash memory that was sold last year. That would put 2002 results falling at around the 28% decline that flash memory sales fell last year.

The California market research firm says it lowered its forecast because of continued weakness in end-equipment segments that are key users of nonvolatile memory devices.

Wireless handsets, "the principle application market for flash memory," is still in the doldrums as phone upgrade sales to end users continue to lag and new 2.5-G and 3-G phones fail to spark consumer demand," points out Jim Cantore, iSuppli analyst.

But that doesn't mean that unit sales will decline. "With flash manufacturing capacity at an all-time high, the industry is awash with inventory," Cantore points out. Unit shipments will grow 17% this year, he predicts, but "flash memory revenues will continue in free fall as ASPs average selling prices decline more than 32% for the year."

"The precipitous decline in ASPs will begin to moderate going into 2003," Cantore predicts. "We currently are forecasting revenue growth of more than 13% during 2003 despite a projected decline of 8% for ASPs."

The longer out his forecast goes, the better it gets. According to the iSuppli analyst, 2004 "looks like an extremely strong year" with flash memory revenue increasing to more than 26% as ASPs rise more than 2%, thanks to a better balance between supply and demand." I'll try to remember.

Here is your latest
growth market de jour

Ready to hear about another classic growth market that's coming your way? Okay, here's your good news fix--a good pick-me-up for these grim winter days. This time it comes from a traditional source, your friendly trade association.

Check out microelectromechanical systems, a rapidly emerging market that should grow between 75% and 87% over five-year, according to the MEMS Industry Group, a group of 22 U.S. vendors that includes Corning IntelliSense, Honeywell, Intel, and Xerox.

They collectively figure that the MEMS market will grow from a $2-to-$5 billion industry in 2000 to an $8-to-$15 billion business by 2004. They estimate that there are already 1.6 MEMS devices per person in the U.S., and by 2004, that number will grow to nearly 5 per person--a compound annual growth rate of 45%.

MEMS-based chips have a top layer of mechanical devices such as mirrors or fluid sensors and are used to make pressure, temperature, chemical and vibration sensors, light reflectors, and switches, as well as accelerometers for airbags, pacemakers, games, and other products. "MEMS will soon contribute to every aspect of our daily life, from palm-sized high-definition projection displays to grain-of-rice size implantable medical devices," says Ken Gabriel, Carnegie Mellon profession who heads the trade group.

Another hard-to-believe,
flash memory forecast

It turns out market researchers can't even agree on the direction of the flash memory business this year, much less predict the size of that market. Now here's a Monterey, Calif., firm saying that flash will rebound and grow by 7.7% this year and by a huge 36% in 2003.

It's no wonder that the outlook is so foggy. In 2000, the flash market shot up to $11.3 billion, a gigantic 148% jump over 1999 sales and a market record, according to Web-Feet Research, which is a new research firm to me. But last year, it says, sales went down almost as fast as they went up in 2000--declining 30% from the 2000 total.

The modest growth expected this year will be driven by a gradual return in consumer demand, especially as the economy recovers from the 2001 recession, Web-Feet says. The boom cycle returns in 2003, it says, but that only gets the flash market back up to where it was in 2000. But listen to this. During the next three years, flash should grow in the "20%-plus" range, producing 2006 revenues of $22.8 billion. I know this is a roller-coaster IC market, but that prediction is difficult to believe.

The two principal flash types will have different growth patterns, Web-Feet says. The supply of NOR parts is expected to continue to exceed demand over the next few years, just like it has for the past year. On the other hand, NAND parts will move into a "chronic shortage" situation once consumer demand picks up later this year, the research firm predicts. And for the next few years, NAND is expected to continue to be in a chronic shortage because supply will not catch up with demand, it concludes.

By 2006, NOR-based parts are expected to account for 55% of the total flash-memory market, while NAND-oriented designs will hold a 45% share, Web-Feet predicts.

