Advanced Micro Devices

Format for Printing

Format for printing

Request Reprints


By duttermohl
February 26, 2004

Posts selected for this feature rarely stand alone. They are usually a part of an ongoing thread, and are out of context when presented here. The material should be read in that light. How are these posts selected? Click here to find out and nominate a post yourself!

I engaged in some panic buying today as I think the HP announcement has put AMD over the top in the investing public's eye. HP was not just filling a gap in their lineup until XeonCT comes, they seem to be fundamentally changing their server lineup with the announcement of a 4-P box and a long term strategic relationship with AMD. Looking at HP's 32-bit benchmarks, the Xeon market just shrank to those companies who would only buy Intel Inside. Still a large part of the market, but, one that is serviced best by Dell. With it's longer pipeline, Prescott will do worse on many of these same benchmarks and a 64-bit OS will help it less than 64-bits will help Opteron - adding data to a saturated data bus won't add performance.

The Itanium market just died. HP has tried to force Itanium down customer's throats for years, to little effect. Now, the very rational for EPIC - a high powered 64-bit RISC chip - will drive it into so small a niche that even massive transfusions of cash won't help it ever become profitable. I expect HP to bite the bullet and announce that they plan to port HP-Unix over to Opteron/XeonCT to "give customers a choice" and quietly abandon Itanium. Intel will keep Itanium on life-support for a couple of years by offering the kind of incentives that only a market leader can, but, the day after Craig Barrett retires, the plug will be pulled.

The HP, Sun, IBM triad puts upward pressure on the value of AMD's Opteron chips. AMD is likely to have trouble making enough chips to fill the orders for these three and all the tier two vendors who are selling Opterons (those Opteron start-ups aren't being bought up by the big three). This capacity constraint has nothing to do with manufacturing capacity, as Dresden is more than capable of covering all the needs of the entire server segment even with the large Sledgehammer die on 130nm (a little more than 10% of the market in units). Rather, AMD cannot abandon it's strategy to meet Intel on all battle fields to prevent Intel from subsidizing one product line with another. It doesn't mean AMD has to win all segments; they just have to field a competitive product.

AMD has already addressed this with a slide from their analysts' conference that showed two pyramids - one upside down, showing the relative R&D spending they planned for three segments- Desktop, formally the largest segment (base of the old pyramid) was now the smallest with Server R&D now the largest. (Mobile stayed about the same).

This puts AMD on the horns of a dilemma. Go for the short-term gain and long-term pain, or stay the course. Any attempt to take full advantage of the server market will crimp their market-share in the desktop/mobile segments - segments they will need to fill their 300mm/65nm in two short years (market share is easy to lose, and hard to regain). But, aggressive competition with Intel in the volume market, with Intel already aggressively converting to 300mm/90nm, will mean AMD has to give away silicon that otherwise could be very profitable. Like most choices in life, AMD is likely to compromise.

A successful conversion to 90nm will help a lot. And, AMD may let IBM fabricate their own Opterons, but I expect AMD's unit share to drop over the next few quarters as they allow their own ASPs in the high volume markets to drift upwards. (This is the real benefit of the HP announcement- AMD will now have the pricing flexibility that Intel has enjoyed so long.)

I do expect AMD to keep their fingers in the pie in all market segments, but they will compete aggressively only in server (high asp) and mobile (fast growth) allowing Intel a breather in the desktop space. While AMD might enjoy premium pricing in one or two "high end" segments, Intel is likely to increase market share in Desktops (AMD's traditional market niche) lose a little in mobile (chiefly DTR as the Pentium4-M disappears) and lose a lot (percentage wise, but still sell a lot of units) in the server space.

As Eachus says, AMD cannot stop Intel from making good money nor can Intel stop AMD - until AMD has their 300mm/65nm capacity on line. Then, unless Intel has substantially changed their line-up (and I expect them to, BTW, Intel is not stupid!) AMD will be able to provide the volume to "hurt" Intel (it ain't gonna happen, but we can dream!)

Buy under $15, sell at $20 (for now).


Become a Complete Fool
Join the best community on the web! Becoming a full member of the Fool Community is easy, takes just a minute, and is very inexpensive.