Hi all; Well, I caught my big fish. I'm afraid your
management at RMBS are not above activities on the web
that are a bit scaly, or maybe slimy is a better word.
And not particularly bright, either. I have to say that
I am stunned and surprised.
I'm sure that the revelation is only going to shock those
people that really believe that Rambus is a totally
ethical, above board and honest company, but there are
a few of those naifs still running around buying stocks.
In the mean time, I am going to concentrate on the
dramreview web site, while it is still there.
On the product page, the one where dramreview manages to
mention every product that is dreaming of shipping with
RDRAM, while missing most of the DDR products out there, a
list of analyst predictions for memory types over the next
four years appears:
Analyst RDRAM DDR
00 01 02 03 00 01 02 03
% + -- -- -- -- -- -- -- --
Dataquest | 12 33 50 62 n n n n
In-Stat | 8 25 45 55 n n n n
Semico | 3 2 1 1 9 20 29 38
I went and did a web search to verify the above numbers,
but couldn't find them. Perhaps they are available only
to people who subscribe to expensive analyst reports.
That would take a lot of money, wouldn't it? Kind of
makes you wonder just who dramreview is, doesn't it?
Anyway, I did find a couple of links. I couldn't locate
a Semico prediction as low as the one above, maybe I'll
go looking again for one, and if I find it, I'll link it
to this post.
In late '99, Semico was predicting 7% for both DDR and
DRDRAM in '00:
Direct Rambus and DDR each will account for less than
7% of worldwide DRAM unit shipments in 2000, while a little
more than three-fourths of the market will be PC100/PC133
devices, according to Semico Research Corp., Phoenix. In
2001 and 2002, DDR and Direct RDRAM will gain ground at
the expense of PC100/PC133.
...
A "macro" forecast might do for a product such as PC133,
which is likely to become mainstream quickly, according to
Leon Shivamber, vice president for product management at
Arrow Electronics Inc., Melville, N.Y. But for higher-end
devices, predictions from individual customers are more
critical, he said, adding, "At that level, there is likely
to be a lot of hand-holding until the products are broadly
adopted."
I may as well rub the following quote from Arrow in the
thread's face, this is what I've been saying for a long
time, of course:
That's apt to be the case for Rambus DRAMs, according to
Shivamber. "But [the uncertainty surrounding the
technology's ramp-up] will not affect what we do until
there is a production schedule," he said.
"Lots of customers are struck by Rambus and are thinking of
designing it in, but very few of the big boys have designed
in Rambus in large scale," Shivamber added.
techweb.com
Back in late '98, Semico was at the low end for predictions
of DRDRAM production in '99 (and was overly optimistic, I
believe) back in late '98:
Semico Research Corp., Phoenix., has cast a dim light on
volume projections, estimating that vendors are equipped to
produce a mere 33 million Direct RDRAM ICs in 1999. High
end processors shipping from Intel alone will require
between 170 million and 200 million DRAM units, according
to Semico analyst Sherry Garber. "You can't get there from
here," she said.
Meanwhile, Rambus was pumping the numbers, and was quite
wrong:
"We've talked to the DRAM companies, and we've added up
the volumes they plan to make, and it appears to us to be
north of 100 million units," said Subodh Toprani, vice
president and general manager of Rambus' logic division,
Mountain View, Calif. "And remember, these are plans being
made in 1998 for products shipping in 1999."
techweb.com
-- Carl |
