Apple has bought a fab

Phynaz

Lifer
Mar 13, 2006
10,140
819
126
http://www.eetimes.com/document.asp?doc_id=1328505

Apple Inc. has bought a wafer fab in North First Street, San Jose, Calif. from analog and mixed-signal chip vendor Maxim Integrated Products Inc., according to a Silicon Valley Business Journal report.

Apple paid $18.2 million for the 70,000 square foot building at 3725 N. First St. the report said, referencing public records.

In May 2015 ATREG Inc. (Seattle, Washington) announced it had been retained by Maxim to help with the sale of its 200mm R&D fab there. ATREG described the fab as being suitable for prototype, pilot, and low-volume manufacturing and as a platform to strengthen U.S. based customers with strategic partners.

The unit was described as operational and included 197 pieces of chip manufacturing equipment from such makers as Applied Materials, Hitachi, Novellus, Tokyo Electron, KLA and ASML suitable for production at nodes from 0.6-micron down to 90nm, with the bulk at 0.35-micron to 180nm.

Apple's fab is immediately next door to an office building belonging to Samsung Semiconductor.

Apple is unlikely to use the fab for production of its main chips, for which it uses foundries TSMC and Samsung, but could use the facility for R&D in other components such as mixed-signal devices, MEMS and image sensors and for work on packaging.
 

ShintaiDK

Lifer
Apr 22, 2012
20,378
145
106
They also really need some private R&D. Rather than having test chips made at TSMC or Samsung for others to look at. When mass production starts its too late, but if you need a good year in advance to fix issues and tune the chip. Then its a problem.

We all know how Samsung works for example.
 

Ferzerp

Diamond Member
Oct 12, 1999
6,438
107
106
What could they be making on such a (relatively) large process? As I understand it, prototyping on a large process doesn't really translate in to something useful for a small process does it?
 

scannall

Golden Member
Jan 1, 2012
1,946
1,638
136
What could they be making on such a (relatively) large process? As I understand it, prototyping on a large process doesn't really translate in to something useful for a small process does it?

The building already being a fab has already got all the environmental and clean room controls in place. Adding new equipment for smaller processes would be fairly straight forward.
 

master_shake_

Diamond Member
May 22, 2012
6,430
291
121
leave it to you guys to turn a hill of sand into something useful...

im sorry that was just awful.
 

Mondozei

Golden Member
Jul 7, 2013
1,043
41
86
Very interesting, but also kind of curious. Apple has always had the philosophy of "focus on your strengths". This is way out west of their traditional areas.

Then again, so was CPU design and we know how well that ended up. Also, the buying of the facility is really peanuts. This is probably a test balloon to see how well they use this facility.
 

ViRGE

Elite Member, Moderator Emeritus
Oct 9, 1999
31,516
167
106
It's an R&D lab, possibly to try some designs out but also for the car.
Agreed. Maxim Integrated is an analog/mixed-signal company; it's completely the wrong kind of facility for digital logic production. Plus it's far too small, even for a test fab.
 

aigomorla

CPU, Cases&Cooling Mod PC Gaming Mod Elite Member
Super Moderator
Sep 28, 2005
20,841
3,189
126
"0.6-micron down to 90nm, with the bulk at 0.35-micron to 180nm."

uhh.... u cant even make NANAD's at that waffer size.... what are they going to make at such a large waffer size?
https://en.wikipedia.org/wiki/90_nanometer
90nm => intel Prescott / AMD Athlon 64 crica.... approx 2004, 11 years ago.

"Apple is unlikely to use the fab for production of its main chips,"
Unless Apple is going to dump billions, and yes its billions, considering a new fab from intel costs almost 5 billion dollars,
http://usatoday30.usatoday.com/tech/news/2011-03-28-intel-manufacturing.htm
"Each chip factory costs Intel $3.5 billion to $4 billion. Intel said it will also spend about $7.3 billion on research and development in 2011."

I really dont see the use of this fab outside RnD for highly specialized projects.
Even then, apple has no idea how to run a FAB.
 
Last edited:

Rakehellion

Lifer
Jan 15, 2013
12,182
35
91
They also really need some private R&D. Rather than having test chips made at TSMC or Samsung for others to look at. When mass production starts its too late, but if you need a good year in advance to fix issues and tune the chip. Then its a problem.

We all know how Samsung works for example.

Sounds like that's what they're doing. No way this shack can produce 100 million iPhone chips.
 

videogames101

Diamond Member
Aug 24, 2005
6,777
19
81
Plenty of companies run small fabs at older nodes for their own specialty purposes, but none of them could be used for mass market manufacturing. I agree with the article's conclusion, I would guess Apple wants to research doing some in-house mixed-signal stuff.

As far as speculating about Apple developing their own fabs: Even if Apple decided a few years ago to pursue their own bleeding-edge digital fabs, the plan would still be many years away from coming to fruition. Can't see any reason they'd go down that route, absolutely no return on investment even in the long run IMO.

In fact, I would venture a guess that in 5 years there might only be 2 leading-edge fabs (for digital logic) left. We're already down to 3, GloFo hardly counts at this point.
 
Last edited:

monstercameron

Diamond Member
Feb 12, 2013
3,818
1
0
Plenty of companies run small fabs at older nodes for their own specialty purposes, but none of them could be used for mass market manufacturing. I agree with the article's conclusion, I would guess Apple wants to research doing some in-house mixed-signal stuff.

As far as speculating about Apple developing their own fabs: Even if Apple decided a few years ago to pursue their own bleeding-edge digital fabs, the plan would still be many years away from coming to fruition. Can't see any reason they'd go down that route, absolutely no return on investment even in the long run IMO.

In fact, I would venture a guess that in 5 years there might only be 2 leading-edge fabs (for digital logic) left. We're already down to 3, GloFo hardly counts at this point.
Why do you think glofo hardly matters?
 

2is

Diamond Member
Apr 8, 2012
4,281
131
106
Very interesting, but also kind of curious. Apple has always had the philosophy of "focus on your strengths". This is way out west of their traditional areas.

Then again, so was CPU design and we know how well that ended up. Also, the buying of the facility is really peanuts. This is probably a test balloon to see how well they use this facility.

They weren't a watch manufacturer until recently either, they're even investing in their own electric vehicle. They don't want to be the next Nokia or Blackberry and since they can't outright buy Google, they need to expand and innovate anywhere they see an opportunity. If they don't, Google will. It's like an arms race between two super powers.
 
Mar 10, 2006
11,715
2,012
126
You say that but in what context are irrelevant. Just saying doesn't let me understand.

Global Foundries couldn't even properly execute putting into production Samsung's 14nm node.

They also hyped up their 20nm and 14nm XM processes to no end, but in the end they turned out to be non starters.

Global Foundries well and truly sucks.