As one of the few analysts with a Strong Buy, I'd like to contribute the following thoughts:
1) When AMD has not had to contend with legal barriers, it has been an effective competitor (e.g., 386 and 486). Of course, those were windows of opportunity created by latching onto Intel's intellectual property.
2) When it was initially on its own for device development, AMD fell flat on its face (the K5).
3) With the (admittedly very costly) acquisition of NextGen, AMD became the owner of some impressive microprocessor design skills. From a decided disadvantage with the initial K-6 offerings, AMD has closed what appears to be all of the gap at the low to mid-range of the MPU market.
4) If the K-7 does what it is designed to, the higher end of the desktop market (including workstations) comes within AMD's reach.
5) If some of what I've seen comes to fruition within the next several months (which I believe), third party offerings will allow AMD to participate in the SMP server market with the K-7 at no disadvantage in performance.
6) Given Intel's Xeon pricing, which charges exorbitant prices for cache memory (by my estimates, Intel is getting about $1,000 for a custom SRAM chip that costs them less than $50), the entrance of the K-7 into the server market would place substantial amounts of Intel profit at risk.
7) In light of AMD's share gains at the low end to date and the likely prospects for future gains, does it really make sense that Intel's market capitalization has been 85-100 times that of Intel? Probably not.
8) Two qualitative issues:
8a) Intel has been, is now and is likely to remain a vastly superior company to AMD by almost any measure (but the measure can in fact become reduced, which leads to a belief that AMD could well become a superior investment); and
8b) Jerry Sanders is not a buffoon. How anyone who has been able to muster a credible challenge to one of the greatest enterprises ever created gets dismissed so easily is beyond me. Jerry's style is radically different than my own, but I find it inconceivable that people disparage the leader of an organization that produces literally millions of working MPUs annually that each incorporate millions of transistors. If these things were so easy to build, everybody would be doing it. Even Intel, with the most impressive design and manufacturing groups ever assembled in the electronics business, suffers significant delays (e.g., the various Xeon offerings and the Merced). These companies are incredibly impressive and deserve our awe (which I can state even while holding that Intel's segmentation strategy borders on blunder).
I'm a little leery of offering this commentary due to my position at Needham. So let me state that on SI, these views represent my own position and should not be viewed as official Needham investment research. I feel precluded from offering estimates or recommendations, but much of the above is merely a recitation of fact. I trust that readers will accept it in the spirit with which it is offered - Tad LaFountain |
