Replies to post #37759 on Intel Corp (INTC)

[chipguy]

Following you line of argument consequently ends up noting Intel did neither invented the transistor nor the use of silicon to fab fab chips.

Brush up on your history books. Intel didn't invent the
MOSFET, DRAM, or MPUs. They just perfected them
and made them manufacturable on a industrial scale.

That is the advantage of Intel - it has the billions per year
to work ahead of the curve and solve the basic physical,
chemical, and mechanical problems in future processes
that impact the manufacturability and economics of new
technology often before the IBMs and Asians of this world
even know they exist. That's why the latter are always
scrambling in catch up mode. The hangers-on and process
tech buyers like AMD are even further behind.

And just this week the third biggest semico in the world
has announced it can't afford to stay on the process
development treadmill past 45 nm.

[wbmw]

Re: Bottom line, i did not mean to suggest AMD invented the wheel, but that it makes nice vehicles out of it.

Why would you give credit for AMD implementing someone else's creation? It's like giving Intel credit for their motherboard business.

True, things like copper interconnects and strained silicon had been researched for years before Intel put forth their own R&D. Intel rarely even investigates a new technology without having some fundamental university research on the subject (and many times, Intel actually funds these universities to do the research). But then Intel has to take that technology, develop it into high volume, tune it to their own process, and ramp it quickly.

Does AMD do the same? No, they get the process recipe from Motorola or IBM, which has already been developed, tuned, and ramped to a particular process. AMD's value add by comparison is quite small, at least if you are using copper, SOI, and strained silicon as examples.

IMO, AMD's best asset in terms of process engineering is the methodology they used to ramp yields at Fab30. Running at over 12M CPUs in Q4 2005 was more than most people predicted. However, I can't say the same about their performance today. Running just over 17M CPUs with Fab30, plus Fab36 and Chartered does not seem to be suggesting the same progress in yields as before. Then again, their most likely problem right now is being demand limited; although, if this were the case, keeping Chartered around at the costs involved seems very perplexing.