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.
Politics : Formerly About Advanced Micro Devices -- Ignore unavailable to you. Want to Upgrade?


To: Rob Young who wrote (56720)4/29/1999 12:47:00 AM
From: Gary Ng  Read Replies (1) | Respond to of 1575191
 
Rob, Re: At one time Intel was a force to be reckoned with in the
DRAM industry... then Samsung came along

I thought it was the Japanese(NEC etc.) not Korean.

Gary



To: Rob Young who wrote (56720)4/29/1999 12:59:00 AM
From: Process Boy  Respond to of 1575191
 
Rob - <At one time Intel was a force to be reckoned with in the DRAM
industry... then Samsung came along.>

Sheesh Rob. That was 15 years ago. As you may know, Intel invented the DRAM. Ihave seen al sorts of commentary from Andy Grove about that time frame.

Anyway, I believe Intel is still a force to be reckoned with in the DRAM industry. Note bolded reference to Samsung.

ebnews.com

Will Intel spark Japan's DRAM turnaround?
By Andrew MacLellan
Electronic Buyers' News
(02/05/99, 05:25:32 PM EDT)

Silicon Valley

As the sun rises on a new fiscal year, Japan's DRAM makers may be looking for a helping hand to assist them in keeping pace with the global chip market.

Short on working capital, these industry icons have lost much of the luster that set them apart during the boom years of the DRAM gold rush. Now, as they prepare their annual budgets for the start of the April 1 corporate year, executives at Japan's top DRAM companies are entertaining options that would have been anathema just a few years ago.

Toshiba Corp. confirmed last week that it is seriously considering an offer from Intel Corp. to help it meet 1999 DRAM production goals, and NEC Corp. said it, too, is open to outside, even foreign, investment. Mitsubishi Electric Corp., by contrast, rejected an Intel offer last week, saying that it is able to fulfill production plans unassisted.

Whatever the outcome, the investment planning currently under way at Japan's DRAM companies could have a serious effect on this year's DRAM market, which is poised to transition to the new, high-speed memory

interface known as Direct Rambus DRAM. Companies willing to expend the capital needed to shift to Direct RDRAM could lay claim to the volume memory-IC market, while those on the sidelines may have to content themselves with trailing-edge, lower-margin business, observers said.

"The Japanese have basically adopted an attitude of extreme conservatism with respect to DRAM, which means the propensity to invest is very, very low," said Victor deDios of research firm deDios & Associates, Newark, Calif. "So anyone pushing a new technology into the market, which in this case is Intel with Direct RDRAM, has to ante up the dollars."

Toshiba, which already produces limited quantities of 72-Mbit Rambus chips, will shift to 128- and 144-Mbit devices later this year. But the company so far has trimmed about $100 million from this year's $1 billion investment budget and may have trouble securing new capital amid a nationwide recession, according to analysts.

"We did receive a proposal to fund production of Rambus DRAM, but we cannot comment further than that," said a Toshiba spokesman, who confirmed that Intel is behind the cash offer. "We are just starting to look at it."

Though it may be the first Japanese company to publicly acknowledge such an offer, Toshiba's position is hardly unique in the industry. Last October, Intel made a $500 million investment in Boise, Idaho-based Micron Technology Inc., which swapped the rights to 6% of its shares for a promise to aggressively ramp Direct RDRAM in 1999.

More recently, Intel extended a $100 million incentive to Samsung Electronics Co. Ltd. in exchange for convertible bonds, representing about 1% of Samsung's outstanding common stock. Toshiba would not comment on the cash value of Intel's latest proposal.

With decidedly deep pockets, Intel has approached other memory vendors as well, according to industry sources. While it would not comment specifically, NEC said it would consider an equity investment if it helped the company in its Direct RDRAM ramp.

"Given the tight situation at the moment, we're prepared to think about third-party investment to help with the production of Direct Rambus," a company spokesman said. "Who, what, and where, we're not prepared to talk about at the moment. ... But I think the Japanese DRAM industry has to seriously think about [outside investment] given the potential demand for Direct Rambus by the end of the year."

In all, Intel could spend as much as $1 billion to grease the industry's shift to Direct RDRAM, which is expected to ship in about 15% of all PCs by the end of 1999, according to deDios. Intel would not comment on its investment strategy.

At least one Japanese DRAM maker, Mitsubishi, formally rejected an Intel offer last week, but said a future deal is possible.

"Although Mitsubishi Electric was approached by Intel with regard to possible financing measures for production-related investments, we declined due to the fact that we are able to achieve our current production plan with our present capital investment levels," said Koichi Nagasawa, general manager of the company's chip group, in a statement.

DRAM makers Hitachi Ltd. and Fujitsu Ltd. would not comment on their investment plans, although Fujitsu said it has not been approached by Intel.

Part of the problem for Intel in pitching Rambus stems from associated costs; for example, Rambus will require new packaging and test equipment. With many of Japan's chip makers expecting to lose money this year (Hitachi last week increased its anticipated year-end loss to a whopping $3.35 billion), companies there may have trouble single-handedly fielding the manufacturing infrastructure required to ramp Rambus.

"The Japanese market, along with everyone else, is hurting because of the ASP decline over the last couple of years," said Sherry Garber, an analyst with Semico Research Corp., Phoenix. "That's why you've heard repeatedly of companies backing off on new investment. In order to bring Rambus to market, the back-end cost of new packaging and testing equipment is substantial-about $100 million and up."



To: Rob Young who wrote (56720)4/29/1999 1:40:00 AM
From: Yousef  Read Replies (2) | Respond to of 1575191
 
Rob,

Re: "Funny thing is, nary a mention of DRAM, RAM, Samsung in "Only the
Paranoid Survive." Now that's funny."

You really need to get your facts straight, Rob ... Samsung didn't have
anything to do with forcing Intel to leave the DRAM business.

Re: "Just how weak do you think Samsung's .18 copper-based process will be?"

I expect that Smasung's .18um process to have FET performance of a .25um
process coupled with a dense interconnect. Not optimum for building
high speed CPU's.

Make It So,
Yousef



To: Rob Young who wrote (56720)4/29/1999 2:13:00 AM
From: Tenchusatsu  Respond to of 1575191
 
<Funny thing is, nary a mention of DRAM, RAM, Samsung in "Only the Paranoid Survive.">

That was back in 1983, and judging from the stellar performance of INTC over the last 16 years, I'd say that Intel was laughing all the way to the bank.

Meanwhile, Samsung is still stung by IMF. (Well, not the IMF itself, of course. In Korea, the acronym "IMF" has become an expression for the current economic depression occuring there.)

Tenchusatsu



To: Rob Young who wrote (56720)4/29/1999 3:07:00 AM
From: nihil  Read Replies (1) | Respond to of 1575191
 
DRAM in "Only the Paranoid Survive" --

Checkout pp. 85- 94 where Grove discusses dealing with the first 10X inflection point and getting out of memories.

In some other source, Moore or someone discusses the big benchmark study in the 80's that led Intel to the conclusion that it was the least effective chip manufacturer of the six in the study. Samsung was the best. This led Grove and Moore to emphasize manufacturing efficiency (which is where Barrett came in.)