RMBS Article:
Thanks rmbs thread!
Analysis: RAMBUS, A Burgeoning Semiconductor Powerhouse by Adam Schwebel
Company Basics Rambus is a semiconductor design house. The company employs a
much different business strategy than other better-known semiconductor companies such
as Intel, Motorola, Texas Instruments, and Samsung. The company makes high-speed
chip-to-chip interfaces. Currently, the focus has been on memory interfaces, but within
the next few years a much wider variety of products using RAMBUS' high-speed
interfaces will be introduced. The company was incorporated in 1990 by two Stanford
electrical engineers, Mark Horowitz and Michael Farmwald. The company went public
on the NASDAQ on May 14, 1997 under the ticker symbol RMBS. Since its inception,
RAMBUS has been awarded 95 patents covering fundamental aspects of chip-to-chip
interfaces; over 90 are still being reviewed by the Patent Bureaus. The first widespread
adopter of RAMBUS technology was Nintendo. The Nintendo 64 utilizes RAMBUS'
first generation high-speed interface technology. Since then, Sony has employed it in its
PlayStation2 game console. Direct RAMBUS, the company's second-generation
technology, is employed in order to meet the high bandwidth requirements of the
advanced graphics engine. The third, and most lucrative to date, platform RAMBUS'
technology is consumed is in PC systems. Since late 1999, Intel has enabled the use of
Direct RAMBUS in its Pentium III line of microprocessor-based systems. In addition,
many networking companies (the likes of Vitesse, PMC Sierra, and Brocade) and other
computer manufacturers (SUN Microsystems and Compaq ALPHA) have designed
RAMBUS technology in their future products, however, very few are currently shipping.
Papa Intel On November 15, 1996, RAMBUS entered into a vast licensing agreement
with Intel Corporation. The terms of this agreement, which have been amended, are
somewhat of a mystery. What is known is that Intel agreed to employ and promote the
technology in its products as best it could. If Intel were to ship 20% (twenty percent) of
its chipsets in two consecutive quarters, the company would be rewarded handsomely.
The allotment of 1,000,000 (one million) RMBS warrants at a strike price of $10.00 per
share would be awarded at the completion of the target. Since RMBS split it's stock 4:1
on June 14, 2000, Intel's warrants are now 4,000,000 (four million) at a strike price of
$2.50 per share. This will amount to Intel effectively owning more than 4% (four percent)
of RAMBUS. There are some details that can't be overlooked. Intel must complete its
target by December 31, 2000, otherwise the warrants become invalid. In the amended
filing, the warrants will not be granted to Intel until 12 (twelve) months following the
completed quarter in which the goal of 2 (two) consecutive quarters of 20%+
RAMBUS-compatible chipsets are shipped. If Intel achieves this goal one new board
member, appointed by Intel, will be selected to sit on RAMBUS' Board of Directors for
that quarter. In addition, upon the completion of the target, Intel will be automatically
upgraded on its license for RAMBUS patents. If they meet an unspecified stringent goal,
they could get royalty-free access to the patent portfolio for a brief period of time. Since
Intel currently has roughly 80% (eighty percent) market share of the PC space, if 20%
(twenty percent) of its microprocessors were RAMBUS enabled, a large sum of
royalties would flow in in the form of chipset memory controllers (2.5-5% royalty rate)
and RDRAM shipments (2-3% royalty rate).
DRAM Industry In 1998, the DRAM industry was in the doldrums. Nineteen ninety-nine
saw the reemergence of growth for the first time in a few years. This year DRAM will
comprise a $29 (twenty-nine) billion market. Next year, the market should see a large
spike upward to $54 (fifty-four) billion dollars. In 2002, we should see a $76
(seventy-six) billion market and a peak in 2003 at $95 (ninety-five) billion. In the
beginning, I said RAMBUS has been focused on their memory interface. All of the top
13 (thirteen) DRAM manufacturers are licensed for the RAMBUS specification. Well,
this year RAMBUS' DRAM interface (RDRAM) is projected to capture 10% (ten
percent) of the $29 billion market. By next year, I expect RDRAM to explode to take
30% (thirty percent) of the then $54 billion market. But here's the wildcard, RAMBUS
has demanded royalty payment for ALL high-speed synchronous memory devices and
their controllers. The company firmly believes these devices infringe RAMBUS patents
and they should be fairly compensated. Over 95% (ninety-five percent) of the current
market comprises these high-speed memory technologies and it will be higher next year.
Intellectual Property (IP)…A License to Print Money RAMBUS engaged Hitachi
through litigation, based on their belief that Hitachi and SEGA, who used the Hitachi
products, infringed 8 (eight) key patents. In June of this year, after 5 (five) months,
Hitachi agreed with RAMBUS and obliged to pay an up front settlement fee as well as
quarterly royalty payments on SDRAM and DDR SDRAM and the controllers that
interface with these memories. Toshiba has also agreed, albeit not through litigation, to
pay these royalties as well as on their proprietary DDR FCRAM. This is only two of the
top 13 DRAM producers, but it is my belief that all of the remaining 11 (eleven) will sign
a similar agreement with RAMBUS, be it through litigation or agreement. The next
manufacturers to sign, in my opinion, will be NEC Corporation and Fujitsu. Once the
memory makers are signed, RAMBUS will receive a 1.0-3.5% royalty on virtually the
entire DRAM market! However, I don't believe RAMBUS will have all the
manufacturers signed up by the end of calendar year 2000 (largely due to ongoing
litigation). My estimates for 2001 are that the company will receive royalties on 65%
(sixty-five percent) of the $54 billion DRAM market. By 2002, I believe the company
will receive royalties on virtually the entire market (98%), with its interface gaining 60%
(sixty percent) market share! RDRAM will garner 85% (seventy-five percent) market
share by 2003, while the total market will be $95 billion.
