Some see a slow recovery this year for the chip industry, and for the entire nation for that matter. That still includes me. For nearly a year now, I have predicted the chip business would not be back in full swing until at least the third quarter. That still sounds about right to me.

Capital spending to fall again
hard this year, says Dataquest

Dataquest didn't have much in the way of good news early in January year for the world's semiconductor production equipment industry.

Because there still is a giant glut of capacity, chip makers will cut their capital spending for plants and production equipment by 24% this year, according to the big market research firm. That would be down nearly as much as it was in last year's debacle when this spending fell 29%. That would amount to a drop of more than $10 billion from last year's $44.4 billion.

Dataquest expects spending on just chip production systems to recover in the second half, but that chip gear revenues this year will still end up down 19% from the $25.2 billion spent by chip makers in 2001, according to the market research firm's new forecast. These expenditures fell 29% in 2001, it estimates.

"A macroeconomic recovery and returning electronic equipment demand should finally bring the 'demand-component' of the down cycle under control," says Klaus Dieter Rinnen, who heads Dataquest's semiconductor manufacturing research. "However, overcapacity remains excessive and still demands industry attention," he adds.

Capital spending cuts was rampant throughout the world last year. But foundries and DRAM vendors made the deepest cuts, which resulted in Asia Pacific-based chip makers reducing their capital spending by nearly 47% in 2001, Dataquest says. And Taiwan's chip manufacturers cut their capital spending by more than 50% in 2001 downturn, according to the research firm's latest estimates.

The blood bath, though, wasn't that bad in the U.S., Europe, or Japan. U.S. chip companies cut their capital spending by 21%, Europeans by 26%, and the Japanese by 18%, according to Dataquest.

R&D spending in U.S. this year uncertain,
and at best will only match inflation rate

The annual R&D forecast from Battelle is usually not only interesting, but is a fair indicator of innovation in this country. But this year's numbers gave me a mixed, uneasy feeling. The overall increase over 2001 was expected to be only about 3.5%, or barely matching the estimated inflation rate.

Total U.S. spending for R&D is predicted to hit $285.6 billion this year, according to the Ohio research firm. Three significant factors have confused and complicated the R&D outlook for 2002 and made it difficult to get a handle on where R&D spending is going. They are the new Republican president, the recession, and the impact of the Sept. 11th terrorist attacks.

"The effects of these three events are not simply isolated and cumulative," declares Jules Duga, a Battelle senior researcher and co-author of the report. "The impact of these significant external forces cannot be underestimated... their effects are somewhat more difficult to deal with."

Analyzing the emerging patterns of R&D in previous years could accommodate almost any singular disruption and present a "reasonable picture of the anticipated effects," Duga says. "However, the concurrent triple-whammy experienced over the past few months creates a degree of uncertainty that is higher than usual and adds a layer of complexity to the forecast."

The big guy in R&D continues to be industry. Companies are predicted to spend nearly $195 billion in 2002, which would be a 3.2% increase. The federal government is expected to spend $75.5 billion, a 4.7% rise over 2001, universities and non-profits will carry out $15.4 billion in R&D this year, about 3% more than they did in 2001.

Even before Sept. 11th, the federal government commitment to R&D was starting to show signs of growth due to the proposed expansions of defense and health programs. But the terrorist attacks are causing federal R&D spending to shoot up even higher, according to Battelle, especially in intelligence information and improved methods for the interpretation of such information. Other hot areas, it says, are electronics for computers and telecommunications

While U.S. industry still is spending a significant amount for R&D, Battelle says the rate of increase is "somewhat less than has been experienced in recent years." This is a reflection, it says, of economic factors such as the downturn and relates to cash flow, discounted interest rates, and to the belief that these difficult times will have an adverse effect upon R&D funding. Historically, inflation-adjusted reductions in R&D funding have resulted in periods of retrenchment that lasted over a few years, Duga notes. "This year," he adds, "is significantly different."

Yankee Group sees IT spending
starting to grow again this year . . .

Information technology spending dropped by 1.1% last year, but it should start growing again this year, according to the Yankee Group, which projects an increase of 3.3% this year.

But the following year should get back into gear, according to a new report from the Boston firm. IT spending in 2003 should show "impressive 11.5% growth," it predicts.

