Rambus controls the RDRAM architecture. Perpetuating it in the marketplace with a series of sustained innovations, should continue to extend the CAP for some time. Similar, in a way, to Windows 3.0, followed by Win 3.1, Win 95, Win 98, Win98se, and Win Millenium. Win XP will represent a discontinous innovation, as I understand it. I would hope that Rambus would be able to follow a similar path.
To digress for a moment: I guess I'd say WinXP is a discontinuous innovation (not so for Office XP, although it's a good product). The trick is to make new versions discontinuous enough to throw competitors off balance and to charge into new markets, but continuous enough to bring your value chain along with you, and thereby maintain your power. At any rate, I believe that it's not just hype when they say that Windows XP is the most important release since Windows 95.
Back to RMBS:
I think your paragraph gets at the crux of the issue, so let me tell you why I disagree. Well disagree may be too strong of a word. Here are my concerns:
1)
INTC is the Gorilla of PC hardware. If RMBS wants to alter the dominant memory standard (i.e. as MSFT changed it from Win 3.1 to Win95), they need INTC's approval. Essentially, the dominant memory standard is whatever INTC says it is. Fortunately for us RMBS shareholders, INTC has chosen RDRAM as the next-generation memory standard.
I believe that they did this for sound technical reasons, and in a sense they had little choice but to choose RDRAM. But I see no reason to believe that INTC will necessarily choose the next great thing to come out of RMBS for the next-next-generation standard (i.e. ~5 years out), other than because RMBS has the smartest engineers. Now, having a great engineering staff is definitely a sustainable advantage, and is necessary to be a Gorilla (pace the opinion of some Gorillas' detractors), but it is not in itself enough to be a Gorilla.
2) Related to this, I see no reason to believe that it will necessarily be RMBS that comes up with the next great memory standard. First of all, I think that the RMBS hawks have been correct in saying that the cost required to retrofit fabs for a new, profitable DRAM standard is small enough to warrant the investment, given the potential profits. That implies that N years from now, should (should, not when) a newer, better memory technology comes about (e.g. one which is more compatible for some reason with INTC's then-upcoming architecture), that technology will have a good chance to win. In other words, I don't believe that the switching costs for memory technology are Gorilla-like.
Second of all (and this is important), even if continuity with present-day RDRAM technology is important, it's still possible that another company will come up with the winning next-generation technology. For proof, look at W-CDMA. Now don't get me wrong; I'm not saying that W-CDMA is indeed superior to CDMA2000 (I don't know, and those who do on this thread seem pretty sure the answer is no). But the point is, this shows in technicolor detail how a competitor can hijack a standard. Yes, W-CDMA involves royalty payments to Q, and "W-RDRAM" would involve royalty payments to R, but there's still a substantial power shift. All of a sudden, you go from 100% architectural control to, say 50% architectural control (and you have to figure the newer 50% benefits its creator in someway)--which begs the question, was the original control 100% to begin with?
Note that this doomsday scenario is much, much easier to envision when your "control" is due to your ability to licence, rather than your huge market share in a particular product. In other words, you'll never see a successful "W-Windows". That's because Microsoft has enormous market share & brand power, which is resistant to erosion in the short-term. So even if a competitor manages to clone the Windows API, they're screwed. Due to effective segmentation, etc., they can get a small minority of the market to go along, say 20%. But then Microsoft just redefines the next version of "true Windows" in a way that undermines the competitor. Due to the inertial power of brand, the vast majority of the market will go along with Microsoft, and all of a sudden, the competitor isn't compatible with the dominant standard, even though they reliably cloned the "old" standard. Why did the new standard evolve in exactly the direction that Microsoft wanted? Because Microsoft doesn't only own the standard, it creates and sells the product that reifies the standard.
On the other hand, what does RMBS do to combat the (mythical) W-RDRAM? They can certainly come up with a new standard (call it RDRAM 2.0), and try to get it adopted. However, since RMBS doesn't actually sell any products, it can't rely on its huge customer base to force RDRAM 2.0 on the market, and thus undermine W-RDRAM. They have to basically hope that the memory manufacturers agree to push RDRAM 2.0 instead of W-RDRAM. That is, shall we say, an unpleasant situation for RMBS. Pinning their hopes on the good will of the memory manufacturers is an iffy proposition.
Now, let me step back a moment. I fully agree with Tinker et al. who say that RMBS looks to have a great market position in the near-term, but they aren't guaranteed to have any revenues once the patents run out. As I said, every two weeks I put new money into the market, and that's 100% RMBS. And of course any specific scenario I outline will sound far-fetched. My goal isn't to make an accurate prediction, but rather to present an allegory. RMBS may not fall, and if they do it may not be for the above reason, but it can happen. Because RMBS has the ability to suck profits out of its industry, but not any unique ability to push the industry in a new (advantageous to RMBS) direction, I call then a Leech, not a Gorilla. In essence, their plum position in today's (or, properly, tomorrow's) market is sure, but such a position in the day-after-tomorrow's (that's lendemain's to you, Boa) market is not guaranteed. Gorillas must continually be born again in order to saved; RMBS may only live once.
Once again, this is hardly a sell recommendation. It is my favorite stock. I should also say that RMBS appears to have strong management, although I'm not sure I'd say stellar. I think that TRFM severely understates the importance of management (see, for comparison, The Warren Buffett Way, by Hagstrom). Shrewd managers will surely work to avoid the above scenario by a) building a brand image, b) undermining memory manufactures, transforming them from powerful customers to a weak sales channel (see MSFT or INTC...), c) developing alliances with supportive OEMs, so as to ensure that new RMBS-friendly extensions of the standard will be adopted (It appears that they're doing this today. Samsung may be RMBS's--how can I say this politely--subordinate in much the same way that Dell is MSFT's & INTC's), and d) managing the hell out of the INTC relationship.
ardethan |
