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 : Rambus (RMBS) - Eagle or Penguin -- Ignore unavailable to you. Want to Upgrade?


To: mishedlo who wrote (52478)9/4/2000 9:36:11 AM
From: mishedlo  Respond to of 93625
 
Semi-conductor growth rates
Comments Zeev?
wsrn.com
=========================================================
New York, Sep 03, 2000 (123Jump via COMTEX) -- Semiconductors are heading toward a record-breaking year. In late October 1999, the Semiconductor Industry Association (SIA) predicted that in 2000 industry revenues would grow 21% to $174 billion. Reflecting the extraordinary growth of the industry in the first quarter, the SIA in June revised its forecast to 31% growth and revenues of $195 billion. Even the revised SIA forecast looks conservative now. In August, the trade group announced that the actual growth of the semiconductor industry in the first half of 2000 was 37%.

Dynamic random access memory (DRAM) is considered to be among the fastest growing segments of the semiconductor industry. In fact, analysts are having trouble predicting this year's growth of the DRAM segment. In January, Dataquest and Semico Research, market research firms, announced expectations for 43% and 40% growth, respectively, and DRAM revenues of $30 billion and $28.3 billion. As early as May, however, both research firms revised their forecasts upward. Semico Research predicted DRAM growth of 47.5% and revenues of $30.6 billion, while Dataquest put growth at 58% and revenues at $34.8 billion.

There are several reasons for the extraordinary DRAM growth in 2000, which could be categorized as demand driven and supply driven. According to analysts, demand for DRAM in 2000 will be generated from PCs, Internet servers and Windows 2000. In the same time period, observers note that supply of DRAM will be low in 2000, due to the limited capital investment in DRAM fabs over the past several years.

In 2000, PCs would generate the bulk of DRAM demand. On the one hand, according to Jim Handy, director and principal analyst of memories worldwide service of Dataquest, PCs account for roughly 80% of the DRAM market. On the other hand, according to International Data Corporation, PC unit shipments would increase 18% to 19% in 2000.

The relation between DRAM demand and PC growth is not straight. This is because the memory of the PC is constantly increasing. According to Dataquest's Handy, the average memory size of the PC is growing by about 45% per year. So even if PC unit growth is zero, DRAM Megabyte demand will be up 45%.

In 2000, analysts expect increasing DRAM demand from servers. According to Wall Street estimates, Internet servers will account for 15% of demand, an increase from 10% in 1999. This would be a result of intensifying Internet commerce that requires more and powerful servers to support Web sites and purchase gateways. "Servers alone are doubling and tripling the average system memory size to 4 or 16 Gbytes," says Farhad Tabrizi, vice president of strategic memory marketing at Hyundai Electronics America, San Jose. "We're even talking to some server OEMs about 64-Gbyte memory."

Windows 2000 is also seen as a catalyst of DRAM demand. It will do so by increasing the Mbytes of memory per box. According to Bob Eminian, vice president of marketing and corporate communications at Samsung Semiconductor Inc. in San Jose, the PCs shipping in the first quarter of 2000 had an average RAM of 96 Mbytes to 100 Mbytes. However, Wall Street analysts believe that as Windows 2000 systems begin to ramp, average memory will jump to 128 Mbytes per box.

Analysts disagree on the growth of DRAM bit demand in 2000. According to Dataquest, bit demand will be up 70%. According to IC Insights, another market research firm, bit demand will be up 75%. Finally, Semico Research has predicted an increase of 85%.

However, demand is just half of the DRAM story. Over the past few months, an increasing number of observers have been talking about limited DRAM supply. "We think the market just entered a real shortage - approximately 3% to 5% unfilled demand in the most recent quarter," said Reiko Soga, senior product marketing manager at NEC Electronics Inc.'s North American memory business. Steve Cullen, an analyst at the market research firm In-Stat Group, agrees. He says, "For the next couple of years, demand will exceed supply." Tamao Kikuchi of Nikkei Market Access has reached the same conclusion, "A serious shortage of DRAMs will continue until the end of 2001." Only Dataquest's Handy seems more cautious, "We believe that 2000 will end with a capacity/demand balance, turning into an undersupply in 2001."

The reason for the shortage of DRAM is limited capital investment in DRAM fabs over the past few years. In 1998, capital investment in the segment hit its lowest point at $7 billion and then increased to $9 billion in 1999. This limited capital investment is a result of the fact that since 1996 most DRAM players have been losing money. According to Handy, the sum of all DRAM manufacturers' losses in 1999 was $5 billion.

