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 : AMD:News, Press Releases and Information Only! -- Ignore unavailable to you. Want to Upgrade?


To: Maxwell who wrote (4778)3/6/1998 12:48:00 AM
From: Joe NYC  Read Replies (1) | Respond to of 6843
 
Maxwell,

9) AMD is ROCK BOTTOM! At $21 with the K6-3D and K7 in the pipeline with a robust 0.25um process? What else can you ask for?

Evidence. Of course when evidence comes in, the price will no longer be $21. Without evidence, with management lacking credibility, with very small margin for error, you are making a bet.

What is the company worth without K6? That's your downside protection. There is a lot of room on the upside, or so it seems, but you can't forget that Intel's costs are going down as well. Intel is moving to .25u process as well. With the glut of chips in the pipeline, the conservative $160 price of K6 may turn out to be very liberal.

The 300 MHz K6 will be $99 by the end of '98

Joe



To: Maxwell who wrote (4778)3/6/1998 12:51:00 AM
From: James Yu  Respond to of 6843
 
Maxwell,
One more point - IBM's copper process is behind AMD. Here is an article of IBM's copper process.

TechSearch - IBM scores CMOS breakthrough
September 22, 1997,
by Crista Souza
Silicon Valley- IBM Corp. has developed a breakthrough CMOS manufacturing process that analysts say could put the company at the cutting edge of the chip industry. Using copper interconnects instead of the traditional aluminum between metal layers, the company resolved one of the metallization problems that has held back development of powerful chips with true line geometries tighter than 0.25 micron. IBM Microelectronics, Fishkill, N.Y., is preparing the process for commercial rollout with PowerPC microprocessors and ASICs starting at 0.2 micron (L-drawn), although the company said 0.16 micron (L-drawn) has been achieved. Known as CMOS 7S, the 1.8-V technology uses
insulated copper interconnects to boost performance over transistor channel lengths as small as 0.10 micron (L-effective). CMOS 7S stacks six metal layers to pack between 150 million and 200 million transistors on a chip. In the ASIC world, that translates to as many as 12 million logic gates. "This is two to three years ahead of the competition," said Handel Jones, an analyst at International Business Strategies, Los Gatos, Calif. "From a system-IC point
of view, it is one of the most important things to emerge to date," Jones said.
The CMOS 7S development will no doubt catch the attention of both high-end workstation and portable-computer OEMs, which stand to benefit from the increasing speed and complexity of chips that are smaller and require little power to run. And it could go a long way toward solidifying IBM's position as a leader in the merchant chip market. Long recognized as a technology innovator, IBM in just the last few years has made its chips commercially available. "The big difference now is, IBM understands how to develop a technology advantage and then how to bring it to market," Jones said. "I think
they could become a major powerhouse."IBM will introduce CMOS 7S today, nearly a year before the chips based on the process will be ready for commercial release. The process is in qualification at the company's Fishkill and Burlington, Vt., factories. Production is slated to begin by mid-1998, although IBM declined to discuss volumes. Using copper interconnects reduces the resistivity that aluminum introduces as signal speed increases. Industry sources
said that developments beyond the 0.25-micron generation have been inhibited by the subsequent deterioration in metal conductivity. "I think this is one of the key breakthroughs that will allow scaling to a much higher density," said Bijan Davari, IBM fellow and senior manager of logic technology development.
Copper metallization has been studied by the industry for some time. It took IBM's research and microelectronics divisions nearly 15 years of collaborative effort to introduce copper into the CMOS process, resulting in a number of patents along the way, Davari said. However, most of the underlying technology, such as the use of damascene with copper, is in the public domain. Between 1994 and 1996, IBM demonstrated the use of copper interconnects for Sematech, a semiconductor manufacturing research consortium. A number of other companies are said to be developing copper processes. IBM, however, may be the first to bring one to market. Copper has a tendency to "poison" silicon, which makes a chip nonfunctional, Davari said. So instead of metal etching, the company used a dual damascene technique to insulate the copper from the silicon. Alternative approaches to copper are also in development, such as the use of dielectrics as insulation between metal lines. Chip makers have also been looking at fluorinated silicon oxide and other more exotic materials. "I think
many of our competitors will go after [dielectrics]," said John Heidenreich, manager of advanced interconnect at IBM Microelectronics. "Ultimately the whole industry will go to a combination of copper and dielectrics. It just happens that we've chosen to go with copper first." IBM claims chips produced with CMOS 7S can support up to 3,000 I/O connections on a single piece of silicon. Its recently introduced C4 package makes that level of I/O density possible, the company said.
Copyright (c) 1997 CMP Media Inc.

Best wishes

James



To: Maxwell who wrote (4778)3/6/1998 1:26:00 AM
From: FJB  Read Replies (2) | Respond to of 6843
 
Maxwell,

My most serious question is about how they know yields at 0.25 in Austin yet. If they haven't shipped any large volumes of product, is it fair to assume the yield numbers from Ashok are accurate? Are the yield numbers based on "rocket lots" and the actual production yields are still to be determined? Would those numbers be the same?

TIA,

Bob



To: Maxwell who wrote (4778)3/6/1998 1:36:00 PM
From: Petz  Respond to of 6843
 
Maxwell, I'm not quite as optimistic as you on ASP's for plain vanilla K6 ($140 vs $160), but K6-3D will probably command $200, so the 0.25 output will probably average $170.

The k7 will be produced using copper technology and 0.18 process in 1999, according to the 10K.

K6-3D WILL BE A HIT!
I'm a little worried about it. The same emailer who told me that yields on 0.25 were 50%, not 70%, told me that AMD would only make a few K6-3D's. I hope he was talking about 1st quarter output, not the 2nd quarter, because K6-3D is supposed to come out at 300 and 350 MHz right off the block.

Unless Kumar has swallowed a lie or AMD suffers a reversal in yields, AMD will be profitable already in the 2nd quarter.

Petz



To: Maxwell who wrote (4778)3/7/1998 11:06:00 PM
From: Yousef  Read Replies (3) | Respond to of 6843
 
Maxwell,

Re: "AMD 0.25um is the most aggressive process in the semi today ..."

This is an incorrect statement ... AMD may have a high density process, but
they have a low performance FET coupled to a high density/Rs interconnect.
This is the reason that AMD needs a .25um process to just achieve > 266mhz.
This is far from a "world class" process.

Re: "AMD is ROCK BOTTOM! At $21 with the K6-3D and K7 in the pipeline
with a robust 0.25um process?"

The way I see it Maxwell ... AMD can still drop $21 in share price !! <ggg> Are
you saying that AMD will bring out the K7 in their .25um process ... This
would be a real joke. Intel will have the Merced running in a .18um process
at about 600mhz while AMD is bringing on the K7 running at 333mhz.

Make It So,
Yousef