Companies invest in technology for two reasons, the report notes: to maintain existing system status quo and to gain a competitive advantage. Status-quo spending grows at a relatively consistent pace and accounts for the vast majority of IT budgets, while competitive advantage spending is "volatile and represents a fraction of the total technology outlay."

"Status-quo technology spending will remain consistent and continue to drive the majority of IT outlays in 2002," predicts Jon Derome, senior analyst at Yankee Group. "As recession-induced risk aversion fades late in 2002 and throughout 2003, spending on innovative new solutions--the IT category hardest hit in 2001--will begin to accelerate," he predicts.

. . . and 60% of companies surveyed
expect to spend more on IT in 2002

Another optimistic forecast on information technology spending came from Gartner and SoundView Technology. They did a survey in November of 1,048 companies from more than 20 industries that spend more than $35 billion annually on information technology. The good news is that 60% of these folks expect to increase their IT spending for 2002, while 40% indicated lower spending.

Most likely to attract new spending this year, according to the survey, are products in security, storage, Web-based applications, PDA, and Web integration services. Low on the priority list are mainframes, contract labor services, and desktop products and services.

Gartner came up with spending predictions for 18 vertical industries. "Petroleum and mining will be the most attractive sector for IT vendors in 2002, while non-durable goods manufacturing will see the largest percentage drop in IT spending year over year." says Al Case, senior vice president.

Petroleum and mining will show an increase in IT spending of 16%, while agribusiness, fishing, and forestry will have the second highest gain at 10% annually, he notes. Showing "minimal" IT spending increases for 2002 will be business services, financial services, media and entertainment, retail, and healthcare industries, Case adds.

Expected to show the sharpest declines in IT spending this year are non-durable goods manufacturing declining 15% and utilities falling 4%, he says.

How Gartner sees 2002 IT spending by industry

Global sales in $ billions

	Change from 2001 (%)
Petroleum, mining, energy	15.9
Agribusiness, fishing, forestry	10.0
Business services	7.8
Media and entertainment	7.6
Distribution and logistics services	7.1
Financial services	5.0
Retail trade	4.4
Healthcare	3.6
Education	3.4
Nonprofit	1.9
R&D	1.7
Durable goods manufacturing	-0.2
Wholesale	-0.4
Transportation	-1.6
Government	-3.6
Utilities	-3.7
All manufacturing	-6.2
Nondurable goods manufacturing	-15.1

Source: Gartner and SoundView survey

Computer platforms losing
clout as semiconductor driver

The biggest driver of chip sales during the past couple of decades has been computing platforms, but chip makers can't count on these revenues any more to drive up their sales.

Market researcher iSuppli is now predicting a 0.6% decline in computing platform revenues to $143 billion from $144 billion last year. These sales cover all kinds of platforms--mobile, handheld, and desktop PCs, workstations, midrange computers, mainframes, supercomputers, and Internet appliances.

But the outlook is a little better for the semiconductors going into these compute platforms. Chip content will grow about 3% in 2002, while chip revenues from compute platforms will grow less than 9% in 2002 to about $55 billion, iSuppli predicts. That ain't bad, but it isn't close to the double-digit percentages that were racked up in past recovery years.

One thing limiting significant growth this year will be average selling prices, which will fall even though unit shipments will be increasing in most segments of the computer platform market. "For example, unit shipments in the personal computer segment are forecasted to grow by nearly 2% for 2002, but fierce market share competition among the leading suppliers will cause ASPs to fall nearly 4%," notes Joe D'Elia, iSuppli analyst. "This will lead to a projected revenue decline of 1.9% for 2002."

Consumer products still power
growing Chinese chip market

Like we've said before, China is still the bright spot of the global semiconductor business--even in this chip recession.

While the IC market isn't expected to show much growth in 2002, China's semiconductor business could post double-digit gains this year, according to Toshio Ohta, who runs Shanghai Hua Hong NEC Electronics, the Shanghai start up that's a joint foundry venture between Japan's NEC and the Chinese government.

He believes China's IC market could grow by about 10% this year. This is higher than others have predicted. Dataquest, for example, estimates this market will rebound in 2002 with 5-to-6% growth.

But Ohta says the demand for chips in China is exploding as the nation becomes a leading manufacturing base for cellular phones, PCs, and other consumer products.

"China is also becoming the world's largest cell-phone market," he tells SBN at the Industry Strategy Symposium in Pebble Beach this week. China's cellular phone market is projected to grow from 89 million subscribers in 2000, to 300 million by 2005, he says. Not bad.

