Carl,
Thanks for taking the time to type out your long reply. I agree with about half of it and think that the other half is a little too extreme. Maybe you are like Dan3 and respond to the pro-Rambus bias of this thread by trying to center it all by yourself. But, in any case, please see my replies to your post below inside << >> symbols.
--KZNerd
--------------------------------------------------
Hi KZNerd; The reason INTC went with RMBS are clear to me, having watched companies do this sort of thing from front row seats.
The sad fact is that when management gives technical instructions to their engineering staff they have to rely on luck rather than skill. What management should stick to doing, is making sure things are running smoothly, trying to find quality employees, and making sure that the employees they have are kept happy.
<< I find it very hard to believe that management dictated the Rambus decision to engineering. Intel is the type of company that pushes decisions down to the lowest possible level. (At least they used to.) I think it is more likely that technical people made the decision and then sold it to management. The engineers liked the idea because it made for a cleaner platform architecture since graphics memory could be unified with main memory. Also, it allowed for future wizzy things like better multimedia processing and solved the granularity problem. I expect that the decision was driven by Gelsinger although maybe you consider him to be management rather than an engineer.>>
What happens in fact, is that management gets it into their heads that they understand the technical issues as well as their underlings, and, in addition, understand other issues that the engineers are too low down in the organization to see.
In the case of the INTC/RMBS deal, upper INTC management bought the bill of goods from RMBS. This means that they went out and ate some expensive dinners with RMBS management, and sat through some slide shows that glossed over the real issues.
<< I don't buy the dinner part but I think that at the time of the decision the real issues and difficulties weren't well understood. They underestimated how hard it would be. >>
I am sure that the big warrant deal that RMBS signed didn't hurt with the INTC brass. But warrants are not an engineering solution, they are just crapper paper as far as actually getting something running. INTC got away from the engineering issues, and into the stock promotion issues promoted by upper management.
<< Nope, can't buy this. >>
The result has been a disaster for INTC. If they haven't fired anyone over there for this fiasco I would be quite surprised.
<< It is on the verge of being a disaster but there may still be time for Intel to save the situation with a diving catch. More on this later. >>
AMD is going to kick their butts over the next year, by having high performance processors for sale that are not tied to the RMBS albatross. For 2 years, this has been a major disaster in the making for INTC, and it won't be finished until a year or two from now.
<< Well, maybe. But don't forget that Intel may still have a bunch of tricks up their sleave. >>
The reason for moving into RMBS so quickly is management greed, plus the natural human tendency to believe what they want to believe. RMBS promised that their technology would be a lot easier to adapt than actually turned out to be the case, and INTC believed them. As a result, INTC has been forced to continue to dump money down this rat hole, even to the point of funding memory makers for the cost of equipment required to make RDRAMs. This is the stuff of desperation.
<< Well, the dump money part isn't true. Intel will make a bundle off the Micron investment and probably off the other investments too. But I am stumped over why Intel allowed this transition to get ahead of itself. I imagine that when the Merced deal with HP was made Intel resolved to have a huge overlap between IA32 and IA64. No one would dare suggest phasing out IA32 before IA64 is already kicking butt (assuming this ever happens). So why didn't management insist on having grade A support for both sdram and drdram until drdram was proven to be totally in place? I just can't understand how the current plan was ever approved. It's just baffling. >>
The sad fact is that even if RDRAM had been ready to go for 2H99 there would still have been massive allocation and pricing problems. Those alone would have given AMD a big step up. But production problems and delays make it even worse. INTC is hurting bad, and so will be the box makers that have avoided dealing with AMD. The pains involved in this have not yet showed up in the share prices. AMD is still quite cheap. INTC has not lost much high end market share yet, but it is now inevitable.
<< AMD's problem is that they can't deliver very many Athlons even if Coppermine is held up due to lack of drdrams. They're only expected to ship a max of 1MU in Q4 and that is not enough to change much. >>
The CPU play of the next 12 months is short INTC, long AMD in my guess, and this is due to the rambus problem.