With limited capital investment, only a few fabs are expected to open this year. According to Tamao Kikuchi, this year DRAM lines will be opened by Nan Ya Technology of Taiwan and by Samsung Electronics and Hyundai Electronics of Korea.

One set of statistics, which are closely followed for clues regarding where the DRAM market is heading, are the prices of DRAM chips. DRAM tags generally dropped during the first quarter of 2000. According to the American IC Exchange (AICE), the 64M (4Mx16) SDRAM PC100 went down 25% to $6.81; the 64M (8Mx8) SDRAM PC100 dropped 32% to $6.13; the 128M (16Mx8) SDRAM PC100 declined 34% to $12.58; and the 128M (8Mx16) SDRAM PC100 went down 34% to $12.31.

In the second quarter of 2000, however, DRAM tags rebounded. According to AICE, the 64M (4Mx16) SDRAM PC100 gained 31% to $8.89; the 64M (8MX8) SDRAM PC133 gained 40% to $8.91; the 128M (16Mx8) SDRAM PC100 added 43% to $18.05; and the 128M (16MX8) SDRAM PC133 added 35% to $17.23.

So far in the third quarter of 2000, DRAM tags have stabilized. According to AICE, between June 30 and August 31, 64M (4Mx16) SDRAM PC100 slid 1% to $8.77; 64M (8MX8) SDRAM PC133 decreased 5% to $8.50; 128M (16Mx8) SDRAM PC100 eased 4% to $17.36; and the 128M (16MX8) SDRAM PC133 inched up 3% to $17.72.

The trend in DRAM pricing generally conforms to analysts' predictions of imminent shortage. According to Handy, DRAM prices do two things: decline during periods of oversupply and remain stable during periods of shortage. The analysts of AICE also provide support to the shortage theory. In their August 8 commentary they noted, "Even with DRAM manufacturers ramping up production, supply is short of demand."

The shortage will result in an increase of the DRAM Average Selling Price (ASP). According to IC Insights, this year DRAM ASP will increase 26.5% to $5.97, on top of a 16% increase in 1999. Further ahead, IC Insights sees an increase of 45% to $8.66 in 2001 and a modest 9% increase to $9.44 in 2002.

Despite analysts' optimism regarding DRAM revenues this year, it is too early to cheer. One factor that may prove the high expectations wrong is the behavior of PC companies. According to James Carbone of Electronics Purchasing, "The PC business is very cost conscious. When memory prices were high in the mid-90s, PC companies were slow to add memory to systems. When prices fell, they loaded systems with more memory." Correspondingly, if DRAM tags increase too much this year, there is the chance that PC companies will cut the memory size of the systems they are selling.

There are already signs that Carbone is right. According to a worldwide DRAM price survey conducted by ICIS-LOR, a market research firm, spot prices are falling. As of August 4, the DRAM spot prices of 128Mb SDRAM microchips (PC133, 16M x 8) declined 1.86% to $16.44 in the United States, declined 0.25% to $16.97 in Europe and rose 0.23% to $16.88 in Asia. The research firm attributed the decline in Europe and the United States to the restraining effects of high prices on demand and not to any oversupply conditions. Further, the researchers said, "Tight supplies will likely continue until October when demand is boosted for the year-end sales season. DRAM prices will probably stay at the current level and thus large increases in memory capacity of PCs aren't expected."

Yet another factor which complicates the shortage situation is the structure of the DRAM market. The market currently is fragmented with five memory types fighting for dominance. These include the SDRAM PC 100, SDRAM PC 133, EDO DRAM, Rambus DRAM (RDRAM) and Double Data Rate DRAM (DDR DRAM). According to Farhad Tabrizi, vice president of marketing at Hyundai Electronics America, PC100 and PC133 SDRAMs currently account for 85% of the market, EDO about 8%, and Rambus and DDR together around 6% to 8%. Fragmentation has made forecasting demand an increasingly difficult task. As it becomes more difficult for producers to forecast demand, it is likely that there will be shortages in some types of DRAM and oversupply in others.

Considering the developments in the DRAM segment from an investor's perspective, it appears that the major players in the segment are likely to split the bulk of this year's revenue increase between themselves. DRAM supply is limited, even while demand is rising. Except for the three-DRAM lines at Nan Ya Technology, Samsung Electronics and Hyundai Electronics, capacity is not going to expand. With this limited capacity, it is clear that the jump in DRAM revenue will be a result of the stabilized DRAM price. Thus, those with the largest capacity will capture the bulk of the revenue increase.