Life beyond DRAM On December 9, 1999, RAMBUS issued a press release detailing
their intentions to expand their patent portfolio and business scope. In addition to
high-speed memory interfaces, the company will pursue other avenues that required
high-speed electrical interfaces. First on the list is communications. This market is already
large and with the growth of the Internet and the convergence of data/voice/video
communications, high bandwidth connections are becoming a necessity. RAMBUS
introduced its first dedicated communications technology on April 5th of this year…a
3.125 Gbps Quad SerDes (Serializer/Deserializer) Cell. This high-bandwidth I/O cell can
be integrated into ASIC (application specific integrated circuits) controllers or network
processing units to provide a high-speed interface for backplanes in network routers and
switches. Underscoring the company's desire to expand its technology offerings and
widen its business scope, a new acquisition and investment department was conceived.
In his speech during the February Annual Shareholder's Meeting, Geoff Tate detailed
RAMBUS' “other” market to be $150 billion by 2002. This “other” market
encompasses many high-growth industries, some of which have not even begun to
blossom yet. Currently, communications is the largest segment of that market, but digital
set-top boxes, 3G cell phones, personal digital assistants (PDA), digital high-definition
television (HDTV), digital video recorders, sealed Information appliances, and basically
any other multimedia device will greatly benefit from high-bandwidth, high-speed
RAMBUS chip interfaces. In order to capture a significant share of those markets and to
solidify its technology lead in the memory interface market, further improvements to its
RSL (RAMBUS Signaling Level) are needed.
QRSL: The Next Generation June 2000 was a very important month for RAMBUS in
many aspects. The pinnacle occurred on June 16th. Not only did Toshiba agree to
license RAMBUS' non-RDRAM patents for its SDRAM, DDR SDRAM, and DDR
FCRAM, but it also became the first licensee of QRSL, which was also announced that
day at the VLSI Circuits Symposium. QRSL, which stands for Quad RAMBUS
Signaling Levels, is the next phase of chip interface technology. It has a native transfer
rate of 1.6 Gbps, twice that of its current generation technology. By using its patented
DDR (double data rate) technology combined with its newly patented MLS (multi-level
signaling) technology, which allows four bits to be transferred per clock cycle,
RAMBUS' new chip interface can yield 12.8GB/sec rates from a commodity 64-bit bus.
However impressive this sounds, the technology has only been announced; shippable
products will not be available for some time. RAMBUS is in the process of melding this
next-gen technology with its proprietary memory interface, which will formulate
ERDRAM (E = enhanced), or some variant. I don't expect this PC memory technology
to be available on the open market for at least 2-3 years, but it does provide a glimpse of
the scalability of RAMBUS technology.
Anticompetitive Claims! When RAMBUS sued Hitachi on January 18, 2000, Hitachi
responded by claiming that, among other things, RAMBUS violated Sections 1 and 2 of
the Sherman Antitrust Act of 1890. With the settlement between the two companies on
June 22, 2000, all charges were dropped. Unfortunately, (other than Toshiba) no other
memory manufacturers or chipset producers have signed similar deals with RAMBUS. It
has come to the surface recently that select memory manufacturers are conspiring to
submit a claim to the FTC in regards to a supposed antitrust monopoly. Based on the
fact that Hitachi used the same defense, I think it is highly unlikely the courts will rule in
favor of the suggesting memory makers. RAMBUS was granted several fundamental
patents relating to the use and control of synchronous memory devices, and after all, a
patent is a LEGAL MONOPOLY for the period of 17 years. If this claim is carried out,
a short-term downtick will occur. However, if RAMBUS wins or settles, it will have an
immediate impact because a large number of manufacturers will sign at once. Either way,
it is pure speculation, but important to consider.
The Bottom Line I mentioned scalability in regards to RAMBUS technology a few topics
back, but now I want to shift to the business model. Find me a better business model and
I'll show you the smartest businessman in the world! There simply is no better business
model than RAMBUS'. The company is a design house that doesn't manufacture
anything, therefore keeping capital expenditures low, while maintaining zero inventories.
Other manufacturers, who license the technology, produce their products. Their
marketing, and specifically, advertising is done by consumer products companies. And
best of all, their mission statement calls for a minimum of 80% (eighty percent) of
employees, at all times, to be engineers, thereby, keeping the BIG B (Bureaucracy) in
check! What this all equates to is this: long-term net profit (NPAT) margins should reach
60%+!!! Perhaps Mark Edelstone of Morgan Stanley Dean Witter (MSDW) said it best
in his revised research report dated June 19, 2000, “Investors should keep in mind that
RAMBUS should become one of the most profitable companies in the world!”
Disclaimer: This release contains forward-looking statements regarding financial results
for future periods. Actual results could differ materially. All assumptions are my own.
They are influenced by many factors and people including, but not limited to RAMBUS
Investor Relations, Geoff Tate, Gary Harmon, Mark Edelstone, Dataquest, InStat,
Samsung, Semico, and others. As Baz Luhrman said in his commencement speech at
MIT in 1997, “my advice has no basis more reliable than my own meandering
experience;” this holds true for me as well. Please take my comments, analysis, and
predictions for what they are: free advice! If you agree with my report, take a portion of
your money and invest in RMBS. As with ANY investment, please make sure it is one
you feel comfortable with that if you lost all of the money, you would not do anything
drastic. Remember, there is a risk-reward ratio to everything in life…investments should
be treated the same. If you don't agree or do but don't feel comfortable investing, do not
throw this report out, but instead, store it in a safe place and when you're cleaning out
that drawer in a couple years, pick it up and see how accurate I was. Good luck to all, I
wish only the best!
|
