SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Technology Stocks : Mattson Technology
MTSN 3.6000.0%May 12 5:00 PM EST

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: Jack Hartmann who wrote (3150)1/10/2001 12:26:44 PM
From: Jack Hartmann   of 3661
 
Longer Term, Semis' New 300 mm Fabs Are Critical
By Jim Seymour
Special to TheStreet.com
1/10/01 12:12 PM ET

The market may be pessimistic about the near-term future for chipmakers, but the management of those companies continues to bet on a bright and busy future. Looking at book-to-bill ratios is the favorite short-term measure of how semis are doing, but watching for commitments to build new " fabs" -- fabrication plants, the places you build chips -- is a better long-term tell.

And by that standard, the big semis are roaring ahead.

Advanced Micro Devices (AMD:NYSE - news - boards), for example, just announced it will build a huge new fab, built around the 300 mm die process, to open in 2004. In an effort to recover part of the humongous cost of fabs -- essentially the most expensive "machines" ever built -- AMD will try to negotiate a share-the-fab/share-the-capital-costs deal with another chipmaker.

AMD needs the flexibility, as well as the production capacity, of another fab badly -- sooner than 2004 if possible, though the lead times on building these "fab"ulously complex buildings, and equipping and testing them, limits AMD's ability to bring more capacity online sooner.

The flexibility of multiple fabs allows chipmakers to customize their facilities, dividing overall production so that product line A is built here, line B there and line C elsewhere. AMD was long the victim of a single-fab mentality, and it hurt: The complexities of running a multiple-product semiconductor plant led to serious product problems, which hurt AMD's reputation and bottom line.

Current AMD facilities in Austin, Texas, and Dresden, Germany, are running near capacity already. And AMD has also announced a shift of production of low-end CPUs from Austin to Dresden, to try to wring more dollars out of its investment.

Chipmakers produce chips on silicon wafers, with as many as possible crammed into the confines of each wafer, or disc. After production, the discs are cut into individual rectangles of silicon, which are mounted in plastic or ceramic "carriers" -- the part most visible to the human eye.

The importance of moving quickly to 300 mm silicon wafers, the next production standard, cannot be overstated. The larger the die, the greater the profits, because the additional costs of using and building on, say, 300 mm wafers, vs. the current industry-standard 200 discs, are relatively modest, while the net final production yield per wafer jumps dramatically. AMD is already a joint venturer with Fujitsu on a 300 mm plant that produces flash memory -- a high-demand, high-return product -- and has, no doubt, been studying the return on that investment, as a guide to further fab design and investment. (See Carolyn Humer's excellent story on the prospects for flash in 2001, on TheStreet.com.)

Larger wafers are also key to getting enough volume for commercial success, especially early on, for new chip designs. The yield of working chips from each wafer is relatively low in early production; chipmakers throw away a lot of bad silicon. With the much larger number of individual chips produced on 300 mm wafers, the higher net yield per wafer delivers commercial quantities much sooner.

In addition to the AMD announcement, Mitsubishi announced Wednesday plans for a new, $1.7 billion fab in Kochi, Japan, due to open in 2003.

Intel (INTC:Nasdaq - news - boards), by the way, has led the way toward 300 mm-wafer plants for CPUs, with a plant now under construction and set to open next year.

None of this will be of much interest to day- and swing-traders, but those looking at the semis for longer-term investments should consider how quickly possible buys are moving into the 300 mm era. The semi business is all about fast paths to market, production flexibility and efficient use of capital; companies that don't move aggressively are history.

Especially with the increasing move toward specialized chips -- such as flash memory and high-speed communications chips -- the production-flexibility advantages of multiple, widely distributed, large-wafer plants are immense

thestreet.com

Looking forward to the CC since MTSN has geared toward the 300mm fab area.

Jack
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext