From: Fred Hager...
By fred hager fredhager.com
Rambus' stock has been under pressure ever since the company met estimates and posted
earnings of $.04 on Tuesday. We anticipated the numbers would be higher, and we were wrong.
Earnings are always difficult to predict, and while not many in the industry will admit it, they are
clearly manipulated by the company posting them. It is often a no win situation for companies, as
Rambus has showcased. Last quarter the company exceeded estimates and was punished for
their accounting. This quarter they remained conservative, and were punished for not exceeding
estimates. Rambus could have posted higher numbers, but they chose not to, in our opinion. A
crazy notion, some might say, but we don't think so. Rambus chose not to include the Hitachi and
Toshiba settlement income this quarter, although they could have. We believe the company did
not want to release the information about settlement income, considering negotiations with many
other DRAM manufacturers is still taking place. Effectively, Rambus let the royalty income dictate
the quarter's earnings with some deferred revenue, and while the last three months have not
proven explosive for RDRAM distribution, the technology is being shipped in volume, and the
volume are increasing.
We've stated our admiration for Rambus' management in the past, and we respect their
conservative approach to their business. In hindsight, it is easy to see why the company would not
want to blow out estimates, as there is no reason to try to gloss over anything not relevant to the
actual growth from royalties. While the stock may continue to experience pressure in the short
term, it does not reflect a change in our outlook for the company. It is worth watching Rambus
closely for the next three months, as the stock is more and more attractive of a buy with every
point it goes down.
Rambus is not only among our favorite stocks, it is also the security of choice of many journalists
and day traders. We prefer to look long-term with the investment, but in the short-term, the noise
from journalists will always spark the traders to throw the stock around. Lately, there has been a
bevy of news surrounding Rambus, Intel, Rambus' technology performance, anti-trust issues, Via,
and host of other controversies. We at Fredhager.com stick to the facts when analyzing the news,
and we suggest to subscribers to do the same. Speculations and assumptions are often wrong, as
we illustrated with our own assumption on earnings this week. It is easy to become fearful of an
investment when you've seen it grow significantly, or the stock has moved up a lot based on
developments and there is all of the sudden new news that jeopardizes the company's potential to
fulfill the original expectations. Qualcomm is a good example of this, and Rambus may become
subject to such speculation.
Anti-Trust Issues Numerous articles this week discussed Intel's involvement in the industry's
efforts to show that Rambus is violating anti-trust laws with their efforts to claim ownership of
core technologies associated with SDRAM and RDRAM. While we do not believe Intel is, or
will be involved in any action against Rambus, it does not surprise us that they attended a meeting
about the issue. Intel, as any company, will always attend meetings they are invited to, to find out
what is at stake. While one article pointed to Intel's involvement, it in no way indicates that Intel is
participating in any actions against Rambus. Intel's participation is an assumption by the author,
but no action has been taken.
The most significant, and in our estimation, the most positive indicator from all the anti-trust talk is
that the industry is effectively admitting that they can not break the Rambus patents, and an
anti-trust campaign is the contingency plan. The industry is indicating that the Rambus patents are
valid and unbreakable and they think they have no chance of winning in the patent courts.
Rambus may be a monopoly concerning their own patents, but that’s what patents are all about -
a legal monopoly. The industry is not attacking Rambus, Rambus is attacking the industry, much
like Rambus went after Hitachi and prevailed. The industry is merely trying to avoid paying
royalties on Rambus IP that they have been using for so long that they think it is public domain.
The real problem is that it took nearly ten years from the date that Rambus applied for their
patents before they were issued. The application date is the one that counts for ownership, and
that was in 1990, two years before SDRAM memory started being used in PC's. Rambus could
not enforce their patents until the issue date, 1999.
While the press has spun much of this meeting talk lately into an anti-trust forum against Rambus,
it is also possible that the industry is working towards an agreement with Rambus, and they
wanted Intel's take on the whole situation. If such a settlement is reached, Rambus will clearly be
the standard in the industry.
Via, DDR, and the Pentium 4 A central issue that is in the balance in all of these discussions is the
future of DDR. We learned this week that Via is going ahead and planning Pentium 4 chipsets that
are compatible with DDR (without the express written consent of Intel.) When asked what they
would expect from Intel if Via tried this again, a Via official stated that history has a funny way of
repeating itself. For those who are not familiar, Via created low-end chipsets to fill a gap in supply
last year. Intel sued and put an end to the illegal action, but not before Via captured a large share
of the market. Now, Via is stating they are going to do the same thing with the Pentium 4, and
thus risk being sued again by Intel. What has not been discussed is the action Rambus will take
with Via, despite Via's comments that they will fight Rambus' patent claims. While Intel did not go
after OEMs who were using Via chipsets last year, Rambus can and likely will do so to protect
their income potential. Intel, who was losing market share to AMD last year likely did not want to
strain relations with OEMs, and let the issue lie. Rambus has less to lose by imposing their
restrictions on OEMs. This is what happened with Sega and Hitachi, and the issue was settled
when Hitachi came to terms with Rambus.
