comment on this article

FinFET yields caused Intel to mothball Fab42

Industry watcher Malcolm Penn gave an interesting interpretation at his IFS 2014 event of why Intel has mothballed Fab 42. The company, he believes, has been too successful in scaling its FinFET technology.

"Yields from its FinFET process have been much better than expected," he suggested. "This means Intel will only need two 14nm fabs, rather than the three it had planned for. It's a better position to be in than not having enough capacity." A further factor has been the decline in demand for pc processors.

Penn pointed out that Intel got out of the planar transistor world at the 40nm node, in contrast to the foundries, which continue to develop planar implementations. "So Intel has had a long run at learning how to make them."

Apart from Intel having this 'long run', said Penn, it didn't have to worry about anyone but itself. "TSMC has to work with a large number of customers and it's much harder to get things right."

Graham Pitcher

Comment on this article

This material is protected by MA Business copyright See Terms and Conditions. One-off usage is permitted but bulk copying is not. For multiple copies contact the sales team.

What you think about this article:

This is simple an issue of how much silicon Intel needs for the PC and mobile device market. Mothballing Fab 42 likely is not the only action in adjusting their manufacturing capacity. I also expect to see an impact on Intel's 450 mm program.

Posted by: Walter Huber, 07/02/2014

Posted by: semi person (28), 04/02/2014
Curious idea that "high" finFET yields led Intel to overbuild by an entire $5B fab...that they now "do not need". Intel has been financially very successful in recent years....and this is not how they got that way! Lower need for PC chips makes more sense at the moment. And, by the way, Intel did not move away from planar CMOS at 40 nm, they made a nicely competitive 32 nm cycle with planar designs. The decision to ramp finFETs was indeed made early(by '08) and has given them a substantial learning curve lead over TSMC, GF, etc.

Posted by: Michael Current , 24/01/2014
"This means Intel will only need two 14nm fabs, rather than the three it had planned for" - So which two plants are coming up on 14 nm Mr. Penn? Oregan and the other fab in Arizona? What about Ireland?

Posted by: Joe Bloggs, 23/01/2014
Did Mr Penn provide any actual EVIDENCE that Intel has better than expected 14nm yields??

I'm an Intel Person and I VERY MUCH DOUBT that Intel planned 3 factories based on a poorly understood process and then overshot its yield calculations by 33percent.

That's not a viable outcome of statistical process control and its not how Intel works.

Posted by: Pedro, 22/01/2014

Add your comments


Your comments/feedback may be edited prior to publishing. Not all entries will be published.
Please view our Terms and Conditions before leaving a comment.

Related Articles

Serving a purpose

In 2010, AppliedMicro made a strategic decision to sign an architectural ...

Bus stop?

While the debate about VME’s future continues, companies are developing new VME ...

Under new ownership

Since its launch by Real Time Engineers some 15 years ago, FreeRTOS has grown ...

Intel to buy Qualcomm?

I suggested the other day that the Qualcomm-Broadcom ‘will they, won’t they?’ ...

Stinging reports

Another week, another report highlighting the industry’s apparent inability to ...