Patrick -
During the same quarter less than 500K servers were shipped.
You are looking at 'industry standard servers' (i.e. Intel-based). Intel already has something close to 100% of that business, by definition.
The rest of the high end market is a target which Intel architecture machines can target, both in IA32 today and IA64 in the future. This is not only High ASP and High margin, but the market as a whole is 4 times as large as the whole of the PC market in revenue terms. This is a huge opportunity for Intel. This is the reason for the large investments of both resource and executive mindshare with HP, IBM, Sun, and others.
The high-end is nowhere for Intel to escape to. It is much to small. A 10% loss of marketshare in the desktop market would have to be made up by a tripling or quadrupling in the server market
CPQ has had more than 60% of it's profit from servers for more than 3 years. So let's see, how many $150 parts do I have to sell to equal one $4000 part? and what about margins? In CPQ's case, with Desktop ASPs heading to $1K and margins near 20%, and server ASPs heading to $20K and margins near 35%, which way will management go?
Your analysis of the relative values in these markets does not square with the thinking of any of the major OEMs, all of whom see the desktop business as an increasingly commoditized space with little profit potential, and the high end as the inevitable next market.
Server shmerver, Intel is making its bread and butter in the desktop.
Desktops are an increasingly poor growth vehicle for everyone. CPQ figures that desktop revenue will be less than 20% of their total revenue within 2 years. It is already a small fraction for IBM and HP. Even Dell is predicting a shift to less than 50% desktop revenue within the next few years. And in terms of margin contribution the desktop segment will be in single digits maybe as soon as next year. Don't confuse unit volume with revenue or margin. |
