|
WaferNews is published weekly by PennWell. To request a free sample issue or to receive subscription information, please forward requests to Christine Tourgee, WaferNews, 98 Spit Brook Rd., Nashua, NH 03062; Tel: 603-891-9174; Fax: 603-891-0574.
SEZ Group Exits Wet Bench Business
The SEZ Group has taken steps to refocus its resources on single-wafer wet surface preparation technology. This move includes the closing of its wet bench production facility in Donaueschingen, Germany, which was acquired by the SEZ Group in February 2001.
The second step is the shift of resources to SEZ's core business, patented single-wafer equipment for microchip production. To further facilitate this refinement the SEZ Group reduced its global workforce by 20%, from 750 to 600 employees worldwide. The SEZ Group's staff reduction includes the closing of the Donaueschingen facility but will not affect the company's service, process and sales support.
SEZ said its decision will not negatively affect IP and products developed by its California Research Center (CRC), which was acquired as L-Tech Technologies last year.
--Rachel Robinson, Senior News Editor
(March 19)
Final Manufacturing -- Moving to the Front
If Ultratech Stepper's Art Zafiropoulo's recent prediction that in the next 10 to 12 years more than 50% of all ICs will be bumped, it will no doubt come at the expense of some in the industry. But are there enough indicators to point to a "real" trend in the merging of backend processing with frontend wafer level processing?
The disaggregation that occurred with the fabless model - too many entities involved in the fabless supply chain - makes it difficult to say if there's a trend suggests Rudy Garcia, strategic marketing manager at NPTest. He thinks IDMs have more flexibility because, even with outsourcing, they still control fabs and the packaging process. Still, for large classes of high volume parts, the stacked die approach is more cost-effective. "For example, the process recipes for creating SRAM memory cells are not the same as those used to create Flash memory cells," explains Garcia. "Therefore, it's more cost-effective to stack die than to try to pioneer a new fab process combining both."
A more critical question is who will do the backend processing on top of foundry-produced wafers? "There are some small foundries that can do backend processes on a wafer before dicing, but these are not the standard activities of companies like TSMC or UMC," states Ludo Deferm, VP of business development at IMEC.
A further complication is who's to blame for test failures of systems in packages (SIPs) and wafer level packages (WLPs)? "The WLP house will blame the foundry and the foundry will blame the WLP house," Deferm said.
The issue is clouded by the fact that foundries typically only test for a select set of electrical tests but do not conduct functional testing - the circuits don't work properly until passive devices are integrated at the WLP house.
Small foundries will have their own challenges. "It's difficult to know what the customer wants until they book the business," notes Claudio Loddo, CEO of 1st Silicon. "We haven't encountered requests for bumping yet. Those customers that need high-end packaging have their own in-house solutions or they will go to a high-end assembly and test house." The foundry already engages ASE for customers who need turnkey solutions with customer concurrence.
Griff Resor, president of Resor Associates, notes that when it comes to processors, Intel will be the driver for packaging and goes on to point out that wire bonding has a frequency limitation, which directly affects high-end processors. "At frequencies greater than 1GHz, you can't use wire bonding because the wire loop forms an inductor that delays the signal [RL delay], thereby affecting signal delivery."
Also noteworthy, "Backend people are trying to bring the productivity scaling factors enjoyed by the frontend, such as batch processing, to the back," explains Resor. "And more batch processing will be reached more effectively for low volume chips and low cost small chips by the use of more strip packaging, not wafer scale packaging."
A major challenge with wafer scale packaging is the form factor (pad pitch, number of pads/die, die size) - changes all the time. "The only place I see WSP being practical from an economic standpoint is in large volume," Resor said.
Another challenge that affects progress is testing. "The wafer is not a good shape for testing. The problem with wafer scale packaging is test, burn-in and the temp cycling."
Some of the frontend tool design attributes that are making it to the backend tools include SECS/GEM interfaces, remote diagnostics, use of FOUPs, and ergonomics, according to James Quinn, executive VP and GM of the advanced packaging business unit at SUSS MicroTec. "Customers want automation but designed specifically for packaging applications," continues Quinn. He also sees a definite trend in that high-end IDMs, which use standards for frontend equipment, don't want to see standards thrown out for backend equipment.
