(Bit-Tech) Intel shelves delayed Fab 42 facility

rainy

Senior member
Jul 17, 2013
382
4
116
#1
Jul 12, 2000
27,225
21
81
#3
According to the article, the problem is Intel's short product life cycle and no need for the 14nm volume.......

Hopefully this coupled with lack of competition doesn't make Intel artificially slow its release cycle any more than it already does.
 

bullzz

Senior member
Jul 12, 2013
383
0
76
#4
I think they overestimated their volume from foundry business
 
Aug 11, 2008
10,387
40
126
#5
So what percentage of 14 nm production was this supposed to produce?
 

Jimzz

Diamond Member
Oct 23, 2012
4,194
2
91
#6
Another reason they should take in more outside fab work.

Problem is intel expects the same margins on all chips they produce. This has allowed ARM and AMDs low power jaguar to have a much larger market share then they would have if intel could get past their wanted margin rates.
 

JDG1980

Golden Member
Jul 18, 2013
1,522
0
136
#7
Ideally, I'd like to see antitrust authorities break up Intel into two companies: a fab and a chip designer. This would let us see who actually has the best designs, without Intel being able to use their process advantage to crush everyone.

A pipe dream, I know... but if it could actually be done, it would lead to more rapid progress in the IT field.
 

Torn Mind

Diamond Member
Nov 25, 2012
3,244
0
91
#8
Ideally, I'd like to see antitrust authorities break up Intel into two companies: a fab and a chip designer. This would let us see who actually has the best designs, without Intel being able to use their process advantage to crush everyone.

A pipe dream, I know... but if it could actually be done, it would lead to more rapid progress in the IT field.
Their profits pay the well-paid engineers they have. Kill off a good chunk of their profits, and it is certain they will pump out parts that are not so good.
 
Jul 12, 2000
27,225
21
81
#9
Ideally, I'd like to see antitrust authorities break up Intel into two companies: a fab and a chip designer. This would let us see who actually has the best designs, without Intel being able to use their process advantage to crush everyone.

A pipe dream, I know... but if it could actually be done, it would lead to more rapid progress in the IT field.
"Antitrust" laws are not there to punish success, they are there to ensure a fair market devoid of monopolies.

It would be impossible to argue Intel has any sort of monopoly in 2014.
 
Jul 1, 2001
20,628
45
126
#10
Heh... I know that this isn't the P&N forum, but it is kinda amusing that Obama visited this factory in 2012 and proclaimed it to be the future of American manufacturing.

Apparently, the future of American manufacturing is to take state and federal tax incentives and build empty buildings with them :(
 
Last edited:
Apr 22, 2012
20,395
0
106
#11
Last edited:
Dec 17, 2008
1,845
6
91
#12
Was 42 supposed to use 450mm or 300mm wafers? I heard of reports that 450mm is going to be delayed due to supplier issues.
 
Apr 22, 2012
20,395
0
106
#13
Was 42 supposed to use 450mm or 300mm wafers? I heard of reports that 450mm is going to be delayed due to supplier issues.
450 ready. Like D1XM1 and D1XM2. The 2 other major 450mm fabs. The plan for fab42 is now 450mm wafers with a 10nm node.
 
Last edited:

Jimzz

Diamond Member
Oct 23, 2012
4,194
2
91
#14
Heh... I know that this isn't the P&N forum, but it is kinda amusing that Obama visited this factory in 2012 and proclaimed it to be the future of American manufacturing.

Apparently, the future of American manufacturing is to take state and federal tax incentives and build empty buildings with them :(
The incentives were to hire 1000 new employees. Intel did meet that spec.