Disk storage systems sales fell hard
last year, and will dip again in 2002

For the first time in four years, the disk storage system market fell last year. And it really fell hard. After years of healthy growth, this global business collapsed, falling 18.2% in 2001--or an astounding $5.7 billion, according to International Data Corp.

"Several converging forces fueled this decline, including the demise of many dot-coms, intense price-competition in the U.S., and recession related spending cutbacks," says Charlotte Rancourt, storage systems research head. Also contributing the revenue fall, she says, "are the efficiencies recently gained from storage and server consolidation."

While the U.S. storage system market is expected to recover in the last half, IDC believes the worldwide market will decline 1.7% in 2002 as the U.S. downturn ripples throughout the world. IDC expects the global market for disk storage systems will grow 5.2% in 2003.

Despite the overall market decline, pockets of growth can be found in some segments of the market. Networked storage systems sales, for example, are expected to grow 8.3% in 2002, reaching almost $8.5 billion in market revenue, Rancourt predicts. "This growth is fueled by the benefits of storage consolidation and the need for efficient management of storage infrastructure resources," she adds.

Danger! Foundry capacity to
outpace even growing market

The silicon foundry market is being touted now as one of the fastest growing, global chip segments over the next five years. And I tend to agree with those forecasts. However, something that some of us may be forgetting is that supply could end up growing much faster than even this accelerating demand.

Not Semico Research, which now expects a massive build-up of fab capacity to upset the delicate supply/demand balance in this market. As a result, this Phoenix market researcher predicts that global foundry capacity will exceed overall wafer demand "at least until 2006." And it will be by a "significant margin," the research firm says.

For the first time ever, the foundry business will experience a prolonged period of excess capacity on average over the next five years, warns Semico analyst Joanne Itow.

"I wouldn't call it a 'capacity glut,'" says Itow, "but the supply side is keeping up with demand in the foundry business. I don't see a major shortage of in terms of capacity for awhile." That could be hard on profit margins.

But typically, the older-generation and mature technologies will have the most over capacity, while foundry capacity will remain tight for leading-edge processes, she says. "There's a lot of demand for wafers at 0.18-micron and smaller right now," Itow says. "The demand has accelerated towards the leading-edge technologies, but we can't see much demand for the older technologies."

What's turned the foundry supply/demand picture upside down is the massive build-up of fabs by both the start ups and the established players. Foundries currently are building 200- and 300-mm wafer fabs at an alarming rate.

While this could have a major impact on profits, the foundry business is still expected to grow rapidly this year. Itow figures it could grow by as much as 35-to-40% this year.

Would you believe cell phone sales
doubling in five years? Me neither

You'd think by now that market forecasters would be more careful about predicting a surging worldwide cellphone market after missing so badly with all the nutty forecasts made last year.

But an optimistic Strategy Analytics now is looking for cellular handset sales to jump nearly 20% this year--from 393 million last year to 469 million units. Not only that, but the Boston research firm sees global sales hitting 870 million by 2006. Like similar forecasts last year, that's a bit too optimistic for me. Other forecasts run a lot lower. Motorola, for example, predicts 420 million handsets will be shipped in 2002.

The 2006 market will be made up differently than current business is. By 2006, 77% of all handsets sold will be replacements, up from 52% last year, according to Strategy Analytics. Despite GSM's overall dominance of the cellular market, the research firm says CDMA will control a disproportionate share of the 3G markets. The CDMA share of global handset volume will reach 25% by 2006, it predicts.

And by 2006, Southeast Asia will account for 40% of the total retail sales, says Strategy Analytics' Phil Kendall. That will be twice the market share of any other region. Western Europe and North America will see their share of handset sales "fall significantly over the next five years from over 50% in 2001 to 36% in 2006, he predicts.

"Inventory buildup was a huge issue" in the global market in the last two years and "some residual is still left to work through in the first half of 2002, especially in Western Europe," says David Kerr, VP at Strategy Analytics. Western European handset sales will remain flat in 2002, and won't start recovering until 2003, he believes.