The Chinese executive also says the PC market, unlike the U.S., is still exploding in China. Shipments are expected to grow from 10 million units in 2000 to 14 million in 2001, according to Ohta. "By 2005, the PC market in China will reach 30 million units," he says. Other hot Chinese chip markets are automotive, consumer, and communications, he says.

Advanced Forecasting claims
its index signals slow recovery

Advanced Forecasting is bragging that its IC recovery index has confirmed another turning point for the semiconductor industry. Claiming that its model has accurately predicted peaks and troughs of IC sales for 16 years, the Cupertino research firm says its index "is now pointing in a positive direction."

The IC recovery index, which is based on economic factors affecting the IC industry's growth and recessions, described a severe decline in IC sales throughout 2001. But is latest data clearly signals the bottom of this recession, it claims. "From April 2001, based on the index, we predicted the IC industry would reach bottom during the fourth quarter," says David Crume, marketing head. The current situation resembles July 1998, he notes, when the model "accurately identified the minimum point of the 1998 recession."

Now the index is signaling the beginning of a slow recovery. The economic factors that determine the rate of recovery in 2002 are less favorable than they were during the 1999 recovery, Crume notes.

Standard logic sales collapse,
will lag industry in recovery

I guess you could call standard logic a "mature market." It really got clobbered in the chip recession last year. Worldwide revenues acted like DRAMs, plunging 46% to $1.6 billion from a record high of $3 billion in 200, according to a new study by iSuppli.

The El Segundo, Calif., research firm believes that standard logic will lag the rest of the industry in the recovery. In fact, this market will still won't get back to its record 2000 level in the next four years, reaching only $2.3 billion by 2005.

Even though standard logic is mostly simple gates and functions, it still is widely used in system designs as interfaces and "glue logic" between ASICs, programmable ICs, and other highly integrated devices.

"Prices for standard logic are now stabilizing at their lows and are expected to remain static through the end of the year," predicts analyst Kelli McNatt, who compiled the report. But suppliers have "done a good job in cost reduction and are now making some money at current levels.

"They have also invested in bringing newer devices and packages with higher ASPs average selling prices to market and that will help them through these tough times," McNatt says. "It would be prudent for buyers to negotiate long term pricing agreements during the next few quarters to take advantage of the current situation."

The iSuppli analyst also warns chip users that some suppliers will not develop any more standard logic families. They also shouldn't use TTL logic devices in any new system designs because of the danger that these product lines will be phased out in the next several years. "They should encourage engineering teams to switch to CMOS equivalents to avoid end-of-life problems that might impact continuity of supply," McNatt adds.

Will 2010 Caddy be
a computer on wheels?

In the beginning, the big prediction was that just as Hal ran the spaceship in 2001, A Space Odyssey, a powerful minicomputer would end up running everything on a new car. Automobiles were going to be the next great market, replacing the military business back in the 1970s. Not quite, and no minicomputers.

Nowadays, enterprising market researchers have kept the excitement going for this market by talking about telematics and the ubiquitous microprocessor. Hear this new report from Telematics Research in Minneapolis. By 2010, it figures, each luxury car will use 50-to-70 microprocessors and have 12-to-15 times the computing power of today's personal computers. Now let's see: 50 times 17 million vehicles in the States alone--that's quite an MPU market.

Here's how they'll be used. Most of the MPUs, between 40-to-50 of them, will go into embedded control, replacing existing electronic and mechanical systems. Emerging telematics gear information, communications and entertainment systems will need five-to-10 MPUs, and another five-to-10 of the processors will go into "intelligent" vehicle systems to handle such tasks as collision avoidance, minimizing the impact of crashes, and monitoring auto operation for improved safety and performance.

"Microprocessor-based systems will replace electrical/mechanical/hydraulic systems as 'X-by-wire' systems become prevalent," predicts Egil Juliussen, Telematics analyst. "These intelligent vehicle systems are just emerging, but will see strong growth in the next decade." Hey, I remember you, Egil. Your Texas company was making those huge PC forecasts back in the '80s.

But those early auto market forecasts were not all wet. Some people say that today's vehicles are equipped with 15-to-30 microcontrollers or embedded processors. Last year, they say, global automotive chip sales ran close to $14 billion.

