Found this posted on The Fool by HowlOnMoon... found it interesting.
Hello all,
I have been lurking in the background for some time now. I flirt around with different
boards (mostly AOL) but this one seems to have its "head screwed on straight". Even
the bashers have a degree of respect. (the Yahoo! board is a waste of time).
First of all I am a Rambus long. Bought the stock in 2/99, 7/99, and 8/99. This
company have made a pile of money for me. Well into the six digit area. I have always
been a long investor. I don't believe much in charts, daytrading, shorting, ect. I DO
realize that a very small percentage of people make money at it but most do not. I
actually took 3 weeks of vacation a couple of years ago to day-trade. I got out with
only a couple of thousand dollars in the hole. I saw an experienced trader loose over
$80,000 on an overnight position (bad news after the market). I saw people make
$1,200 per day. But I never saw anyone make a killing over the short haul.
Ok, enough about why I am here ....... let's talk Rambus!!
First let's look at the facts!
The price of RDRAM has been in a free fall.
On 2/25/2000, RDRAM cost 6.8x what PC133 cost. On 8/1/2000 RDRAM costs
only 2.4x what PC133 cost. This is an incredible price reduction in just over 5-months.
Yes, their have been a number of message board posts noting the recent $80 decrease
in price of RDRAM on Dell computers recently. Anyone remember just a few months
ago when Dell charged over $800 more for RDRAM computer? Now it is just
$400.....like a 1/2 off sale in record time! Now don't forget that Dell includes a nice little
retail profit margin for themselves here!
Now, a reasonable assumption from the FACTS of rapidly falling prices.....
Rambus and Samsung (all well as others chattering in the background) have stated that
they expect the RDRAM premium to be only 10-20% by the middle of 2001. No
quarantees, just expectations based on current trends.
Why? Well, two main reasons. The first is the ramp-up in volume production. This has a
tremendous effect on pricing. Not only in memory production but everything else too.
but the second is just as exciting.....new advancements in production techniques. Yes,
companies are finding ways to improve PC800 yields, reduce die size, decrease
packaging costs, reduce testing costs.
Some more facts now...
Don't forget now....a year ago and it was Rambus and Intel mainly. Now, it is many,
many, many companies. And not just memory manufacturers. Testing companies,
packaging, software, telecommunications, etc. Regardless of what folks may speculate,
these people are not idiots and do not throw hundreds of millions of dollars and even
billions in some cases into Rambus R&D and production capabilities foolishly. These are
smart people folks.
They see the benefits and necessity of Rambus technology for the FUTURE. I am not
going to argue the point that Rambus offers little incremental imporvements with current
technology and the aging Pentium core. Let's face it....not much improvement, at least
using the standard "benchmarks".
But the P-4 is a whole new system. Intel has spent YEARS and mucho dinero
developing this. And yes, all with Rambus in the forefront of their major R&D $$$$$.
Ok, reasonable assumptions again....Let's go on the high side and say the pricing
premium is 20% higher for RDRAM rather than 10%. So what does this mean? Well,
on a "standard" computer with 128K of RDRAM this would cost (at wholesale) a mere
$32 more than a computer equipped with SDRAM. For you power users, this is a
beath taking $64 more for 256K. Whew! I just don't think $64 is a deterent for a
computer that will last 3-4 years. I spent $91 in gasoline for my boat this past weekend
and $59 on a lunch for 3 people. (Anyone ever been to Canyon Lake in Texas?)
The P-4 debuts at 1.4 gig with 1.3 available...FSB is 400m. RDRAM will be needed to
make this chip perform to capacity.
Ok, reasonable assumption again....using Moore's Law....the performance of the chips
will double every 18-24 months. This means that a 2 gig Intel chip is very likely by June
30, 2001. Chips at 1.5 gig will become mainstream, and less expensive, which in turn
offsets some of the premium price for RDRAM. When the chipset for SDRAM is
available, only the value players (consumers) will want it. I would predict (whoaa! is that
not what our beloved analysts do?) that SDRAM finds a niche market (no offense Mr.
Kumar). Sure, there will be a market for SDRAM but it will rapidly shrink in market
share. DDR? Heck, Ihave no idea except that you kinda have to believe Intel when it
says that DDR has a place in servers.
But also, the DDR argument for servers was largely based on the latency problems that
were inherent in the EARLY versions of RDRAM. The gap on this latency issue has
closed quite nicely. Who now knows what may happen now?
Does it matter?? Looks like it may not! Rambus has three memory manufacturers now
paying royalties on SDRAM & DDR with another one (Mitsubishi) admitting they are in
"negotiations" with Rambus. Ok, all you shorts jump up and down now and holler about
how the Japanese companies just "cave-in" when dealing with U.S. companies in court
or potential legal proceedings. Oh yes, you FUDsters don't forget about these
companies being "minor" players in the market. Face it, the market is fragmented with
consolidation occuring. Currently, some 15 memory makers control about 90% of the
market. Many, many, others are lined up for the other 10%
Will all the memory makers cave and sign? Likely not. I would just be that a legal action
or two comes out of the patent claims. I like my chances so far though!
Ok folks, must close now. Ihave a meeting tomorrow morn concerning the health of the
real estate market in Texas.
One more thing......what do I personally see as the biggest threat to Rambus right now?
Intel of course! No, no, not what you immediately thought. Intel has a nasty history of
engineering and production problems. They always have a snafu or tow that delays
introduction of their product (ok, i love Intel and have owned it since 1995 but let's face
facts) I just hope they have no problems getting Willy (P-4) to market on time.
On October 1st intro would be just dandy....giving Rambus a good 9-12 months head
start to reduce the price premium, demonstrate it's SUPERIOR technology, and meke
the patient ones handsomely rewarded in monetary terms.
Don't become disillusioned by the gyrations of the market. Look out two years, not two
weeks or two months. Sure, some shorts and daytraders may get big-time lucky. But I
just bet I am way ahead of 99.9% of them in the long haul.
Good health to all.....yes, even you Vamp!
HowlOnMoon |
