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 : PC Sector Round Table

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: Yogi - Paul who wrote (1052)10/16/1998 2:58:00 PM
From: Mark Oliver  Read Replies (2) of 2025
 
Intel Corp. (INTC) 84 9/16: chip maker is making a $500 mln investment in Micron Technology (MU 31 3/4) and will own a 6% stake in the company; says investment in MU to help speed up Rambus (RMBS 61 1/2) memory chip development.....

In-Stat In-Sights

Getting Ready To Rambus

By Steve Cullen From the October 12, 1998 Issue of Electronic News

Senior DRAM Analyst, Semiconductor Product Services

How do we get There from Here? At the Intel Developers Forum last month, the DRAM talk was almost exclusively about Direct Rambus, the technology that Intel has identified as its primary DRAM thrust from mid-1999. Intel will be ready. They provided Forum attendees with the full complement of electrical, physical, and thermal design information necessary to design the RIMM modules that will take the Rambus DRAMs and the motherboards that will hold them.

They also reiterated their intention to focus future chipset memory interfaces on Rambus, starting with chipsets intended for the high-end desktop from mid- 1999.

Rambus will be ready. They provided 14 DRAM suppliers with design implementation packages in February. They have assured that there are suppliers for all of the supporting parts from clock drivers to sockets. And at Intel's Forum, they announced a validation program to minimize compatibility problems by confirming the specification compliance of these components.

The DRAM designs will be ready. Four suppliers (LG Semicon, NEC, Samsung, and Toshiba), representing over 40 percent of the DRAM market, showed roadmaps for their Direct Rambus DRAMs. An additional four suppliers (bringing the total market share represented above 75 percent) joined in for a panel session. There was a consensus on the panel that demand for RDRAM would be about 15 percent of the DRAM market by the end of 1999, a number that is consistent with the Cahners In-Stat Group forecast as shown below.

And a number of potential users, including Compaq and Dell, have announced their intention to use RDRAM in future PC product designs. So what is the problem?

It could be DRAM capacity. Capacity? Isn't there over-capacity in the DRAM industry? Yes, there is. And wafer fab capacity is not the problem. RDRAMs can be made in the same fab lines that make EDO and SDRAM. There are certainly more than enough of those around. Just ask any DRAM supplier. But Rambus requires some unique back-end equipment for assembly and test. Where most SDRAMs are packaged in TSOP packages, most Rambus DRAMs will use a micro BGA package. This means new assembly equipment. Where most SDRAMs run below 150MHz data rates, Rambus DRAMs will start at 800MHz data rates. This means new test equipment. All this new equipment is expensive and can have lead times of six months or more.

In the good old days, customer intentions might have been enough for the DRAM industry to make the investment in this needed back-end equipment. If it didn't get used today, it would get used tomorrow.

But in today's DRAM market, with its non-existent profits and tight capital expenditure budgets, suppliers are reluctant to spend this kind of money without some sense of real commitment on the part of their customers. So some suppliers are installing a minimum set of back-end equipment and waiting for demand to build before installing more.

Why take the risk of being a leader when being a fast follower may be good enough? And what are the customers doing? They have been enjoying the soft DRAM market. Need DRAMs? Order them today, get them tomorrow. No need for long-term planning here.

But if customers continue to follow this strategy, they may be in for a surprise when they try to place volume orders for Direct RDRAM. DRAM suppliers who have not installed sufficient back-end capacity will be unable to meet demand, regardless of how much fab capacity they have. And considering the lead times for the back-end equipment, this problem will not be resolved overnight. Time is beginning to run short. PC OEMs, if you want Direct RDRAMs in 1999, it is time you had a serious talk with your DRAM suppliers. Better yet, give them purchase orders.
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext