OUTLOOK AS OF APRIL 2002
1. Earnings - Improving, lots of room for growth
2. Float - Tiny, lots of room for stock splits
3. Broker Coverage - Minimal much room to support #2 as #1 arrives
4. Goodwill/JEDEC - Significant contributor to industry standards work
5. Developing Meter & Embedded Markets - Europe/Asia/US
6. Royalty income - Multiple licencees developing products
7. Product income - Value added approach, aids in developing new markets
8. Unique Products in FRAM and ESRAM - both markets in infancy
9. Competitors still behind - OUM & MRAM still in development; Flash and E2PROM offer more limited performance
10. IP Litigation - Minimal: Settled EDRAM suit v. NEC amicably; NSM litigation going well to date.
11. Research Investment - Growing.
I have not seen, and probably won't see, figures but with IFX, TI, Toshiba, NEC, Fujitsu, Samsung all spending along with RMTR and its other licensees, Symmetrix and its licensees, as well as the non-licensees like Agilent, my most conservative SWAG is that there are more than 15 companies spending an aggregate of $300 mil annually on FRAM specific research in 2002.
In 1999 Dataquest was estimating that it would take 10 years to get FRAM to the point of encroaching on DRAM densities and annual investment close to the $1.5bil. See:
ebnews.com
<<This is where the economics of the argument come into play. If DRAM manufacturers are running
out of room with their current dielectrics, and if they intend to spend $10 billion on developing a
new generation of DRAM, then it is highly likely that the aggregate spending to develop a
manufacturable ferroelectric process could be on the order of $1 billion.
Given that this expenditure will probably be necessary for the 1-Gbit or 4-Gbit density, then it
becomes straightforward to conclude that the DRAM business will be the first to develop highly
manufacturable processes for ferroelectrics.
Dataquest predicts that once DRAMs using ferroelectric materials are in volume production,
FRAMs will not lag far behind. Within two years of the introduction of a ferroelectric DRAM of a
certain density, we should expect to see a non-volatile FRAM of the same density. This
development is likely to bring about the death of all floating-gate technologies within a few years.
Dataquest doesn't want to make today's non-volatile-memory manufacturers too upset about this
sequence of events, though. We don't anticipate this development for at least 10 years. >>
Dataquest has predicted a lot of things, some right and some wrong. It's likely done both in the forgoing.
But if Fujitsu is already working on 1Mb this year:
ebnews.com
Toshiba and IFX are going for 32Mb FRAM at 0.20 cell size by Q4 2003:
eetimes.com
NEC moves cell size to 0.25 this year or next:
ebnews.com
And if TI is moving as fast as they say:
ebnews.com
<<Other companies are researching embedded memory, too, but with different technologies. Texas
Instruments Inc., Dallas, is developing ferroelectric memory (FRAM) for use in its communication
ICs, according to Hans Stork, vice president and director of the company's silicon technology
research unit.
The world tried to embed more and more DRAM for system-on-chip products, but the cost was
just too high, Stork said.
FRAM has cells that are about one-quarter the size of SRAM cells, feature read-write speeds
comparable to DRAM, and are nonvolatile, so the memory consumes less power than either
SRAM or DRAM. Stork said TI has already produced test chips with 64Mbits of embedded FRAM
and expects to see initial products as early as 2003 with on-die memory densities about three
times greater than current levels of embedded SRAM. By 2004, TI could be shipping
communications chips with as much as 80Mbits of embedded FRAM, the company said.>>
Then I have to guess that the rate of investment is picking up considerably. I'm guessing that we'll pass the $500 mil a year mark no later than 2004 and hit $1.0 billion no later than 2006. That won't get us to 1Gb chips, but heck the "sweet spot" is 1Mb to 256Mb as far as I am concerned.
So I guess the question is... how high will the stock price go by 2006? ...And can I hold that long? <Hoo Hoo>
0|0 |