While AMD captured a big portion of the high-end market last year, due to problems Intel had
with the 820 delays and availability issues, it is possible that Rambus will enable Intel to regain the
high-end crown with the P4. Intel stated in their conference call that they were comfortable with
the availability of the P4 processors, and they would have plenty on hand for the launch. Due to
the performance benefits of Rambus in the high-end, Intel will conceivably take the lead, forcing
AMD to play catch-up with whatever DDR is available (the latest on DDR availability is January).
Intel will have plenty of time to capture market and reduce the costs. AMD, who we know has
Rambus engineers in house, could be in a tight spot without Rambus technology of some sort.
Any announcement from AMD and Rambus, or Via and Rambus will move the stock. In the
meantime, many companies are going to delay with Rambus as long as they can, while they wait
for DDR to become available. We would anticipate that Rambus will start to take action with
some of these companies, if settlements are not reached.
Performance of RDRAM vs. SDRAM - More On The Intel Benchmarks
In addition to Fredhager.com Staff researcher, Eric Jhonsa's discussion of the RDRAM vs.
SDRAM performance issue last week, it is worth commenting further on the topic. Recently,
benchmarks were published indicating a PC with SDRAM outperformed (by a very small margin)
a similarly configured PC with Rambus RDRAM.
This performance discrepancy occurred due to two reasons: A. The benchmarks and methods
used were the ones that have always been used; they were running by themselves with nothing
else running, whereas in the real world, people run more than one application. (Have you ever
been working with a spreadsheet while a file is downloading from the Internet?)
When a single program is running by itself, most of the time the data it needs is found in the high
speed memory cache, but when another program is also running, it will cause an access to the
main memory to load the data it will need into the cache. Then, when the first program tries again
to access data, it will have to do the same thing - access main memory. Thus, when more than
one program is running at the same time, main memory speed has much more of an impact on
performance. While some journalists still try, not many respectable industry folks will disagree
with the notion that Rambus RDRAM memory is much faster than SDRAM memory.
Real world benchmarks that emulate how people really use their PC's will have to be run to show
the speed superiority of Rambus RDRAM.
B. They do not require large amounts of data. Rambus shines in programs that require huge
amounts of data that must be constantly fed to and from main memory, such as video streaming
and voice recognition (these type applications are going to become more and more in demand as
they improve because of the speed increases in PC's). For more on this see last week's update.
As stated, Intel's new Pentium 4 is on schedule and Intel has confirmed that the PC desktops will
ship only with Rambus RDRAM memory. The new bus runs at 400 MHZ matching Rambus
RDRAM which already runs at 400 MHZ. We have read that it will start shipping in the fourth
quarter and the slowest one will be 1.3GHZ. In their conference call, Intel was asked about
Rambus, and the company officials stated that they felt Rambus was the best memory solution
available. Also, as Intel has stated in the past, they will always look at alternative technologies.
These comments were circulated as anti-Rambus, but it is nothing the company has not said
before. Intel has not changed their position with Rambus, in our opinion, and they are as behind
the technology now more than ever. One thing was noticeable during the conference call, and that
is more questions than ever before were asked about Rambus.
Rambus stated in their conference call that Dell is launching a $1500 Rambus computer, which
indicates freed up supply. In Playstation 2 news, the hype surrounding the device's stateside
arrival has prompted some retailers to pre-sell the devices for an extra $100, or a 30+%
mark-up.
In closing, we feel that Rambus will eventually sign SDRAM and DDR licensing agreements with
most of the memory manufacturers, and some of these licensing contracts will be announced in the
coming months. In the meantime, as usual, there will be negative articles on Rambus as the
memory manufacturers try to avoid the inevitable. These articles, some of which will be totally
false, will scare some investors who will make the mistake of selling, and cause the volatility
always associated with Rambus.
It is difficult to predict what will happen in the short-term with Rambus. We believe by year-end,
the stock will be worth $200, and our long-term target stands at $500. Some events that could
affect the stock positively short-term are: announcements of settlements on SDRAM and DDR
with memory manufacturers, Micron beginning production or licensing agreements, Intel Pentium
4 news, and Sony Playstation II news, among others. |