<< That would be pretty darn risky in my view. Instead I would pay with AMD out of the money calls (I already own some) and consider INtel out of the money puts. This way you can't lose much if you're wrong. >>
As to what this means for RMBS, I do not know. I do know that RMBS major production continues to be delayed, and their major customer is replacing/delaying their product, while the stock price is still quite high.
<< I don't know either. I am long rmbs having bought back in Thur morning near the bottom after that -10 point gap down at the open. Right now I'm kind of wishy that I had sold when it got back up to 95. But it is really hard to figure out this stock. >>
A factor which is probably hard for most people to fathom is just how much the DRAM makers hate RMBS. They really don't like to have to pay royalties. They feel that INTC forced this stuff onto them, and now they are wasting money trying to put out product, while INTC delays its use.
<< You are right-on here. And I expect that this hatred means that the dram makers are not about to break their butts to bail out Intel and Rambus. >>
My guess is that the DRAM makers would love to get rid of rambus forever, so they would naturally prefer to support alternative technologies. With INTC forcing the industry to go to rambus, this was not possible. But with INTC wavering, and the competition putting out fast new CPUs that don't need rambus, you can bet that the DRAM makers are shifting production out of rambus stuff, and probably pushing their development of DDR.
<< Of course, don't forget that if drdrams are in short supply then the price can go even higher and dram makers will put their hatred on hold if they can make some bucks. >>
Naturally, none of this is going to get through to the individual investor, or even the analysts. Rambus's great strength is stock promotion, not engineering. Their competitors really don't care about press releases. I should explain about this, as it is not going to be obvious as to what is going on.
Rambus exists primarily to sell stock. To that end, the company puts out regular press releases. Press releases are sent out to the general public, the public that is in the market for stocks, not the market for DRAM technology.
<< Are you getting a little too cynical here? But it is true that Rambus is shameless in manipulating the public. But they do such a skillfull job of it that most people have no clue that they are being manipulated. >>
The thing to remember is that DRAM makers do not sell product to the general public. For this reason, they are not real big on wasting their time on public relations. There are only a few major customers for new DRAM technologies, and you don't have to put out press releases to contact them. You can do it by picking up the phone and calling up a handful of design engineers at a handful of chip designing firms. The loss in design wins for rambus is not going to be visible to the general investing public, there is no reason for it to be visible, no one is trying to sell the investing public stock based on DRAM technology except RMBS. The sad fact is that the analyst community has as little visibilty into the engineering design community as the general public does, and the same applies to the institutional investors.
<< I have a feeling that some of the big funds have some pretty good visibility. And they are swinging the stock around on days when there doesn't seem to be any news. But these guys don't talk about it. >>
So there is no reason why RMBS's stock price has to be tied to future prospects for their technology. The people buying and selling this stock probably know more about the practices of the Yanomamo than that of the designers who will choose to specify one or another memory technology into their CPU designs. Consequently, I would say that RMBS will continue to be a great trading stock until it is obvious that INTC is in trouble, and that they are losing market share to AMD. My guess is that this will be obvious about six months from now.
<< Well, that is one scenario. But maybe Intel can still provide a chip set solution for sdram to take the pressure off the drdram shortages. Maybe BX (although they are already short) or 810e or something that is already available can be used with Coppermine. Maybe these deals are being made as we speak. There is just too much at stake for Intel to be caught flat here. I expect that the box makers have been bitching so loud to Intel that Intel will launch a new "Operation Crush" to escape the trap that they've laid for themselves. After all this is still Intel even though they seem to have forgotten that momentarily. Of course there is the scenario where Intel pulls sdram chip sets out of their butts just in time and Coppermine ships mostly with sdram but with some drdrams and the rare drdram systems get the reputation of being the really cool machines. It could happen! >>
-- Carl |
