SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Technology Stocks : Intel Corporation (INTC) -- Ignore unavailable to you. Want to Upgrade?


To: Mary Cluney who wrote (140129)7/25/2001 8:48:46 PM
From: Windsock  Read Replies (1) | Respond to of 186894
 
Mary - Re:"First you start with a mixture of false assumptions and then you draw conclusions that fit your agenda and hope that people get tired of trying to unravel your false assumptions."

The PeckerHead Logic is a common distortion technique. Dan3 uses the same method. A series of distortions and circular logic to give the impression of a numerical proof. When you arrive at the end the conclusion appears logical and plausible. But the conclusion is total Bull$hit. It works like this.

Start with A -- an irrefutable fact

If A then B
If B then C
...
If (N-1) then N
If N then Not-A
Therefore, Not-A is clearly true.

QED



To: Mary Cluney who wrote (140129)7/26/2001 4:37:19 PM
From: pgerassi  Read Replies (1) | Respond to of 186894
 
Dear Mary:

You are still far away from 25% of IDC. It is closer to 2% than 25%. 13.5% is halfway. Thus, anything between 13.5% to 0% is closer to 2% than 25%. And anything above 13.5% is closer to 25%. Now simply take a look at Intel. The marginal cost to process a 0.18u 200mm wafer has been estimated by many here including people in the processing section to be about $2K. Taking the normal estimates of yield, packaging costs, etc. leads to a marginal cost of between $25 and $50 a P3 CPU. The cost to an Intel OEM averages to some where between $100 and $200 for that CPU. Thus the markup for MG&A, R&D, engineering, FABs and profit becomes around 4x the marginal cost. That is typical. The number comes up again and again.

Now the OEMs, those that build servers, take X in component costs, add in MG&A, engineering, R&D, manufacturing plants and profit and typically get 4X as sales. Thus X is no more than 25% of the sales (revenue). X covers all components, cases, racks, PCBs, connectors, cables, disks, PSs, etc. Chips are one of those components. Do you really think that all else is not larger than the main CPU chips (or in the case of Slotted CPUs CPU modules)? Even in PCs that is not true. Thus, we are already below 12.5%. And thus, 2% is closer.

Now SD shops do not do all of that engineering, have little advertising, and low cost of plant (the store covers both). Thus they sell you the components at retail or wholesale, and sell you the labor at 4x to assemble your system (at retail some of the labor is folded into the component costs). They can survive on the 3x labor gross profits. They do not ask you to pay for the fancy digs, the big engineering team, the glossy ads, the name recognition, and all that jazz because you and they know it does not exist and not wanted (at least on your part).

HOw many servers have you specified? How many have you opened up and looked inside? Did you spot the CPUs? What on the open market could you get for them (to savvy buyers)? Now how much did the system cost? Even compute servers (the ones with large CPUs and memory and little else) would the cost of the CPUs be 25% or more of the cost of the system. And they replace systems with the same CPU cost but far higher costs due to engineering and unneeded expansion possibilities. And that is the upper bound of CPU revenue margin servers. Most will be far below that number. And as has been shown, the lower bound is under 1%.

Could the percentage of CPU costs grow? Yes, it probably is. But, it will never get to the 25% stage (unless you think that a totally integrated SOC would do). It will probably go to the same level with a plain jane high end PC.

Pete