"Intel had actually hired sufficient new employees to receive state tax benefits, beating a planned target of 1,000 employees since 2011, when construction on Fab 42 started."
 

jdubs03

Senior member
Oct 1, 2013
377
0
0
#15
450 ready. Like D1XM1 and D1XM2. The 2 other major 450mm fabs. The plan for fab42 is now 450mm wafers with a 10nm node.
Makes sense, considering the building itself cost $1.7B, the tools are the vast majority of the cost fab costs, ($5B total), so saving that for 10nm makes a lot of sense, it in effect means there would be no need to start construction on another building. My guess is that lowered CapEx expectations, and that could explain a rise in their stock price since the azcentral article came out.
 

jpiniero

Diamond Member
Oct 1, 2010
5,916
81
126
#16
Well they already closed that Ireland plant that I thought was also for 14 nm. With Intel going towards smaller chips, I guess they don't need as many fabs anymore.
 
Mar 11, 2013
66
0
16
#18
i think this will be good for quarterly earnings being announced tomorrow after close
 

Homeles

Platinum Member
Dec 9, 2011
2,585
0
0
#19
Ideally, I'd like to see antitrust authorities break up Intel into two companies: a fab and a chip designer. This would let us see who actually has the best designs, without Intel being able to use their process advantage to crush everyone.

A pipe dream, I know... but if it could actually be done, it would lead to more rapid progress in the IT field.
All that would result in would be subpar parts from both AMD and Intel. Technological progress would be slowed tremendously.
Tokyo Electron, Applied Materials
They're now one in the same. But there's some truth to what both you are saying. Intel isn't the only one out there, but they do push the envelope from time to time.
 
Last edited:

jdubs03

Senior member
Oct 1, 2013
377
0
0
#20
All that would result in would be subpar parts from both AMD and Intel. Technological progress would be slowed tremendously.

They're now one in the same. But there's some truth to what both you are saying. Intel isn't the only one out there, but they do push the envelope from time to time.
Intel pushes the node with the help of their suppliers, at the same time pushing their suppliers. ASML and EUV for example.
 
Mar 13, 2006
10,089
47
126
#22
Intel pushes the node with the help of their suppliers, at the same time pushing their suppliers. ASML and EUV for example.
Intel's spending on R&D is so large that the industry would not be able to continue without them. Look up the article "Moores' law ends when Intel says it does". You can find it here in the forums from about a year ago I think.
 

beginner99

Diamond Member
Jun 2, 2009
3,916
56
126
#23
I think they overestimated their volume from foundry business
Or maybe it's the reason why there will be no Broadwell on Desktop. One fab less, not enough volume what to do? Don't do 14 nm initially on desktop. Seems a logical choice. Remember Westmere, Arrandale and 32 nm, only mobile and enthusiast platform got such a part initially until SB arrived later.
 
Dec 17, 2008
1,845
6
91
#24
450 ready. Like D1XM1 and D1XM2. The 2 other major 450mm fabs. The plan for fab42 is now 450mm wafers with a 10nm node.
You are correct (did a minor amount of research to confirm what you said now that I am at home and not my lunch break). Originally fab42 was 450mm fab ready where they would produce 14nm 300mm wafers in the very near future and eventually switch to 10nm 450mm wafers.

This makes no sense to me for the old plants were not 450mm ready that is why D1XM1, D1XM2, Fab 42 etc were being built for they would be 450mm capable. It also makes sense to never make these fabs 300mm fabs for not all the billion dollar 300mm equipment can be converted to 450mm. If you are not already using these fabs at 90% capacity it makes more sense to repurpose your existing 300mm fabs to 14nm instead of 22nm or 32nm, save that money and just spend that money on fab42 when it ready to do 10nm 450mm wafers.
 

Homeles

Platinum Member
Dec 9, 2011
2,585
0
0
#25
Or maybe it's the reason why there will be no Broadwell on Desktop. One fab less, not enough volume what to do? Don't do 14 nm initially on desktop. Seems a logical choice. Remember Westmere, Arrandale and 32 nm, only mobile and enthusiast platform got such a part initially until SB arrived later.
Except there will be Broadwell on desktop.
 


ASK THE COMMUNITY