Silicon Investor (SI) -- The First Internet Community

STOCKTALK

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.

Politics : Formerly About Applied Materials -- Ignore unavailable to you. Want to Upgrade?

To: 16yearcycle who wrote (14980)	1/22/1998 10:25:00 PM
From: SMALL FRY	Respond to of 70976

Eugene, Re: 5::1 DOW/NAZ You're a genius! What a concept...you won't mind if I borrow your strategy do you? I'm trying to build a model to "get rich quick" and I'll incorporate your idea and see what happens. Thanks, SF PS: Did you just move to Redmond? How d'ya like it so far?

To: 16yearcycle who wrote (14980)	1/22/1998 10:31:00 PM
From: davesd	Read Replies (2) \| Respond to of 70976

This is about MU, but it's got some stuff about FAB pushouts...be sure to read the fab pushouts in the middle of the post if you don't read the whole post. 8:59am EST 21-Jan-98 Credit Suisse First Boston (Geraghty, Jack (212) 325-308 AM Call: MU: We Are Upgrading The Stock To Buy CREDIT SUISSE FIRST BOSTON CORPORATION Equity Research-Americas Industry: Electronics/Semiconductors John M. Geraghty, C.F.A.-212/325-3085-john.geraghty@.csfb.com Viewpoint We are upgrading MU from Hold to Buy for the following reasons: Semiconductor capital spending will increase at a minimum rate in 1998. Our estimate is approximately 2%, as detailed in our broadcast of December 24, 1997. Since that time, and through the end of 1997, we have outlined approximately $22 billion in fab push-backs and cancellations. These show capacity being cut back worldwide, but about $14.5 billion (or 66% of the total) is in DRAM fabs. Many of the DRAM fab cancellations were due to Korean companies running out of cash and simply canceling. As the list notes, not all the DRAM players are Korean. Fujitsu has canceled a Scotland plant and Toshiba is also delaying some expansion plans. It also appears that Oki Semiconductor may be leaving the DRAM area and concentrating on Logic ICs. Finally, Taiwan is also thinking about DRAM capacity. Powerchip is re-directing a substantial portion of its capacity towards foundry services and other companies, such as Nanya and Vanguard, are also moving in this direction. Of the total cancellations, the U.S. was only 25% of the total, while the remainder was in Asia, including Japan. Korea was 25% of the total, and this is a predominantly DRAM market. Usually, when prices decline, companies continue to produce units in order to alleviate some of the depreciation costs associated with a very expensive fabrication facility. This has been the case in almost every cycle. Now, it appears that some leading semiconductor manufacturers are scaling back production, as well as capital spending plans, because of a fall in the price of DRAMs and the economic slowdown in Japan. Hitachi has announced it will cut production at eight domestic plants in an attempt to stem losses. The company plans to halt production for between four and seven days a week in February and March, cutting output by about 20%. Mitsubishi Electric plans to close one of its U.S. facilities, where it manufactures older 4Mb. DRAMs. This is believed to be the first closure of an overseas semiconductor factory by a Japanese company. Fujitsu is scaling back capital investment by 30%, Toshiba has postponed construction of an advanced fab (mentioned previously) and NEC now appears to be consolidating production of its 16Mb DRAMs at its U.S. facility. NEC will cease production of 16Mb chips in either Japan or the UK, wherever production costs are higher. If this action spreads to other manufacturers, the overcapacity situation would certainly be alleviated earlier than towards the end of 1998, as is the common forecast. Demand continues to be generally good. PC units should increase approximately 15% in 1998 and the move to Pentium Pro (Pentium II) should also increase the amount of memory per unit. In Table 3, we have a rough outline of how this process should continue. In 1997, we believe supply outstripped demand considerably, at least in the PC market. Additionally, PC DRAM as a percentage of total supply was well below the 50% level we use as a rough gauge of supply and demand parity. In 1998, if we are correct, that balance should begin to shift more towards the demand side, if supply is somewhat restricted. Our DRAM market forecast is given in Table 4. The numbers are not complete for 1997 and the 1998 forecast is subject to revision, especially with the current capital spending cutbacks and production cutbacks now under way. If the bit rate growth, which has averaged above 80% in 1996 and 1997, would slow down to the low 70% range, this could be an extremely positive development. Table 1 Fab Cancellations or Pushbacks in 1997 Company/Location/ Delay to Cost Fab Name ($in millions) Samsung/Austin/Fab 10 Indefinite $450 PH 2 Alpha-TI/Bangkok Indefinite 700 LG Semicon/Chongju/C3 7/1/99 1600 Fujitsu/UK Candeled 1200 Dongbu/Korea Indefinite 2000 Nippon Steel/Japan Indefinite 150 Hyundai/Scotland 12/15/99 1700 Samsung/Kihueng/Fab8 1/30/99 1500 Toshiba/Iwate 3/30/99 1300 Oki/Miyagi Canceled 1100 Powerchio/Taiwan 12/1/00 1350 Texas 1/30/01 1500 Instruments/Avezzano, Italy Submicron 12/30/98 1200 Tech./Bangkok Intel/Israel/Fab18 1/30/99 1200 CTI 3/30/00 400 Semiconductor/Korea Intel/Ft. Worth, 6/30/00 1300 Texas/Fab 16 Interconnect/Sunnyval 4/1/98 180 e/California Interconnect/Malaysia 7/30/99 1300 Texas 12/30/99 1600 Instruments/Dallas, Texas/DMOS6 Hitachi/Irving, Texas 1/30/99 400 Pricing on a near term basis has turned up. Prices for the volume products, such as the 1x16 EDO DRAM have lately begun to turn up on the spot market after a plunge from above $9.50 in March to below $2.00 in early January. This may be a near term aberration. There are some reports that the Koreans are now shifting production to the 16Mb. synchronous DRAM as opposed to the EDO model. The former is a faster part and is better utilized with more advanced microprocessors. Yet, synchronous pricing has also turned up recently. We show this pricing in Chart 1. Production cutbacks in Japan may spur the Koreans and others to also limit some production. As previously noted, lowering production levels is highly unusual for manufacturers. If the Japanese producers are beginning to scale back, it could be an example for other manufacturers in Korea and Taiwan to also limit production. We also believe that further dramatic price cuts from Korean manufacturers are unlikely. As the Korean government seeks to reach an accommodation with the IMF and other lenders, it is unlikely that Korean companies would once again become disruptive on the international scene by cutting prices again. It is in their best interest to see the price rise in order to improve their cash flows.