Replies to post #20306 on Advanced Micro Devices Inc (AMD)

[I_banker]

Chipguy, as an Intel or AMD investor, the relevant figure is not how much the systems cost, but rather, how much the processors sold for. Intel doesn't benefit differently if HP sells a $2 million dollar system or a $20 system.

What matters is how much revenue the chip is creating for the manufacturer, as this will determine its profitability. In the case of Itanium, it generates more profit per chip for Intel than Opteron does per chip for AMD. Using this metric would be a meaningful comparison, but is a magnitude different in the end results as compared to system price comparisons. Using system prices is irrelevant unless you happen to be a HP/IBM/Sun/et. al investor.

[CombJelly]

"And how many Opterons would it take in the form of
IBM e325 econo-boxes to match the ~$2m of hardware
revenue from those 8 systems powered by 54+ I2s?"

I dunno. I don't really see where it is a badge of pride to point out a sale where Intel supplies but a small fraction of the overall value. Heck, the company that made the RAM in that system probably had a larger percentage than Intel did...

Why is it relevant how much money SGI or IBM sells their systems for? Last time I looked, the name of this forum is "Advanced Micro Devices", not SGI, IBM, Cray or even Sun. While it might be interesting to see a post about the larger sales, a post about a single system is pretty dull stuff.

Besides, it's fun to plink at subzero...

[yourbankruptcy]

chipguy, you can repeat your words again, but replace Opteron with Pentium, IA64 by Sparc and 3Q by 1995:

Pentium can only press its nose against
the glass and look in at Sparc's lucrative market
segments

The history repeats itself.

[SilentBob]

Chipguy,

Given that IT budgets are only just now beginning to recover, I cannot see how the fact that a customer has to spend $1m to get a small number of systems containing 22 - 32 processors is great news for Intel. IMO there are few companies willing or able to make that level of IT investment in 2004. I would expect most companies to buy their compute resources in smaller bites for the next 18 months or so.