The backend challenge can be reframed by recognizing that the industry can't stay on Moore's Law if backend costs don't decrease faster than they have. "The cost of assembly and test has come to dominate the cost of some device types," comments Tim Boyle, VP, CTO, and GM of expansion projects at Electroglas. "Strip testing is one approach that addresses the cost issue in which packaged devices don't have to be singulated before test and which enables final test to be very fast. You're basically treating the lead frame like a wafer, vs. the old way of testing in which individual packages were tested or only a few were tested in an array."
Another trend that Electroglas has seen recently is that almost all of the company's customers test and ship some known good die (KGD), even if it's just a fraction of their business. "We've seen a large demand for testing in the diced wafer format," states Boyle. "When you ship KGD, the backend almost disappears into the frontend."
While most are rather cautious in their assessment of the backend/frontend merge issue, Phil Smith, strategic marketing manager of Teradyne's semiconductor test division, is a major optimist and expects increasing backend/frontend developments in the future. He bases his sentiments on trends he's seen including the fact that, at least for microprocessors, the package is already the limiting factor with respect to increased performance, and, in general, the interconnect itself is an issue for many device types. Two emerging trends are further evidence: high-speed serial data links and optical interconnects. "People are beginning to design wireless cores into their consumer ICs now," says Smith.
Echoing Resor's thoughts on frequency, Smith goes on to say that as frequencies keep going up, wires won't work. "Wires are the least valuable part of the device. Moore's Law is scaling silicon performance faster than we know how to scale interconnect performance. Systems will keep getting cheaper and better and the wires will have to go away -- the issue is in front of us now."
--Debra Vogler, Contributing Editor
(March 12)
Gordon Moore Honored by SIA
The reputation of business titans has suffered greatly in recent months, so the tribute to Intel co-founder and chairman emeritus, Gordon Moore, at SIA's 25th anniversary forecast and award dinner in San Jose on November 6 offered a glimpse of greatness to those too young to remember less cynical times. A number of speakers called Moore a moral man and referred to the unusual combination of being technology trained as well as a businessman.
Moore was humble as he received his Lifetime Achievement Award and chose to show the audience images from past technological achievements in the semiconductor industry and the amazing feats occurring today as the industry continues scaling down circuit features. Moore's belief that things evolve at a faster rate than originally thought was borne out in his retrospective presentation.
As a point of comparison, Moore noted that in 1968, there were 1E9 transistors shipped/year; today, the industry ships 1E18 transistors/year. Plotting the number of transistors shipped per year resulted in a line that showed smooth growth - despite the cyclic and often dramatic dips of industry forecasts. Commenting on the ability of the industry to provide ever-increasing performance at ever-lower costs, Moore called semiconductors a spectacular value to the consumer.
Perhaps Moore's greatest wonderment is for the achievements of photolithography. He showed images of 50nm resist lines that were printed with 193nm light and exclaimed, "We're breaking the laws of physics!"
He went on to say that, even more surprising to him than the shrinks accomplished in the x and y dimensions, is the scaling down of the thickness of the transistor and offered up the example of Intel's tri-gate transistor.
Known as a humanitarian, Gordon and his wife, Betty, started the Gordon and Betty Moore Foundation in November 2000, and while he did not tout this accomplishment during the evening, his concern for others was evident. Asked to comment on issues of concern to him during a pre-event press briefing, Moore acknowledged that it takes an educated populace to take advantage of technology, and he fears that many are being left out of modern society. He is particularly concerned about the dearth of minority and women engineers.
With his particular interest being the education of engineers, it is not surprising that the Gordon and Betty Moore Foundation targets higher education. He views the K-12 education problems in the US to be so pervasive and of such a magnitude that only government is able to have a major impact, noting that a lot of local governments have tried but have been unsuccessful in making huge changes.
