Intel........
I'd like to thank Saukie, Merlin, Slacker, Tinker, BB, Thomas, and others for the thought-provoking discussion of Intel's gorilla/king nature. While Intel's genes may seem obvious to some, since the playing field is dynamic, and because of the addition of new participants here on the Thread, I think it is a very worthwhile exercise to periodically go over the in's and out's of TRFM.
I especially like Slacker's response related to Intel's "control over the CPU's architecture, rather than specific IPR".
While all of you were discussing Intel yesterday, I was away, but had PM'd Tenchusatsu from the Intel (& Rambus threads). Ten works for Intel, where he is a Chipset Validation Engineer. I posed the question of Gorilla or King to him, and asked him about IPR or control over architecture yesterday AM. He has stated in his PM to me that I can post it here now to share with the Thread:
Question by Apollo:
Does Intel’s microprocessors represent open proprietary architecture, and hence is this reason for Intel’s 80% market share. If so, what specifically for non-techies is it about the MPU by Intel that is so special? Or alternatively, is it that Intel is a King, a company with so much going for it that it can make and sell a commoditized chip and execute so well as to maintain high margins and blow away the competition?
Ten's response:
"In my opinion, Intel is now a King.
For all practical purposes, the x86 architecture is now an open architecture that Intel can no longer keep to itself like Microsoft can with Windows. The only thing Intel can do is add enhancements to their new x86 processors, like SSE2 for the Pentium 4, but these enhancements won't be "must-have." Worse yet, AMD will be implementing these SSE2 enhancements in their Hammer processors beginning in 2002. (How AMD can do this without Intel's blessing is beyond me.) But ultimately, processors have advanced so far beyond the needs of software that pushing the performance envelope (as Intel is so good at) no longer matters much anymore. Sooner or later, bottom-feeder companies like Via will be coming in from below with cheaper, lower-margin products which are slower in comparison, but at least "fast enough" for most users.
However, one thing that Intel does have going for it is its huge manufacturing capacity. This helps Intel to drive the economies of huge volume, thereby crushing the competition with lower costs and higher leverage in gaining OEM design wins. Another thing Intel has going for it is the partnerships with these various OEMs in helping to deliver the total PC system, not just the processor. Intel also helps out OEMs by selling their own motherboards and chipsets, laying out new open standards for PCs, designing software APIs and compilers, and constantly asking OEMs for end-user feedback. In comparison, notice how AMD has stumbled in Athlon infrastructure support, which continues to hobble Athlon's acceptance in new markets besides low-margin retail.
Via, on the other hand, has the potential to really challenge Intel in a commoditized market. And there are a handful of smaller competitors ready to make their own x86 processors specifically tailored to niche markets. True, the chances of success for these smaller competitors are slim (witness the demise of Cyrix, IDT, and Rise), but the opportunity is still there.
However, up to now I've been talking about x86. With Itanium (also known as IA-64, but Intel wants to move away from that designator in favor of Itanium), Intel has the potential to become a Gorilla in the high-end server and workstation market. Currently, these markets are dominated by vertical solution providers like Sun or IBM. But Intel wants to transform these markets to a horizontal model, where the prices are lower by an entire order of magnitude, and where Intel becomes the owner of an open proprietary architecture. Although the first Itanium generation is late, Intel has so much support from industry partners like HP, Compaq, IBM, NEC, Bull, etc., that success is pretty much guaranteed once Itanium is finally released. (Heck, Itanium is currently released in a pilot program right now, so the wheels are already turning.)
Whew, I said a mouthful there. Hope you understood everything I was trying to say. If not, I'll be glad to clarify anything you found unclear.
(additional comments by 2nd PM) I guess a King's huge market share helps a King to be self-sustaining. In Intel's case, what would make it King would be the huge manufacturing capacity, combined with excellent execution on the part of the guys who squeeze more speed out of a given process. This capacity brings economies of volume to Intel's bottom line, allowing Intel to continue being King."
Tenchusatsu
___________________________
To finish, there are 2 things that I would reinforce:
1. Among many of Intel's strengths, the FAB capacity may be its greatest, allowing for outexecution.
2. Itanium has the potential, as a new architecture and with a new value chain, to induce a huge tornado. Remember, that Intel has recently stated that the world only has something like 1-2% of the servers it will need in the next few years; again, this need is Internet-driven.
Apollo |
