You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
Intel’s regularly scheduled analyst day was moved from November 2016 to February 2017 leaving everyone to speculate why. This could be good news or bad news of course but the word on the street is that it is good news, very good news.
Intel is the number one semiconductor company in the world and has been for many years. Samsung is number two and is now applying serious pressure on Intel by releasing a 10nm process in Q4 2016 with production chips from Qualcomm and Samsung appearing in the first half of 2017. TSMC also started 10nm production in Q4 2016. AMD is putting additional pressure on Intel with their next generation ZEN CPUs expected in the first half of 2017.
As of April 2016 Intel 10nm chips were scheduled for Q4 2017 but chatter indicates that the Intel 10nm process has in fact been moved from Oregon R&D to Israel for production. This poses a serious threat to the AMD Zen part which is manufactured by GlobalFoundries at 14nm. It also puts Intel back in the process density lead, until Q4 when TSMC releases their version of 7nm:
View attachment 18727 Bottom line: If Intel does announce 10nm production during their February analyst day presentations Wall Street and others will certainly take notice.
The first Zen is a desktop chip, the second will be a server chip. Intel's 10nm cannonlake desktop chips won't appear until at least 2019, 2 years after the first Zen is released. Zen will be competing against Intel's Skylake (still quite stiff competition though).
If Intel release Cannonlake ahead of time it's going to be a low volume, Y-series product (like they did for the first 14nm CPU). Because of the poor yields, they won't want to damage their margins. In other words, it'll be a pure PR move, with the volume products appearing much later in the year.
It sounds strange to delay an analyst conference by two to three months just because one single pending good news? There are plenty of other important things Intel can talk about and Intel can always have additional events to explain any additional good news. Why Intel feel it's necessary to postpone a regular analyst day? I'm sure if Intel waits for another 12 months there will be more good news they can announce. Why don't they postpone it even longer? It doesn't make sense.
If they had to wait until Feb to release good news, it means they didn't have much good to talk about in November. In the world of PR, you want to have good news to distract from bad news, so I wonder what the bad news is. We will be able to guess by paying attention to what's conspicuously absent from the discussion.
From what I've briefly read on Wikipedia and Semiaccurate and wccftech (below), 14nm Coffee Lake (a hasty insertion?) to arrive after 10nm Cannonlake this year and coexist with it, an indication of earliness for the 10nm.
I get the sense like Intel is pushing ahead with 10nm before it's really ready for mass production so the marketing department can continue to claim process advantage. My view has always been that Moore's law is limited more by economics rather than physics, and pushing ahead of the curve might be technically feasible but not an economically sound thing to do.
I get the sense like Intel is pushing ahead with 10nm before it's really ready for mass production so the marketing department can continue to claim process advantage. My view has always been that Moore's law is limited more by economics rather than physics, and pushing ahead of the curve might be technically feasible but not an economically sound thing to do.