--Debra Vogler, Contributing Editor
(March 5)
'Million Dollar Man' Emerging in Flanders
Europe's Flemish research organization IMEC, Leuven, Belgium, kicked off its 2002 Annual Research Review Meeting (ARRM) -- billed as "Seeds for Tomorrow's World" -- with the launching of two programs that provide an intriguing view of future wireless networks and the quality of human life.
The ultimate goal is to produce new types of networks and a series of non-invasive sensors (e.g., distributed MEMS and bio-sensing transducers) that add up to "ambient intelligence" (AI). With AI, computing would be ubiquitous in our lives - seeing, listening, and feeling - and the associated wireless IT proactive but hidden (i.e., no conventional keyboard or menus).
The first applications envisioned for AI include support for athletes, diagnostics, and repair in health care, personal communications, and entertainment.
In athletics, for example, realization of IMEC's goals behind the new programs could change today's common use of a wired heart rate monitor to a body area network (BAN) where sensors for heart function, blood pressure, blood composition, leg and foot position, and force communicate wirelessly to a wearable digital assistant (WDA) that is linked to coaches and a medical team. It is not difficult to envision extension of this technology beyond athletics to hazardous industrial and military situations for humans.
Rudy Lauwereins, IMEC's VP of integrated information and communications systems, says, "The world is evolving into ubiquitous smart environments, where AI will allow electronic access to information and services any time and any place. But making these smart environments a reality requires a personal wireless network, hundreds of electronic devices residing in the background of our daily lives, and devices that are sensitive and adaptive to people." He adds that many of the services envisioned for the future demand high bandwidth and very low power 100µW operation without batteries. Power requirements could be done with an electrostatic generator or polymer based photovoltatics, both being developed at IMEC.
The specific programs announced at IMEC are M4 and Human++:
The former targets developing a multimedia multimode terminal. This is the first step in evolving to a WDA that communicates via wireless distributed transmissions. IMEC engineers envision the WDA fabricated in SoC and SiP technology and capable of IF inputs that see, hear, and feel; IF outputs that speak, show, and stimulate; and RF communications to satellites and various base stations, including GSM, GPRS, UNTS, and WLAN. One WDA/person would provide GPS computing, global connectivity, biometrics input, health monitoring, and ambient control, as well as a host of infotainment options.
The latter targets a prototype BAN capable of networking more than 100 sensors/person. The sensors will be based on a generic platform that includes power supply, a wireless transceiver, network control, and packaging.
As revealed at ARRM 2002, IMEC is already making progress with specific sensors. For example, it has a Ta2O3 gate-dielectric pH sensor that is applicable to glucose measurements. IMEC engineers are also working on a magnetic detection genechip that is capable of detecting biomolecules related to genetic diseases (e.g., cystic fibrosis specific mutations on DNA and RNA).
Other work at IMEC is with specific MEMS that are capable of micron-calorimetry for microphysiometry; a poly-SiGe based bolometer for thermal IR sensing; CMOS integrated FETS and amperometric sensors that provide blood-gas data and a silicon-based impedance sensor that provides affinity based immuno- and geno-sensing.
While it is far-fetched today, Lauwereins does see the eventual potential for humans being fitted with IR vision, ultrasound hearing, vital signs monitoring, and, perhaps an outside chance, even "giant upload capability" -- the capability to handle an individual's daily barrage of data.
--Pieter Burggraaf, Technical Editor
(February 26)
Top Execs Reveal Downturn Strategies
The semiconductor industry may be in the middle of its most severe downturn yet, but the mood was solidly upbeat at the SEMI New England Breakfast Forum panel, "A Strategic Perspective: Managing Through the Cycle," held in Burlington, MA, on Nov. 20.
While none of the panel members, executives from MKS Instruments, Teradyne Inc., ATMI, and Axcelis Technologies, would hazard a serious guess as to when the current doldrums may end, they did share details of the strategies their companies employ for making it through the difficult times.
Moderated by Bob Haavind, editorial director of WaferNews and Solid State Technology, the panel detailed how their companies tackle the challenges of a downturn and direct their resources to guarantee future success. Acquisition and consolidation were two of the top strategies cited as essential to a company's continuing financial health. Other key approaches discussed included an emphasis on R&D, lean operations, and a value-added product policy.
