InvestorsHub Logo
icon url

alan81

10/02/05 11:39 AM

#63147 RE: bobs10 #63140

What will Intel do with all that capacity?
what's INTC planning to do with 4, 65nm plants
If we look at the past 12 months we see total INTC revenue of 35.6B, while CPU revenue is about $26B. This puts the revenue in the OTHER INTC businesses at a little under $10B.
A few years back I looked at companies like TI, STM, Phillips, etc... that make products similar to what INTC makes in the "other" business units. What I found was they were able to get about $1B/year revenue for every factory they had. This is pretty consistent with what AMD can get out of flash as they currently have a couple of modern factories and a couple of older factories and get about $2B/year in revenue out of the business. If we set this as a benchmark, and then expect the "other" to grow at about 20%/year, that is about $2B/year. That means in the next two years (the time between each process node) they would need about 4 new factories just to keep "other" growing at a good clip. As pointed out, newer INTC factories can certainly count as a couple of older factories. You need to somewhat counter this with the fact that two of the "new" 65nm factories are really old factories refurbished.

Also, keep in mind that CPU die sizes will be about the same at 65nm as 90nm due to the shift to dual core.

In terms of "SOI and DSL"... sorry for the comment, it was condecending, but there was a point. It was a trap and I apologize for that. From your link DSL is
"a transistor manufacturing technology that the companies say can provide up to a 24% speed boost without increasing power requirements....The companies say the "dual stress liner" technology is a second-generation strained silicon process " <- article dated December 2004.
This is as opposed to the INTC strained silicon process which can provide up to a 24% speed boost.
ftp://download.intel.com/technology/silicon/Datta-Chau-foils-IEDM-2003.pdf
slides from December 2003 IEDM....

In terms of SOI, Intel did a nice presentation on SOI at spring 2005 IDF. I was able to find one key slide here:
http://www.behardware.com/art/imprimer/556/
In summary, at 130nm Intel reports SOI as gaining between 10% and 15%, this drops to between 12% and 6% at 90nm, and down to between 3% and 10% at 65nm. The difference between the two numbers is a requirement in some circuits for a history guardband. So Intel has said there is performance gain in SOI, just not enough for them to justify the cost. Of course, with the gain from SOI an ever decreasing number it does not make much sense for intel to start using it at the 45nm node.
--Alan
icon url

wbmw

10/03/05 1:57 AM

#63187 RE: bobs10 #63140

Re: Well if INTC has such a great process producing tons of huge chips with nary an error what's INTC planning to do with 4, 65nm plants which will give INTC in effect 12+ processor fabs compared to fab30?

You're talking out of your non-technical butt again. 300mm wafers offer 2.5x more die per wafer, but not 2.5x more output per fab. Because the equipment is physically larger, and due to manufacturing differences, the wafer throughput for 300mm fabs is actually a lot smaller than 200mm fabs. If you look at wafer throughput, most large 200mm fabs can output between 6-7k wafers per week, but if you look at even the largest 300mm fabs right now, they are topping out around 4-5k wafers per week (16-20k per month). Overall, that probably makes die throughput of a 300mm fab more like 50% greater than 200mm, as opposed to the >2x used in your assumptions.

So four 300mm fabs may be capable of producing ~6 times more output than fab 30. Imagine the CPU market will be bigger next year and the year after that, than it is today, but this would still enable Intel to supply almost the entire world's demand for microprocessors. Add in AMD's capacity, and there will likely be an oversupply in the market. Intel can only control prices so far, legally, before they can truly be accused of abusing their monopoly power, but if AMD initiates a price war, I fully expect Intel to follow with price cuts of their own, and like 2002, this could mean a lot of red ink for AMD if they choose to enter a price war.