http://blogs.barrons.com/techtraderdaily/2016/08/16/intel-hell-freezes-over-with-arm-partnership-for-chip-manufacturing
http://www.forbes.com/sites/patrickmoorhead/2016/08/16/intel-foundry-rounds-out-ip-lineup-with-arm-adds-new-customers-at-idf-2016
(with an extra large photo of clown shoes BK)
10nm for ARM and LG with a few bucks profit, or 10nm server/x86 with huge margins.. Hmm I'd hate to be LG....
Leading edge capacity is the most valuable thing, can't believe Intel is going to produce 70million phones for LG on ARM and shortchange their own servers and cliient...
If they do that than ARM servers will come to, than the intel has NO competitive advantage and TMG is just a cost center competing with TSMC, Samsung, they will get killed... ouch for TMG people
http://www.oregonlive.com/silicon-forest/index.ssf/2016/08/intel_will_make_mobile_chips_u.html
“It’s a very good indication of the supremacy of Arm, and the sticky position Intel has gotten themselves into,” said Richard Holway, chairman of analyst TechMarketView. “They have to admit their own designs aren’t up to scratch, the smartphone market is a lost generation for them.”
https://www.ft.com/content/3d3662f2-645e-11e6-a08a-c7ac04ef00aa
This article suggests they will be fabbing ARM silicon for IoT and 10nm ARM cores for LG phones.
So this is how BK plans to get into IoT. Not a bad idea, but this is definitely not good news for Atom and Quark guys.
Big fat deal............ Altera already has ARM SoC (besides just FPGAs) so Intel would end up making ARM-based stuff even without this.
This is completely not a surprise whatsoever.
so Intel still has ARM architecture license from xscale/marvell days and ran those chips on what amounted to the 8" flash process; Altera has ARM cores all over the place at 28nm but is now two years behind Xilinx at 20nm. Seems like a non-starter here.
Even though the financial pundits have suggested a foundry spin-off, the Fab Gene is too much part of Intel's DNA to separate. @1naf also nailed a lot of other reasons. Altera will eventually come inside, but that's taking a lot longer than would seem reasonably possible.
so, Altera?
Too little too late...
1) 10nm ARM action already at every 'real' foundry
2). Anyone who uses intel will always need to decide if intel can support their prototyping and early ramp in their small Hillsboro fab. It is too small and inflexible to support both huge x86 and internal work as well as outside needs. Compare to Samsung and TSMC giga fab intel is a small d--k
3). Great they got density but they don't got yield. 10 and 7nm TSMC is already yielding. What good is density with no functional chips?
4) cost.. Intel real cost is really expensive. It will be hidden in x86 accounting but they got to pay the piper. More layoffs are coming in 2017 and 2018 as client x86 will fall below 200 million and intel will need to cut more to get costs in line with volume.
Prelude to a spin-off of the foundry business? I do see a potential for it happening in the next 2-3 years. Please comment.
Does this mean F42 will no longer be an empty shell?
Okay, best I can tell is that the good news is the licensing of ARM IP. None of the customers listed in Morehead's article are new - LG is the largest but has engaged for a year - http://www.fool.com/investing/general/2015/11/24/intel-corp-reportedly-trying-to-play-fab-for-lg.aspx - the other three are old news.
So, a step in the right direction. Perhaps a run at Apple from a different angle?