John Bertucci, chairman, CEO, and president of MKS Instruments, was the first panel member to champion acquisition and consolidation as survival tools, describing how his company has broadened its range of markets since the 1970s by adding products surrounding its core gas-monitoring technologies and expanding its role in the supply chain. Judicious acquisition of complementary companies has allowed MKS Instruments to offer more advanced and integrated product lines.
Mary Puma, Axcelis president and CEO, emphasized that a company must be very selective in adding product lines through acquisition, cautioning that no company can be good at everything. ATMI's President Doug Neugold agreed that companies need to focus on key capabilities in order to succeed. Once a product expansion plan is in place, stick to it even during a downturn, he added. ATMI has spent $70 million on high-k materials over the last 10 years, for example.
Neugold also considered consolidation inevitable, especially for small companies in need of a global service and marketing infrastructure. Although acquiring a global reach is expensive, it is absolutely necessary for staying competitive through industry cycles.
Money for such consolidation is tight right now, according to Puma. This is the time, then, to create and nurture partnerships for the future when, hopefully, there will be more cash (and better stock prices) to fund closer relationships.
Puma also underscored the importance of an aggressive R&D spending plan. Last year, Axcelis increased R&D spending by 10%, despite a 46% decline in revenues, introducing six new products/applications. There was little disagreement on this among other panel members. Innovation in R&D, developing new materials, technologies, and services, are all necessary activities if a company is to prosper, according to Neugold.
Since R&D and establishing and maintaining a global infrastructure take a lot of money, the rest of an operation must be run very lean, noted Puma. All processes should be examined for cost and efficiency, and modified or streamlined. Mark Jagiela, semiconductor test business group VP at Teradyne, agreed. He compared the present downturn to 1998, saying that, while volumes are up 40% since then, ASPs have fallen and the cost of labor is 20 to 30% higher. Rather than "racing to the bottom," setting lower prices to sell more products and gain market share, Teradyne has instead addressed labor costs by outsourcing many of its manufacturing operations to low-cost areas in Asia.
Adding value to products is another way to avoid the pitfalls of dropping tool prices in response to a sluggish market. Puma emphasized that it is not just the price of a tool that makes it a good buy; customers recognize and pay for efficient equipment that can lower total costs and improve productivity. Neugold said that ATMI regularly recognizes and rewards its employees for innovative ideas in this area, ensuring that the whole organization recognizes the importance of this piece of the manufacturing process.
Although none of the panelists were making predictions regarding the market after the downturn, Bertucci of MKS Instruments noted high-capacity use in the backend - 90% compared to 60 to 70% capacity utilization in frontend operations. All were enthusiastic about the long-term benefits of SoC technology, since, as Bertucci observed, complexity drives growth in the electronics industry.
Bertucci also reminded the audience that industry forecasts are often inaccurate and should be used with caution in planning for the future. As Jerry Colella, MKS Instruments VP of global operations, states, "The forecast is wrong ... get over it, and get nimble."
Perhaps the most upbeat expression of the four companies' long-term strategies came from Puma, who reminded the audience that a company's focus should always be on the future, even in a downturn. "The good news is there's always an upturn; we just don't know when it will happen." In the meantime, it behooves the company waiting for the next killer app to be prepared.
--Grace Jeromski, Assistant Editor, Solid State Technology
(February 19)
Amkor: Bold Moves, Bright Future?
The year 2001 marked a series of moves for backend firm Amkor, which grew out of its traditional bases in Korea and the Philippines to include new operations in Taiwan, China, and Japan.
The moves may have set up Amkor as a packaging powerhouse in Asia - with its actual success being linked intimately to the economy. Though the planning process was undoubtedly made during the last upturn, analysts note, the actual deals took place during the ongoing downturn. If Amkor survives the current economic conditions, it'll be well poised for the coming upturn, analysts Jan Vardaman of TechSearch International and Jim Walker of Gartner Dataquest agreed.
Companies carrying heavy debt are being abused by the market, noted Vardaman. Amkor is no exception, suggested Walker, with heavy day trading of the stock at between $1.25 and $2.25/share. "I think that the acquisitions, had we not been in a downturn, would have been absolutely brilliant; they would have basically been buying customers," said Vardaman. "But in this economy, it's turned out to be a bit more difficult."
