|
Semicon Europa dope: You can't get away from the numbers |
|
|
|
Oct 10, 2007 at 12:39 AM |
Statistics presented at Tuesday's Semicon Europa 2007 press conference reinforced what chipmaking industry watchers have known for awhile: the fabs are moving to Asia, 300-mm wafers will soon make up the largest capacity category, and the leading semiconductor manufacturers are spending the big bucks on advanced geometry tools.
Still, there were some surprises nestling among the market data.
For instance, a couple of charts brought additional insights into the much-discussed trend toward siting fabs in Asia. Since 2000, Chinese chipmaking capacity has grown at a nearly four-digit pace---approximately 859%--while Southeast Asian capacity has climbed 310%, South Korean 291%, and Taiwan 246%. Yet Europe, the US, and Japan also saw capacity growth, but their respective 91%, 87%, and 77% don't look very impressive compared to the Asian numbers.
As for the data on fab capacity by wafer size and by region, SEMI believes the crossover point---when 300-mm wafer starts will exceed 200-mm starts---will occur sometime in 2008, with the larger substrates garnering 40% of the total (as measured, ironically, by 200-mm equivalents: when will the stats be counted in 300-mm equivalents?). SEMI pres/CEO Stan Myers also cited 300-mm percentage growth numbers of 59% in 2007 and a forecasted 27% in 2008, which would represent a doubling of big-wafer capacity in 24 months (which is no real surprise, given the few dozen new fabs ramping).
The most advanced process node in production---45 nm (that's 0.045 micron for you traditionalists)---has started to push aside 65 nm, in terms of equipment spending. That trend will accelerate in 2008, according to SEMI, with the 45-nm likely to account for more than half of the expenditures. No surprise that foundry power Taiwan spends the most on new gear, while also maintaining the biggest chunk of 300-mm capacity.
But installed base does count for something. Japan, home to more exisiting fabs than any other country, leads the regional pack in wafer-fab materials purchases, with nearly $6 billion expected this year and close to $6.6 billion next year. And who's number two? North America, with just over $5 billion in 2007 and $5.4 billion in 2008. Taiwan trails the top pair with nearly $3.9 billion and $4.2 billion expected for the two years, despite its 300-mm dominance.
The polysilicon sector is one market area where the only current downside is its lack of capacity. Stan uttered the magic words "no downward price pressure" when showing the up-up-up revenue trends of the past few years. With crystalline-silicon solar cell manufacturers accounting for at least half the demand, the global revenues for poly have climbed to about $600 million, more than double the dollars seen in early 2005.
Given the poly supply squeeze (btw, I have some stray silicon that I'm considering putting up for auction on eBay), companies are scrambling to expand their exising facilities or build new poly factories. Announced expansion plans by both old and new suppliers should more than triple capacity over the next couple of years---from under 40,000 metrics tons per year to more than 160,000 metric tons in 2010. Stan said much of the added capacity will come from 15 new players that have thrown their hats into the poly ring, many from China and Russia.
Speaking of silicon, during the panel discussion following the attack of the market stats, Deutsche Solar's pres/CEO Dieter Woditsch (whose SolarWorld affiliate consumes copious quantities of the sandy stuff), noted how fast solar-wafer thicknesses are thinning. He said the substrates were around 330 microns thick two years ago, moved quickly to and past 240 microns to their current 210-micron thickness, with 180-micron-thick slices imminent.
He mentioned these numbers in answering a question about the lack of standards in the PV industry, something which he says will be very difficult to rectify for the time being. When asked what kind of standards might be instituted earliest, the company exec noted that those for thin-film PV may come sooner, but that overall, the first standards will need to determine "what are the parameters that manufacturers will have to have to get good, efficient solar cells." (I suppose it would be useful to have a sturdy standard for conversion efficiency, one of the main metrics used in the PV world....)
As SEMI tries to become a key player in the photovoltaic manufacturing industry, one key pathway to its success lies in the standards arena. Given the eminent Dr. Dieter's comments, don't expect a rash of solar-related standards meetings to hit the docket anytime soon.
|
|