"GSM/W-CDMA based systems will continue to dominate the worldwide cellular market over the next five years," Kerr predicts. But CDMAOne/CDMA2000 sales "will gain share, passing 200 million units in 2006," he predicts.

No telling when workstations
will regain their momentum

Not only was the global workstation market lousy last year, but it won't start booming again anytime soon. That's the way it looks to Dataquest, which estimates that shipments last year fell 11.2% to 1.4 million units.

That marks the first time ever that the workstation business has experienced a decline in shipments from the previous year, according to Pia Rieppo, analyst for the San Jose research firm. "The market remains fundamentally weak," she says, "although it might have bottomed out in the third quarter after four consecutive declines."

Making the workstation market recovery tough, she says, are the "depressed economy, the lack of key new technology drivers, and cautious buying behavior." Dataquest suspects that there are "still a few quarters to go before the workstation market regains its momentum."

Dell extended its lead as the No. 1 vendor in worldwide workstation shipments, according to Dataquest, as the Texas firm's market share grew to 32.4% See table. Only other Top 5 vendor to show a rise in shipments from 2000 was IBM. Sun Microsystems, Compaq, and Hewlett-Packard all had falling sales--HP's drop was dramatic.

Preliminary figures show the U.S. workstation market hitting a three-year low in unit shipments. While shipments were ramping up in the third quarter, the fourth quarter moved in the opposite direction, according to Dataquest. "We believe U.S. shipments will fall under 160,000 units," says Serena Hsu, workstation analyst for the San Jose research firm.

Next-generation SONET will
quadruple market in 5 years

Pushed by service providers to make SONET/SDH more efficient for data delivery, vendors have come up with next-generation devices that will "revolutionize the way that service providers architect their metro networks," says the Yankee Group.

As a result, the global market for next-generation SONET switches is expected by the market research firm to shoot up as a compound annual growth rate of 34% over the next five years. The Yankee Group forecasts that today's $2.7 billion will hit $11.7 billion in 2006.

"As bandwith demand increases or equipment becomes obsolete in metropolitan networks, service providers will be purchasing next-generation SONET/SDH equipment in order to save initial costs, long-terms expenses, power, and space," says Marian Stasney, Yankee Group analyst. "Next-generation SONET/SDH switches will succeed because they provide a cost-effective solution and give the service provider the flexibility to offer additional services to generate new revenues."

The economics are compelling to make the upgrade, according to the research firm, since capital expenditure savings should be in the 30-to-70% range. Limited live traffic has been carried since 1998 and Yankee Group expects wide-scale adoption to begin in the second quarter with equipment starting to be deployed by mid-year.

Static RAM may not
pull out of its nose dive

It was not only the flash memory market that collapsed last year and still was not expected to turn around this year. Static RAM revenues plunged more than 41% last year as rapidly falling demand from wireless and wired networking OEMs created major oversupply conditions. This collapse significantly depressed both unit shipments and average selling prices, according to iSuppli.

SRAM revenues, which amounted to only $3.8 billion last year, will show continuing weakness in these end-equipment markets, the market researcher projects. Revenues are expected to fall 5.7% to slightly more than $3.6 billion.

"The tough market conditions for SRAM are being compounded by other market trends that actually threaten the viability of the SRAM business going forward," iSuppli concludes. "New technologies like Fujitsu's 'FCRAM'and low power DRAMs are actually displacing SRAM in many traditionally strong application areas," it reports.

At the same time, iSuppli says, "more and more SRAM is being embedded in SoCs, displacing the sockets for discrete memory devices." These two trends will dampen SRAM revenue growth for the next five years and may even "force consolidation within the SRAM supplier base," iSuppli concludes.

Only thing that could put new life into the moribund SRAM market in the short term would be improved wireless handset upgrade rates or faster adoption of the new 2.5G and 3G handsets, iSuppli says. Otherwise, it concludes, "we do not see a supply/demand balance that will stabilize ASPs occurring until 2003." That could bring back double-digit revenue growth to the SRAM market, but SRAM suppliers still won't be able to get back to their 2000 revenue highs by the end of 2006, according to iSuppli.