But listen to the latest prediction: "The automobile is in the process of becoming a sophisticated computing center, involving numerous computers and subsystems, all linked via LANs and wireless technologies," says Phil Magney, another analyst at Telematics. Sounds sexy enough to me to sell another big batch of market forecasts.

Venture-backed IPOs edge back,
but '01 still off big time from '00

After almost disappearing in the third quarter, the venture-backed IPO (initial public offering) market turned up in the fourth quarter. While these offerings were still down more than one-third from the year-ago period, the fourth quarter was the largest quarter of 2001 with 14 companies raising $1.48 billion, points out Thomson Financial/Venture Economics and the National Venture Capital Association.

The year ended with a total of 37 venture-backed IPOs raising $3.2 billion, a fraction of 2000's record volume of $21.08 billion but in the same neighborhood as 1998 when $3.6 billion was raised. But 2002 looks like another step up off the floor. "We expect the pace of IPO offerings to increase this year as the economy and the technology industry recover from the two-year economic downturn," says Jesse Reyes, Venture Economics vice president.

IC market for cell-phones:
lousy now, superb in 5 years

Here's the standard weather forecast these days for the handset baseband and RF semiconductor markets with apologies to the National Weather Service. Yesterday's blizzard is still with us and we're still freezing to death. But just wait until next summer--the sun will warm us up.

This prediction, more or less, comes from San Jose's Frost & Sullivan, which has sent us a flash that the chip industry is experiencing a "significant slowing" due to the reduced handset demand, unprecedented inventory correction, and the resulting price drops for baseband and RF integrated circuits. Yup, I'll go along with that.

The problem, as we now only know too well, was that "explosive growth in handset shipments during 2000 generated fears of a component shortage and prompted most OEMs, component producers, and distributors to overstock," notes analyst Veerender Kaul. "Many leading manufacturers were maintaining inventory levels of well over 120 days, compared to normal levels of about 30 to 45 days." This hoarding caused the prices to really shoot up for handset chips.

While this might have helped to boost chip revenues in 2000, it was a disaster for 2001. The sharp increase in inventories caused chip revenues to run to into a "significant contraction" due to a "massive inventory correction."

The market for baseband chips for the handset market now looks like it hit $9.5 billion last year, but Frost & Sullivan has that good old-fashioned optimism. It sees this revenue shooting over $22 billion by 2005.

Chip makers will get help in reaching this record level from third-generation networks that will solve the twin problems of saturation and capacity bottlenecks. 3G nets will provide additional spectrum capacity and new revenues streams, according to Kaul. Launching the 2.5G and 3G services will move the mobile handset from a voice-only (with limited data) device to a voice, data, and video device, he says.

The world handset market for RF chips totaled about $5 billion in 2001 and will "rapidly climb through 2005," according to Frost & Sullivan, which apparently doesn't feel up to making a revenue forecast here.

Kaul cautions that current chips can't support the requirements for 2.5G and 3G systems. He says that RF chip suppliers will have to come up with "significant performance upgrades to deliver the aggressive form-factor shrink, ultra-low-power consumption, and high-speed signal processing required by next-generation standards."

That kind of effort will have to come from chip makers that can provide hardware, software, development tools, and customer support, he says. "This does not mean that suppliers of individual parts will be left out of the marketplace. Through strategic partnerships," he notes, "these suppliers will contribute to the total solutions packaging trend."

Cell-phone didn't hit 400M units last year,
but Yankee predicts 10% growth in 2002

Here's another way to look at the global cellular phone market. Unit sales slipped slightly in 2001, but this year they'll get back on track and grow 10.1% to 435.7 million units, according to the Yankee Group.

And the momentum should keep going for at least another three years, according to a new study by the Boston market researcher. Shipments next year should increase 14.4% to 498.5 million, it predicts, and rise 8.9% to 542.8 million units in 2004. By 2005, global unit sales should hit 596 million

Moderate growth, Yankee Group calls it, but it's still a mammoth number. Fueling this growth will be the deployment of new cellular phone services and applications, it says. Fastest handset unit growth will come in Asia.

Some 395.8 million handsets were shipped in 2001, Yankee Group estimates, somewhat below most forecasts.

Sales of IC making materials
will grow faster than IC sales

I could never figure how demand for semiconductor manufacturing materials could grow faster than the overall chip market, but that's what the big guys are predicting this year. Bare silicon wafer revenues will climb by 9.5% to $5.72 billion this year, according to a new forecast from the SEMI trade group.

