Intel to manufacture ARM SoCs
Pretty big news really, Intel entered a licensing agreement with ARM, basically this allows parties like LG, Qualcomm, Apple, and Samsung, to fab ARM SoCs at Intel fabs.
The two chipmakers, whose designs and technology dominate in computing and mobile, unveiled the agreement Tuesday at the Intel Developer Forum in San Francisco. The accord will let Intel offer third-party semiconductor companies its most advanced 10-nanometer production lines for manufacturing the complex chips usually used in smartphones reports Bloomberg:
Intel, which gets the majority of its revenue from making personal-computer processors, has failed to gain ground in the larger and faster-growing phone market -- the stronghold of ARM’s technology. Under Chief Executive Officer Brian Krzanich, Intel is trying to persuade other chipmakers to use its factories for their production. Adding licenses for ARM’s technology could open up that business to fabricating chips based on those designs for companies such as Qualcomm Inc. and Apple Inc., which now have their chips produced by Taiwan Semiconductor Manufacturing Co. and others.
Intel’s embrace of the competing technology comes as the PC market continues to decline and growth in the lucrative server-chip market slows. The Santa Clara, California-based company began offering foundry services -- outsourced manufacturing -- under Krzanich’s predecessor, Paul Otellini, but hadn’t announced any major customers placing orders in large volume.
Indicating that those fortunes may be changing, Intel announced that LG Electronics Inc., South Korea’s second-biggest phone maker behind Samsung Electronics Co., will use Intel’s foundry business to manufacture 10 nanometer mobile-phone parts.
Separately, Intel said it won’t use extreme ultraviolet lithography as a manufacturing technique in its 10-nanometer production. The technology isn’t ready for the next generation of production, 7 nanometer, Intel also said, and the company won’t use the chipmaking technique until it delivers the promised efficiency.
ASML Holding NV is the main manufacturer of EUV machinery and has been trying to make it ready for full use for more than a decade. ASML’s stock fell as much as 3.5 percent in U.S. trading Tuesday following the comments by Mark Bohr, Intel’s head of process architecture.
“I can’t say whether it’s a year from now or three years from now,” Bohr said of the tehnique at the Intel conference. “I am hopeful.”
Intel to launch 540S-ssd with TLC nand flash memory - 03/30/2016 05:31 PM
Intel is to launch 540S-ssd with TLC nand flash memory. The 540S specs can be found on a German price comparison engine. The SSD series will receive 16nm tlc-memory from SK Hynix, the units will be av...
Intel to Bring the Best in Gaming to the World Stage Intel Extreme Masters - 02/12/2016 09:45 AM
Celebrating the 10th season, Intel Extreme Masters returns to Katowice, Poland, as the world will watch the best professional players compete for over US$600,000. Intel will be present to show the bes...
Rumor: Intel to release three 10nm generations of processors - 01/20/2016 12:40 PM
Intel, according to rumors, is to release three generations of processors that will be fabbed on 5nm after Cannonlake and Icelake. Yesterday we already talked a bit about 5nm at TSMC. This new rumor p...
Intel to release third Core i line at 14 after Skylake - Kabi Lake - 06/26/2015 09:14 AM
Intel will likely release a Core i-processor lineup in 2016 based on the Kaby Lake-series. This would be the 3rd generation processors that Intel fabs at 14nm....
Intel to push Broadwell - dates revealed - 10/07/2014 09:25 AM
Intel is planning to push its 14nm Broadwell processors, which will be released by the end of 2014, into three major notebook segments: entry-level non-touchscreen traditional notebooks sized 10.1- to...
Senior Member
Posts: 1821
Joined: 2011-10-09
Senior Member
Posts: 1285
Joined: 2006-09-02
Just a hedge against lost revenue from AMD CPU competition with Zen and Zen +.
Senior Member
Posts: 22421
Joined: 2008-07-14
Intel is going to be manufacturing ARM based processors for other companies. Intel has no interest in developing ARM based processors. The profit margins aren't large enough.
AMD is still working on ARM based processors. They recently released an ARM based Opteron and there's rumored to be an ARM based Zen processor in the works.
What lost revenue?
Moderator
Posts: 12481
Joined: 2000-05-23
Intel isn't doing just x86 or 64 bit x86.
They were in RISC cpus as well. Even if it didn't work out as well.
(i960 for example; i think it has some Arm in it's DNA. Had that chip in my first DSL modem.)
Senior Member
Posts: 8099
Joined: 2014-09-27
Not really. The 14 nm process is good from Global Foundaries/Samsung, and TSMC on the shrunken node as well. Intel does have 10 nm but currently they haven't got it working quite right. So far, instead of getting high performance 10 nm Cannonlake processors we're getting a modified Kaby Lake (Coffee Lake) instead, on 14 nm. Cannonlake is now reserved only for the low end and slower parts. The news of a 6-core Coffee Lake (Kaby Lake Refresh, and Kaby Lake a Skylake Refresh was already an added processor due to Cannonlake delays) overshadowed this fact.
It's not even close actually. The Intel 14nm is almost what the TSMC/GloFo 10nm is going to be.
Give it a read here.
“Not all 10nm technologies are the same,” said Mark Bohr, a senior fellow and director of process architecture and integration at Intel. “It’s now becoming clear that what other companies call a ‘10nm’ technology will not be as dense as Intel’s 10nm technology. We expect that what others call ‘7nm’ will be close to Intel’s 10nm technology for density.”
It wasn’t always like that. Traditionally, chipmakers scaled the key transistor specs by 0.7X at each node. This, in turn, roughly doubles the transistor density at each node.
Intel continues to follow this formula. At 16nm/14nm, though, others deviated from the equation from a density standpoint. For example, foundry vendors introduced finFETs at 16nm/14nm, but it incorporated a 20nm interconnect scheme.
Technically, the foundries didn’t introduce finFETs at a full node (14nm), but rather at a half node. TSMC, for one, calls it 16nm.