Outlook for litho tools in '02
is for another down year . . .

I sure get tired of those market forecasts that say while the short term outlook stinks, the long-range forecast is looking great. Yeah, sure. Early on, I was told that marketing types don't take long-range predictions all that seriously because by the time they come due no one will ever remember having read them.

I've just got another example of this kind of forecast-this year's photolithography production tools market. Several analysts tracking this business now believe this year's lithography market looks bleak, but looking into the future, they expect litho gear to rebound in 2003 and 2004.

Worldwide unit shipments of litho tools are now expected to fall by as much as 38% this year--declining from 800 systems shipped last year down to 500-to-550 units in 2002, says Noriko Oki, who tracks this market for Morgan Stanley Dean Witter in Tokyo.

"2001 was a bad year for equipment makers in general," Oki says, while "2002 will be a terrible year for the stepper makers," she says. The lithography-tool market looks weak in most regions, but particularly so in Japan.

Dataquest isn't quite as bearish as the Morgan Stanley analyst. Worldwide shipments of lithography tools are expected to fall by 19-to-20% this year, according to the San Jose market researcher--from 800-to-810 tools shipped in 2001 down to 650 units this year.

. . . but not everyone sees
down year for litho gear

Hey, if you don't like Morgan Stanley's or Dataquest's bearish forecasts for the global lithography systems market this year, I've got another prediction you may like better.

Robert N. Castellano, president of The Information Network, is actually bullish on this year's market. He says that a rebound in the economy coupled with higher average selling prices for tools will result in the lithography market growing--yep, I said growing--by 6.8% to $5.8 billion this year.

He estimates the worldwide lithography market fell 17.9% last year--from $6.7 billion in 2000 to $5.5 billion. Castellano says that Nikon sold the most lithography tools last year with a 42% market share, Canon was second, followed by ASML and Ultratech.

Nikon's hikes lead in litho market,
but ASML may take over in 2002

Nikon increased its market lead in lithography tool sales last year, while Canon surprised industry watchers by passing ASML and moving from third to second place in global sales. But that lineup may not last all that long, according to market researchers.

Japan's Nikon remained on top the heap as the world's largest supplier of lithography tools for IC production, followed by Japan's Canon, Holland's ASML, and Ultratech Stepper from the U.S., according to researchers from Morgan Stanley and VLSI Research.

But ASML is poised to leap ahead of its Japanese rivals and become the world's largest lithography-tool vendor in terms of unit shipments for 2002, according to the two market researchers.

It was a wild year for this market. Worldwide shipments of lithography tools fell 34.7% last year, dropping from 1,232 systems shipped in 2000 to 805 in 2001, the new research report estimates. Sales will go down again this year, it predicts, dropping 24% to 612 units.

Nikon increased its global market share in 2001, with its share climbing from 35% in 2000 to 41%. It delivered 434 litho tools for chip production, but its shipments this year are expected to drop by 24% to 330 tools due to the current IC downturn.

Canon took over the second spot by increasing its market share from 23% to 31% in 2001. Its 2000 sales totaled 284 litho tools. But its shipments this year are expected to fall 12% this year to just 250 units.

ASML's worldwide market share fell from 29% in 2000 to 24% in 2001. Its worldwide tool shipments fell 47% last year, from 362 units in 2000 to 192 in 2001. These totals include shipments from Silicon Valley Group, which was acquired by ASML last year.

In fourth place was Ultratech, whose market share dropped by 6% in 2000 to 4% in 2001. The U.S. company shipped 77 tools in 2000 and only 33 system last year, according to Morgan Stanley and VLSI Research.

ASML is expected to rebound and become the leader in the lithography market this year by grabbing a 33% market share, shipping 204 tools in 2002, a 6% increase over 2001, the researchers estimate. "Based on design wins at Intel, IBM, TI, Infineon, Sony, Dongbu, and SMIC, ASML will gain share and become number one in both dollars and units in 2002 or 2003," predicts Steven Pelayo, chip-equipment analyst for Morgan Stanley.

