Errr, before you get too excited about those Chinese contracts, you may want to look at the global supply lines to get an idea of why the financing and network infrastructure decisions are being made a certain way. Naturally, the network infrastructure decisions are reinforcing the prescient moves made by Nokia in the global standards process........and creating a powerful undertow that will most probably drown the most gullible investors, the weakest CDMAOne operators as well as the most shrill technologists before it is all over.
The average wireless handset has about 500 components with a semiconductor content of about 30-35% (expected to exceed 50% in 3G). Nokia sold 79 million handsets last year -- with handset replacement accounting for 40%. It is expected to sell around 100 million handsets this year -- with handset replacement accounting for over 50%. Those kinds of numbers create the kind of manufacturing visibility that allows different parts of the supply line to confidently add capacity. That kind of supply line leverage influences carriers' decisions on what technology to adopt. Handset delivery tends to lag the network infrastructure delivery by about a year.
The transition from 2G TDMA/GSM to 3G WCDMA is clearly being planned as evolutionary, partly because the supply lines have to catch up. Every year, Nokia improves the options available to global carriers at increasingly lower costs to itself because of the increasing modularity of its designs. The increase in the annual handset replacement cycle will only give Nokia more options in scheduling new technology improvements all throughout the transition especially when done in conjunction with carriers organizing themselves into buying groups (GPRS operators, EDGE operators, WCDMA operators). All that translates to more clout in shaping 4G because it typically takes 10 years for a new standard to go through the global process.
Chip industry gurus upgrade outlook amid explosive growth
By Mark Hachman and Bolaji Ojo
Electronic Buyers' News
(06/08/00, 01:40:13 PM EST)
The semiconductor industry is so hot even its top association and research firms can't pin down the market's annual growth rate.
In the last weeks, at least four research groups have raised their estimates for the industry by as much as 12% as exploding demand for chips obliterated earlier projections.
The Semiconductor Industry Association, the leading organization representing chip makers, now sees total 2000 revenue for the industry rising 31%, to $195 billion, from a prior estimate of $174 billion. Additional double-digit growth in the next three years will push the industry's annual revenue to $312 billion by 2003, the SIA said.
"We're seeing unprecedented user demand for communications of all kinds-basic data, and, of course, Internet access," said John Dickson, chief executive of Lucent Technologies Inc.'s Microelectronics and Communications Technologies Unit. "This burgeoning customer demand is a major factor changing the paradigm in the world of semiconductors."
In other words, the industry should expect additional upward revision of current estimates before the end of the year. In May, Dataquest, IC Insights, and Merrill Lynch jacked up their revenue growth projections, citing strong demand across all sectors.
While communications chips have led the recent upturn, optoelectronics has "become the new 'belle of the ball' with strong growth driven by exploding demand for bandwidth," Dickson said.
It's not only strong demand that's driving the upswing, however. Average selling prices (ASP) are on the increase throughout the industry even as unit shipments spiked higher in certain segments, analysts said.
"At the beginning of the year, we set a growth estimate for 2000 of 24.7% -- we now believe that estimate is significantly too low," said Merrill Lynch, in a recent research report. "We're now raising our estimate to 31.5%. That estimate is based on forecast unit growth of 25%, with the remaining improvement in dollar revenues coming from ASP."
The double-digit increase may be the only thing that the research firms agree on, however. Their actual revenue projections differ widely-by as much as $28 billion in the case of Dataquest, IC Insights, and the SIA. While Dataquest estimates the chip industry's revenue will rise to $222 billion this year, IC Insights puts it at approximately $194 million.
Meanwhile, in its midyear forecast issued last week, the SIA predicted a continued rosy future for the chip industry until late 2002, when capacity increases could drag the memory industry down yet again.
The SIA predicted chip sales will grow 25%, to $244 billion in 2001 and see a 14% expansion in 2002, to $279 billion, before slowing to 12% in 2003 when sales should crest at $312 billion.
Sales across all product segments will remain strong, including optoelectronics, which the SIA tracked for the first time at the behest of Lucent's Dickson. Sales of optoelectronic components should grow about 32%, to $7.6 billion this year and rise steadily to about $15 billion in 2003.
As George Scalise, SIA's president, and other semiconductor executives have noted before, the latest revenue uptick is being driven by communications, not the PC.
"When I visit the PC companies, there isn't one of them that isn't trying to reconstruct the PC as a communications platform," Dickson said. He also believes the world market is being tied more closely together by communications infrastructure, minimizing the economic separation between geographical markets.
This development is driving up sales in other parts of the globe too, the SIA said. For instance, chip sales in the Americas during 2000 will be up about 24%, to $59 billion. Sales in Europe will rise approximately 32%, to $42 billion, while Japan's sales will increase 31%, to $43 billion.
The Asia-Pacific region, excluding Japan, is expected to lead global chip revenue growth with a 38% spurt, to $51 billion. The SIA predicts the Asia-Pacific segment will be the fastest-growing sector from now until 2003.
In its predictions, the SIA has forecast steady growth for most logic devices, including microprocessors, microcontrollers, and the hazy system-on-a-chip segment that the SIA calls "MOS Logic."
The growth rate of microprocessors will steadily decline, however, sinking to a 5% growth pace, to $42 billion, in 2003 after jumping 16%, to $32 billion in 2000.
Even passive components, driven by shortages, should enjoy a healthy 22% increase, to $16 billion, during 2000 and a 15% jump, to $19 billion, next year.
Analog products showed the sharpest expected increases, 35% during 2000, to $30 billion and continue to grow at a healthy rate through 2003, when sales of $48 billion are forecast.
As expected, the most volatile segment will continue to be memory, including DRAM and especially flash memory. While the DRAM industry is expected to grow 42%, to $29 billion in 2000, and jump 44% to $42 billion in 2001, a precipitous drop should occur at the end of 2002.
DRAM sales growth for 2002 should be a modest 5%, with revenue of $44 billion, largely a carry-over from the healthy DRAM market of 2001.
Flash memory will also be affected. While an electrifying 116% growth rate may occur in 2000, bringing the total market up to $9.9 billion, sales will dramatically tail off beginning in 2002, the SIA said. At that time, flash sales growth should be only 9%, bringing total sales to over $14 billion.
ebnews.com |