Each move was made for specific (and different) strategic reasons, according to Eric Larson, exec. VP of corporate development. Larson explained to WaferNews that Amkor's origins are in Korea, with Anam Semiconductor. Korea matured and wage rates increased, the Korean won grew expensive, and competitiveness diminished, Larson said. In 1989 Amkor took the opportunity in the Philippines to buy a facility previously owned by a customer. Little movement was made for about a decade, when Amkor took a good look at market trends and charted its future moves.
Taiwan
In the late 1990s, Amkor began considering a move into Taiwan. Amkor was serving mostly IDMs, but the fabless-foundry model was beginning to gain steam.
"That was OK for us, we received a lot of wafers out of those foundries - we serviced a great number of fabless IC companies, most of which started in Silicon Valley," said Larson.
But then the local manufacturing base for motherboards began to grow in Taiwan, and the supply base of product going into the PC industry began to stay on the island.
"We felt that there should be an opportunity for us to participate more in the packaging that was happening on Taiwan," remarked Larson. "Plus, frankly, there were some competitors that were developing in Taiwan.
"Our real strategy was to capture some of the available market for the silicon that stays in Taiwan and, second, we wanted to take it to our competitors."
There were 40 to 50 assembly and test firms in Taiwan and Amkor looked at each as a possible acquisition. There was a mature market on the island, and Amkor wanted to go in immediately as a going concern, rather than as a greenfield.
"It always takes you longer than you think it will to develop a business base," Larson noted. "We wanted to be in there quickly."
They settled on Sampo Semiconductor Corp., acquired it, and bought out a JV Amkor had with Acer and TSMC - Taiwan Semiconductor Technology Corp. - for a total investment of $125 million.
Revenues in 2002 from Taiwan are an estimated $85 million.
China
Also in 2001, Amkor (like everyone else in the sector) turned its gaze across the Taiwan Straits toward China.
"There is a tremendous amount of semiconductor and electronics activity in China, and it will only grow as the trade barriers drop," said Larson.
But Amkor saw a different landscape from Taiwan. Instead of a developed, mature industry, China was a developing market, with a small number of very small players in the assembly and test sector.
"We looked for a greenfield site - this business was going to take a while to develop, and we had a little bit of time," said Larson. "Also, there was no good target to acquire."
They settled on a site in Pudong, Shanghai, with an initial investment of $30 million and anticipated revenues of $5 million this year. While revenues are relatively small, said Larson, Amkor's presence in China has strategic roots.
Several of Amkor's large IDM customers suggested that they wanted China content on some fairly advanced packages - BGA types of packages, Larson said. Second, Amkor viewed a presence in China as part of a basic intercept strategy.
"We clearly recognize the trend that there's going to be large amounts of wafers in China. We wanted to develop a base, and develop the relationships in China so when large foundry efforts begin to come online, we'll be in a good position to get our fair share of the wafers. Companies like Grace, Shangai Belling, SMIC, HHNEC - all these large fab/foundries are up and running or will be in the next 12 to 18 months."
Because China is a smaller operation, with fewer customers, it's being used as a testing ground for MP systems, etc. It's easier there, said Larson, because Amkor doesn't have the legacy systems to deal with.
Japan
Business in Japan had always been moderate, said Larson, as there wasn't much outsourcing, and the vertically integrated companies tended to take care of assembly and test themselves.
"All the silicon manufactured in Japan was staying on the Island," explained Larson.
But as the financial situation in Japan continued to deteriorate, the amount of outsourcing increased. Japanese chip companies were getting pressure from the government to restructure, and Amkor began having conversations with the chipmakers.
Amkor made a decision to enter into a JV with Toshiba - "Basically, a staged buyout of the facility. It operates as a JV during a three-year term, then Amkor ends up owning the facility, we take over the people, etc."
Amkor now owns 60% of the biz, and in January 2004, will buy out the remaining 40%. Investment so far has been $77 million, with an estimated $200 million revenue this year. The deal includes a five-year supply agreement - assured revenue for Amkor.