According to IDC, in 1999 three companies controlled nearly 60% of the DRAM market in terms of projected output of 64Mb DRAM equivalents. Hyundai (combined with LG Semicon) was first with 23.5% market share, Micron Technology (NYSE:MU) was second with 17.6% market share, and Samsung was third with 16.8% market share. The ranking was slightly different in terms of revenues. According to IC Insights, in 1999 the top five DRAM manufacturers were the following: Samsung with 22% of the market and sales of $4.6 billion; Hyundai with 21% of the market and sales of $4.3 billion; Micron with 18% of the market and sales of $3.7 billion; NEC with 10% of the market and sales of $2.0 billion; and Infineon (NYSE:IFX) with 8% of the market and sales of $1.7 billion.

The recently-reported financial results of Infineon and Micron confirm the prediction that the big six will split the bulk of this year's DRAM revenue increase. For the six months ended June 30, 2000, Infineon reported memory revenues of 1.4 billion euros ($1.2 billion). This was a 115% jump compared to memory revenues of 629 million euros ($560 million) for the six months ended June 30, 1999. Micron reported semiconductor revenues of $2.7 billion for the six months ended June 1, 2000. This was a 106% jump compared to semiconductor revenues of $1.3 billion for the six months ended June 1, 1999.

Investors have responded quickly to the growth prospects in the DRAM segment. As of August 30, Micron's share price has gained 120% to $83.7 from $38 on January 3, adjusting for a 2-for-1 stock split in May. In the mean time, Infineon's stock has had a roller-coaster experience. It ended its first trading day on the NYSE, March 13, at $70. Soon after, on April 5, the stock reached a low of $48.50. Then the stock rebounded, reaching a high of $88.25 on June 27. Currently, Infineon is trading at $66.87.



To: mishedlo who wrote (52478)9/4/2000 9:38:40 AM
From: mishedlo  Read Replies (2) | Respond to of 93625
 
RMBS TA
Message 14297309
Comments Zeev?
Thanks to Jim Willie
====================================================the MA's are supporting RMBS between 80 and 83
my faves are the 18-day MA, 30MA, 50MA
notice how RMBS closed right at the 50MA, at 80
the 50-day MA is the lowest of these clustered moving averages
when clustered they contain collective supporting power
these are clustered now

got some trendlines at work, it seems
the failure just over 90 two weeks ago happened in line with the two prior highs of 120 and 110
the bounce this morning at 76 happened in line with the two prior lows of 37 in May and 65 in early August
I love the power of trendlines
one down and one up indicate a tightening in price, kind of like a "wedgie", which means likely less volatility soon

relative strenk was trending nicely upward until reversal at 90
now it sits at midrange 50 before today, probably still there

daily socratic indicator shows we are in the throes of the downcycle, daa
it dont tell me anything more

money flow really shows no material changes in over 6-7 weeks
they be buying, they be selling, in same magnitude

my gut says the two weeks of lateJuly/earlyAug activity between 65 and 75 will support this stock
also, my gut says the trendline will mean support at 75
if it fails, then 65 will be even stronger
any "rear entry" between 65-77 seems a good move
we should have a couple more "attempts" to sell off

these lawsuits are often silly, but this may have substance
if Rambo claims IP from mutual research, it could be trouble
not an expert in this regard
you cannot claim your neighbors' children as your own

did Volt really say he is gone until "The Buss" hits Route 200 ???
/ Jim Willie



To: mishedlo who wrote (52478)9/4/2000 11:27:06 AM
From: Scumbria  Read Replies (2) | Respond to of 93625
 
Michael,

I like your photographs!

1) Hate RMBS because it is a threat to AMD
2) Hate RMBS because you do not like its business model
3) Hate RMBS because you are paid to hate it
4) Hate RMBS because you do not undertstand it

If AMD had developed said technology, acted in exactly the same manner with INTC supporting DDR, you would be arguing the other side.


I'm not sure why these questions keep coming up. I started posting about the problems of DRDRAM performance on a CPU motherboard over two years ago. I did this because I am an engineer, and I understood the problem. I have no financial interest in Rambus either way, and am not impacted by the outcome.

I was the first to complain about the lousy architecture of the K6, and my only interest on these threads is the facts.

Scumbria