Nikon is expected to fall to second place with a 31% share this year, shipping 190 tools, a 42.4% fall from 2001. Canon, which is projected to fall to third place with a 29% share in 2002, is expected to ship only 180 tools, a 28% decline. Ultratech's share will jump to 6% this year as it sips 38 tools in 2002.

Despite recession, PDA market
keeps growing, but at slower rate

The weak global economy and the erosion of U.S. consumer confidence still wasn't enough to stop the growth of personal digital assistants (PDAs). Global shipments grew 17% to 8 million units last year. But that did represent a "significant slowdown" in growth rate from the previous year, according to Cahners In-Stat.

That kind of growth should continue this year. "Shipments in 2002 for PDAs will increase about 18% as new manufacturers enter the market and a wave of wireless functionality takes hold in a big way," says Neil Strother, In-Stat analyst. "The future will be all about making the PDA a more ubiquitous device and more useful to both the mobile business user and the on-the-go consumer."

"Key PDA market drivers for the coming years," Strother says, "will be more robust devices with newer and more reliable applications, backed by better wireless data and phone networks; wider use of Bluetooth and other short-range wireless technologies; and lower unit prices.

In-Stat sees a "solid" global long-term market for PDAs, with annual growth through 2005 running in the double-digit range, peaking in 2004 at a 30% rate, then tapering off in 2005 as the market matures. The current market leader is still Palm, followed by Pocket PC and Windows CE.

Shift in cellphone design
will impact IC markets

There's a major shift going on in wireless communications design. Direct conversion ICs are displacing the superheterodyne circuit that's dominated receiver designs for more than 75 years.

Direct conversion receivers will be used in 95% of all GSM cellular handsets by 2005, up from about 40% last year, according to iSuppli's Market Intelligence Services. Direct conversion technology also is making its way into CDMA cellular handsets, WLANs, and digital cordless phones, the research firm says.

This shift means the market for direct conversion ICs will grow from 120 million units worth more than $400 million in 2001 to nearly 660 million units worth more than $2.2 billion by 2005, iSuppli predicts.

"Cost reduction pressures are clearly powering the shift to direct conversion," says Scott Smyser, iSuppli analyst. Direct conversion ICs eliminate the intermediate-frequency stage required in superheterodyne receivers, which provides a direct savings in component cost, he notes. Switching to direct conversion in GSM cellular handsets reduces component costs as much as 30%, he says, and cuts CDMA cellular handsets costs by as much 70%.

But the shift will not be good news to SAW filter suppliers. "Makers of SAW filters, key components in the IF section of a superheterodyne receiver, will see 20% of their market disappear over the next few years," Smyser notes. But SAW filters will continue to be used in the RF front-end section of all direct conversion receivers, so vendors should be able to survive the storm, he adds.

Do rising blank wafer shipments
in 4th quarter signal turnaround?

Are blank silicon wafer shipments a leading indicator for the semiconductor industry? If they are, then the recent semiconductor downturn may be over. Global wafer shipments grew sequentially by nearly 7% in the fourth quarter of 2001, after falling steadily in the current IC downturn, according to the SEMI trade group.

But last year was a real bummer for the blank wafer business, which includes virgin test wafers, epitaxial silicon wafers, polished and non-polished silicon wafers shipped by suppliers to fabs. Annual shipments by wafer area fell 29% and revenues dropped by 31% in 2001, reports SEMI in its year-end analysis.

Wafer shipments last year amounted to 3.94 billion square inches, down from the 5.551 billion square inches shipped the year before. Revenues fell from $7.5 billion to $5.2 billion in 2000.

"Silicon area shipments last year declined to the lowest annual level since 1998," notes Stanley T. Myers, SEMI's CEO. "The silicon industry experienced considerable consolidation and capacity reduction last year," he notes, "and most new investment was made in the area of 300 mm wafers."

(Click here for last month's SBN Outlook Column.)