RE RMBS
Your write up has concentrated on the relationship between RMBS and INTC. I have not heard anything, from you or anyone else, about AMD and its roadmaps. What will AMD use down the road? They are currently using SDRAM, and I have not heard of any projections that they will use RDRAM.
If that is the case, INTC maybe really shooting themselves in the foot. Not only their chips are slower and more expensive than AMD's, they also use more expensive memory (once they get to P4s). This could cause them to lose a large share of the PC/workstation market. Since they already complain about Rambus collecting toll money from memory manufacturers, they see the memory price as a negative for their own plans.
Anyone knows for sure about AMD's plans?
AMD has licensed RDRAM, and RDAM is in negotiations to get licensing fees for all DRAM controllers being used by AMD (these DRAM controllers actually pay about 2x the % as do DRAM so very lucrative). But AMD's plans for RDRAM can be summed up with this paraphrase, "RDRAM is the memory of the future, little doubt of that, but that is in a year or two or three." So that is AMD's plans on that front.
In regard to INTC being at a competitive disadvantage I see that as an erroneous interpretation IMHO. It is short-term. (1) AMD will need a high speed memory, DDR is still not commercially viable and DDR when and if it comes out in mass quantities will be more expensive than RDRAM. It is the law of supply and demand and the law of the learning curve; (2) the latest reviews by game sites and by sites using INTC's latest chips with RDRAM almost to a man (or woman), excepting Tom's Hardware of course, have concluded, again paraphrased, "golly, this RDRAM stuff really does kick butt." So it is simply not the case that RDRAM is slower than SDRAM or that INTC's latest chips are slower. IT depends on its use. If you want to multi-task, play games, stream-media, nothing touches the new INTC products. I have seen such benchmarks giving INTC an over 30%+ speed advantage depending on the application.
As for the PC workstation, and the chances of losing market share there, over 70% of workstations shipped (I believe this is in the last 6 months) have been shipped with RDRAM. That seems to be a growing domination of the segment.
As for the premium price of memory, yes, RDRAM is more expensive, but the price has come way, way, down since it was first introduced. In addition, the premium of RDRAM over SDRAM is predicted by Rambus, and I believe I saw Intel with this quote (but don't count on the INTC part as this is from memory) that RDRAM will be within 20% of SDRAM prices by the end of 2001 and Samsung stating within 15% by 2002.
These problems are overstated, and very easy to get confused about. I have looked into each and every bit of what is called FUD because one gets fearful time and time again. But almost to a last news quote these problems turn out to be non-problems, or temporary at best.
Finally, what I see is Intel playing Gorilla. Intel has control of its architecture. AMD has finally caught on to the game and has equalled INTC in its ability (after years and years of trying) to produce processors that work with SDRAM. They have also produced the first 1 Ghz+ chips.
Like a gorilla, INTC has decided to play with the architecture. They moved the game from SDRAM to RDRAM and have taken pre-emptive steps to get RDRAM on the market and to learn how to work with it upfront. The transition to RDRAM for INTC has been the most difficult product transition INTC has ever made. How quickly do you think AMD will be able to adapt RDRAM, particularly at 2 Ghz+ speeds, when AMD is planning to move beyond DDR? Much more difficult of a time than INTC had transitioning at the lower speeds.
What I see from the foes of RDRAM is typical monkey behavior. The memory makers and AMD were happily content with the SDRAM architecture. They have used it for years, are heavily invested in it, and it suits their purposes to stifle any innovation that disrupts their easy lives. It is only the industry leaders INTC, Samsung and DELL, with a few supporters like the Japanese MM like NEC, who supported RDRAM. Everyone else was seemingly against it. Why? Because it puts them at a competitive disadvantage.
As an example, Samsung has more money and probably more skill than any of the other MM (memory makers). They can most easily make the necessary investments and transition. Result: They are currently earning extraordinary profits on RDRAM in what is otherwise usually a price driven commodity market. The other MMs had neither the capital nor talent to so quickly move to the next innovation, and remain competitive with Samsung.
INTC gains similar leverage over AMD. Gorillas don't think out about how the next quarter will be. They plan long-term. It is my bet, my prediction, that 12-18 months out, AMD will be under the INTC proverbial thumb again and once again behaving as a monkey. They have to start all over again with RDRAM.
One of the major factors that pushed me into RDRAM this January was analyzing these exact relationships. It was astounding to me how only the best and brightest of the industry were investing in this RDRAM technology, and investing big time. Samsung, if memory serves, built an entire new fab for $4 billion + to accommodate their ramped up RDRAM production. Dell put up with one bad quarter and based nearly their entire workstation line on the success of RDRAM. INTC has practically bet the company on RDRAM. They are so supportive that at this point in time that are offering to pay the price differential between SDRAM and RDRAM to the MM's who produce RDRAM for INTC chips.
As a quote "advanced gorilla game investor" it was these collateral gorilla signs in the value chain, of the visionaries moving the industry forward and putting their money where their ideas were big-time that allowed me to move into Rambus in what I thought was a rather risk-free way in what most others saw as a risky stock at the time.
Have things changed? Yeah, there is this litigation. Litigation is unpredictable. The only claim that I see that has merit is this JEDEC claim. It is a cognizable legal argument that is not attempting to undo Rambus patents. Instead it is basing the claim that Rambus abused the open-standards process and knowingly (or negligently, or otherwise, I don't know the exact legal standard not having read the FTC opinion that the suit is based on) allowed Rambus patented technology to become part and parcel of an otherwise open standard.
From the evidence I've seen Rambus disclosed their synchronous technology sometime around 1992 or 1993 (ie, synchronous, SDRAM, and disclosed other patents around 1995 in their resignation letter. I understand the basis of this legal argument but I think RMBS made sufficient disclosure and that JEDEC was negligent in not following up, nor has JEDEC made any claims that they could have drafted a standard that would not have violated Rambus' patents, something else that the plaintiffs will have to do. And they will have to do so to over 90 patents (although they don't all cover SDRAM, I think most of them do cover DDR however - but this is conjecture and not absolute fact on this latter part).
So I guess what I'm saying is that in Rambus I see a much more powerful gorilla being built here than most respondents on this board have made out. I also have looked at most of the negative evidence against Rambus and have concluded that the very vast majority of it, if not all of it is no more accurate than claims from the telecommunication industry that CDMA was impossible to produce commercially, or that CDMA performs less well than does GSM or TDMA, etc. I'm sure long-term QCOM die hards can attest to these sort of rumors. Long-term Bussers have lived through the same.
So I am very confident in Rambus' gorilla stature.
What I am only somewhat confident in, and I say this with any litigation, is the outcome of the litigation. If not for this litigation I would say that Rambus has the DRAM and the DRAM controller industry (DRAM controllers possibly being even more lucrative than the DRAM itself (something which was not included in the $72 billion market by the end of 2002 stat I brought up) wrapped up without threat of competition, at least not on the foreseeable horizon.
Tinker
Yeah, I'm very high on Rambus. But as to the litigation I will only state that from the incomplete picture I have ascertained, I believe the case against Rambus is weak, with the above JEDEC cause of action being really the only possibly dangerous count, and as for that I have seen evidence supporting Rambus' position in regard as well as evidence that NEC, and something like 19 other parties have already settled with RMBS in regards to SDRAM, DDR, and RDRAM, without venturing into court, or otherwise settling their lawsuit.
So post-law suit, RMBS a pure-bred, enormous pongoid, pre-litigation RMBS a gorilla in-waiting, but boy is this gorilla pumped with nostrils flaring. |
