The fallacy in chipguy & wbmw's argument is that they present HP's replacement of PA-RISC, mips and Alpha as if it were new sales by a variety of companies. It is not. It is simply HP replacements for their own market.
Wrong. SGI sold about $100m of IPF based gear in Q4
even with Altix 350 and Prism still in early/pre product
ramp stage respectively. I expect SGI will at least double
its IPF based sales this year to over $600m even with the
Montecito still not playing a major factor due to its late in
the year release.
NEC, Bull, and Fujitsu have spent millions on chipsets to
implement large scale IPF servers and they wouldn't have
done that if they didn't intend to strongly push IPF into
their existing customer base as well as go after IBM and
Sun, the two remaining RISC players. Dell is not a major
IPF player yet but Montecito will bring a huge boost to
their two and four way 8870 based products, far more
than dual core Xeon will bring to their Xeon line, so IMO
one should keep an eye on Dell as an IPF dark horse.
That will reach its peak penetration in the current year and IPF's market penetration will be a flat line for several years thereafter. The IPF market will not grow or shrink for the forseeable future after 2005.
Yawn. HP has said that IPF sales will cross over 50% of
its high end system sales by year end and exceed 80%
of high end system sales by the end of 2006. Do the math.
Montecito production ramps in early 2006 and the 65 nm
Montvale follows up only a year behind Montecito.
Can you say n-i-t-c-h?
LOL, is that anything like niche? BTW, it is pronounced neesh.
Market Data 
Markets 
AI