Revenues for all wafer fab materials will rebound by 11.3% to $14.14 billion in 2002 after declining 20% in 2001, says SEMI. The trade association also expects worldwide chip packaging materials also to grow faster than the chip market this year-a 10.9% increase to $9.25 bullion. This follows a 29% plunge in sales last year.

The turnaround is clearly shown by monthly wafer shipments, SEMI says. The group's index for monthly wafer shipments to chip makers has August as the lowest point in the industry downturn, which began in the second half of 2000. Wafer shipments have climbed up slowly for three months in a row, says Daniel P. Tracy, SEMI analyst. "November blank wafer shipments are up 13% from the low point in August," he notes.

And this year is just a prelude to what these markets will do in the next few years. Revenues for silicon wafer substrates are expected to surge 15.7% in 2003 to $6.62 billion, followed by another 10% increase to $7.28 billion in 2004, SEMI says. Revenues for other semiconductor material substrates, including gallium-arsenide (GaAs) wafers, are expected to grow 7.5% to $602 million in 2002 followed by 14.3% rise to $688 million in 2003 and a uptick of 7.6% to $740 million in 2004. Another one of those good-looking long-range forecasts.

Digital set-top boxes
get back on fast track

It wasn't just cell-phones that had overproduction and inventory problems last year. Despite fast-rising sales, builders of digital cable set-top boxes overestimated how quickly customers would switch over to digital and soon found themselves with a huge inventory problem.

So these set-top box vendors--Scientific-Atlanta, in particular--sharply curtailed their production runs of digital set-top boxes, comments Allied Business Intelligence

But now these vendors are looking for a rebound as large numbers of customers are now migrating to digital tier service, the market researcher says. As a result, it predicts that the overall market will show significant growth.

Worldwide digital set-top box sales--including cable, satellite and terrestrial platforms--are projected to reach 45 million units this year, up from 36 million in 2001, and 28 million in 2000.

And it won't stop there, Allied says. It expects global sales to reach an amazing 88 million boxes in 2007, with market revenues of $13.9 billion.

Recordable DVD drive might be
today's hottest electronic product

You want to hear growth? I'll give you growth. Recordable DVD drive sales were on a roll this holiday season. Unit sales hit 1.3 million units last year and momentum is building.

Sales will jump an incredible seven times this year to 9.8 million units, according to Jon Peddie, president of Jon Peddie Research. That kind of growth will continue, he says, with an estimated 30 million units shipped in 2004 and more than 50 million in 2005.

Making this happen, he says, are compatibility and lower prices. The DVD Forum's new DVD multi specifications have been released and there are a growing number of affordably priced multifunction recordable DVD drives and recorders now on the market. Drive prices will drop below $400 by the end of 2002, Peddie predicts, with recordable DVD media dropping to less than $3. This buys a write-once 4.7- gigabit, DVD-R disc that can play on nearly every DVD drive and player, he claims.

IBM grabs 3,000 U.S. patents,
leading for ninth straight year

Although I can't prove it, this assumption sounds intuitively right: Companies that account for more patents than their competitors have better R&D. And those people who have better R&D should be coming out with more innovative products that become successful. Right?

Well I think so, and I know darn well that IBM thinks so. For the ninth straight year, IBM was awarded the most U.S. patents and it's proud of that accomplishment. A record 3,411 were issued to Big Blue by the patent office, first time any company has received more than 3,000 U.S. patents in a single year.

IBM was way out in front, topping the next closest company by 75%, it says. IBM, in fact, received more patents last year than the combined total awarded to the 12 largest information technology companies: Hewlett- Packard, Intel, Sun, Microsoft, Compaq, Dell, Apple, EMC, Oracle, EDS, Accenture, and AOL. Impressive.

Last year's patents confirm that innovation is thriving in the R&D labs, says Nick Donofrio, IBM senior vice president. "One of the biggest myths of 2001 was that innovation was dying along with the dot.com bust," he says.

IBM's patents break down to 1,500 patents for infrastructure technologies--software, servers, and storage systems--and 1,200 patents for core component technologies such as microelectronics, storage devices, and displays. More than one-third of the technologies that IBM patented last year already are being incorporated into IBM products and services.

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