repost from MU thread re CC (my emphasis added):
From: Carl R. Tuesday, Sep 29 1998 2:00AM ET
Ok, now I've heard the conference call so this post will be much better than the last one. <G>
1. They have $285 million available to spend on semiconductors
2. The majority of wafer starts are currently at .21 micron. By next year the majority will be starting at .18 micron.
3. Sequential bit growth over next three quarters will be low double digits. Last quarter produced bit growth was up 22%.
4. Sept pricing is strong, but due to shortages caused by strong demand from OEMs MU will be nearly 100% contract prices and absent from spot market. Normally they are 25% spot. Spot prices are 10-15% above OEM prices which is unusual, so prices will probably be firm in October as well, and may rise a bit more.
5. Capital budget for the next 12 months is about $800 million to $1 billion.
6. No short term borrowing last quarter.
7. Capital budget for upgrading TI is another $1 billion.
8. TI fabs must be refitted to bring up to MU standards. Another $1 billion to be spent here. The TI technology will be scrapped and replaced with superior MU technology. In Japan and Singapore this will probably be done on the fly.
9. TI venture will reduce earnings by about $.25/quarter for the next year.
10. Capital expenditures for the DRAM industry this year will be only $4-5 billion, down from 10-15 Billion /year in 1995, which is less than is needed to sustain the industry.
11. No comments on INTC rumors.
12. PC100 lead time is now at 8 weeks.
13. Samsung is just ramping up .25 micron according to an analyst, implying that MU is a generation ahead. MU says that they are confident that they are clearly a generation ahead, and that the fact that they are more financially sound than the competition will allow them to remain ahead.
14. Joint venture arrangements that TI had in place will remain in place after MU takes over.
15. TI deal will probably close within a couple weeks.
If the bit growth will be "low double digits", then costs will fall by less than my prior post estimates, probably because some of the cost improvements are already in place. But prices may be firmer than my prior estimate. Plus I shifted things a month. It looks like at least 2/3 of this quarter (sept/oct) will have firm prices. I believe that the costs may already be at $8 and may fall another 10-20% this quarter, but it probably won't get below $6.50-$7. Also if they are only increasing bit growth in "low single digits" then they probably need more like $1.50 positive margin per chip to be profitable. A cost of $6.50 and and ASP of $8 would do it.
With older capacity coming off line, and reduced capital spending worldwide, combined with the fact that RDRAM will require 15-30% more silicon, it seems unlikely to me that prices will continue to fall at the rate of the last two years. If MU at .21 micron can barely hope to break even, competitors at .25 micron must be really hurting, and competitors with 6" fabs and .3 micron technology are dead meat. More closings of old fabs are probably likely. Even though some shrinks will be done, as shifts take place from SDRAM to RDRAM requiring more transistors, the total worldwide supply of memory will probably not increase much at all in the next several quarters, so demand could catch up (especially if I am right about a surge in PC sales over the next year). Thus prices may well be fairly firm for the next year, and could even rise, which could be why INTC wants to fund MU to further increase DRAM production. <VBG>
I still believe that MU will be a winner in the DRAM consolidations, and that losers such as TI and Nippon Steel Semiconductor will continue to fall by the wayside.
Carl
|