The deal was structured that way for several reasons. First, Toshiba felt strongly that it wanted to remain involved at some level to assure its people were being treated in a way consistent with what they were accustomed to.
Amkor liked the deal because it gave Toshiba a vested interest in making sure the supply agreement actually happened, and provided a stable, large customer.
"We really wanted a period of time to get to know one another," said Larson.
The deal has put Amkor on the map in Japan as a major outsourcer and supplier of assembly and test, said Larson, and has also increased the off-island business for its facilities in Korea, the Philippines, and Taiwan.
"We see the Japan market as being a $7 to $9 billion assembly and test market," noted Larson. "We want to have our fair share of that."
"I think that's going to be one of the best things that they did," suggested TechSearch's Vardaman. "This is the era of the new Japan, where everyone is going to be outsourcing. Basically, they've secured a lot of business."
Larson said Amkor was looking for more such staged buyouts in Japan. One underway was nixed recently, however, due to continued down economic times. Amkor had planned to buy 67% of one of Fujitsu's assembly and test units, Kyushu Fujitsu Electronics Ltd.
Noted Gartner's Walker: "They've done some good things, they saw an opportunity in Japan, and they targeted a couple of companies to go after to get some of the business. They came up with the JV structure - that's a good move. None of the other packaging companies have done that."
Today and tomorrow
Larson said Amkor believes it's in the right places, geographically, but it continues to look at possible opportunities for acquisitions in Malaysia, Thailand, or other countries in the area.
"We kick the tires every once in a while, but with the Philippines operations, we've got a very good cost structure," he said.
They also keep an eye open in Japan and China, he noted, where there might be an opportunity to pick off a competitor or join up with a domestic firm to get more scale quickly. While there's sometimes talk of putting in operations in Europe or the US, plans never seem to be very serious, he said.
In each move, Amkor took advantage of regional operations by bringing in employees from up-and-running facilities to change the density of equipment, examine layout, invest in new equipment, etc.
"We applied, in very quick order, all of the learning resident in the rest of Amkor, and we did that by putting people there," explained Larson.
These transition teams get the new facilities up to snuff, and then use local management. They cultivate a team, and then keep it there, he said.
Amkor's future depends greatly on the length of the downturn, the analysts suggested.
"They have to be able to ride it out. They've got all the banks to agree to help financially. Theoretically, they should be able to do it," suggested Vardaman. "They should be in a good position - all they have to do is make it through the downturn. But everybody else has to make it through, also."
Walker noted that Amkor's attempt to diversify its business from its telecom- and IDM-heavy customer base appears to have been successful. While the acquisitions seem sound, they may still backfire.
"They might have bitten off more than they can handle in a downturn economy, which is reflected in the stock price," said Walker. "They've had major layoffs, too, and they've aggressively dropped prices in BGAs in the past month or so."
The positive side is that the Japanese market is stabilizing, and the JVs there may produce some real revenue, he added.
Amkor Investments
Taiwan
July 2001: Purchase of Sampo Semiconductor Corp., total ownership of Taiwan Semiconductor Technology Corp. Investment - $125 million. 2002 estimated revenue - $85 million.
Japan
Jan. 2001: three-year JV with Toshiba. Amkor owns 60%, buys out remaining 40% in Jan. '04. Investment - $77 million. 2002 estimated revenue - $200 million.
China
April 2001: Greenfield factory in Pudong Shanghai. Investment - $30 million. 2002 estimated revenue - $5 million.
--Matt Wickenheiser, Contributing Editor
(February 12)
Hope Exists, If You Know Where to Look for It
There's enough bad economic news these days to keep the Grinch tickled pink, but there's also a few positive signs of hope out there, if you know where to look.
Things are unquestionably shaky. Possible war with Iraq looms. Unemployment hit 6% in November. Corporate spending still hasn't picked up. Visibility remains limited.
But still, there are a few bright spots, hinting at some hope for 2003.
"Consumer confidence was up for the first time in eight months. And preliminary retail sales figures came in great for after-Thanksgiving shopping, up 12% over last year," Semico's Joanne Itow told WaferNews. "Wal-Mart had its best-ever sales day."
The superstore chain reported more than $1.43 billion in sales at the domestic Wal-Mart chain for Black Friday (Nov. 29) alone, a good comparison to the $1.25 billion on the day after Thanksgiving 2001. Top-selling categories included the silicon-rich home electronics group, followed by small appliances and toys.
Another sort of beacon of hope is consumer confidence, which, according to the U. of Michigan's latest report, rose to 84.2 in November, from 80.6 in October. The same report also had consumer expectations rising from 92.4 in October to 93.1 in November.
Tracking the consumer remains important, with the dearth of corporate spending - consumer spending now accounts for two-thirds of all economic activity. Consumers are expected to spend roughly $300 billion this year on the holidays. That would be about a 4% increase over the same period last year (acknowledging that post 9/11 was one of the worst holidays for retailers in history).
However, according to one analyst contacted by WaferNews, corporate spending might be right around the corner.
"Historically, there is a good correlation between earnings and spending, this can be quantified if you look at the earnings for the S&P 500 and business-fixed investment data from the BEA (Bureau of Economic Analysis)," said Jay Mellen analyst at RBC Capital Markets in Minneapolis, MN. "If the consensus-forward estimates for the S&P 500 EPS are correct, we should see a nice rebound in corporate spending in 2003."
But he's also seeing some segmented strength today.
"The two areas of end-market strength that we have seen recently are consumer electronics and wireless communication - specifically cell phones, not infrastructure," noted Mellen.
There are a few happy signs on Wall St., including the first major technology initial public offering (IPO) since July 2001, when consultant Accenture Ltd. raised $1.9 billion.
Seagate Technology Holdings, the biggest US maker of disk drives, went public Dec. 11 offering 72.5 million shares at $12/share for a value of at least $870 million.
Those in the know believe Seagate's IPO could help thaw access to capital markets for technology companies that are starving for investment. Investment in the disk-drive sector could reflect an opinion of investors that technology as a whole has seen the bottom and is bouncing up.
In fact, if Seagate's IPO is seen as successful, it could lead other tech firms to try their own IPOs in hopes of raising funds through stocks.
Not to be ignored is the recent news from Intel and Advanced Micro Devices Inc. (AMD), both of which upped their projections for 4Q revenue - a possible sign that the PC sector is recovering from the slump. Intel now expects revenue for 4Q02 to be between $6.8 and $7 billion, about 5 to 8% more than the $6.5 billion it reported in 3Q02. AMD expects 4Q02 sales growth of 35% up from 3Q02, up from a previous projection of 20%. It must be noted, however, that the increases may be a seasonable effect, rather than a significant upturn signal.
In a recent report from the generally conservative Gartner Dataquest, analysts note a number of risks and uncertainties, but also point to some coming causes for relief - specifically for the chip gear sector, for which they forecast 15% growth in 2003.
"Given improvements in supply-side fundamentals, Gartner Dataquest holds increasing optimism for 2003: Solid improvements in end-user electronics demand should rapidly lead to an increased need for new capacity," the analyst house reported. "Our more optimistic outlook for the annual revenue picture is based on the significant number of shells and only partially completed fab projects, combined with short equipment lead times."
Companies these days are unwilling to hold inventory for any amount of time, Gartner noted, and that may spell good news for 1Q03. While 1Q is normally soft from inventory overhang from the holiday season, that likely won't be the case in 2003, Gartner suggested. However, Gartner warned, the industry "should not expect a pop as in 1Q02."
On the fab equipping side, Gartner reported that the chip industry will enter 2003 "significantly under-invested."
"Ample unfinished projects are available. In fact, we estimate that as much as 200 millions of square inches (quarterly capacity) of 'non-equipped' fab capacity exists, which will allow rapid moves on equipment once demand picks up," Gartner said. "So, when we combine all these factors, we believe there will be motivation and location to put new equipment in 2003."
--Matt Wickenheiser, Contributing Editor
(February 5)
| 
Web Exclusives 
Add RSS Feed
