Intel likely paid a high price to secure ARM’s support, something their foundry wing desperately needed. Make no mistake about it, Intel went crawling to ARM, not the other way around.
SemiAccurate is aware that this view contradicts the common wisdom but stop for a minute and think about where that common wisdom came from. For example Intel is very fond of touting how their fabs are at extremely high utilization despite an ~33% drop in the PC market volume. Add in some crushing losses on the server-side which don’t show up in the marketshare figures because the non-Intel replacement chips aren’t considered ‘servers’ by the company’s math and you have a big hole in volume.
Like the curious accounting techniques used to not show a dropping server unit marketshare, fab utilization also has some logic that would make a contortionist wince behind it. While it is out of the scope of this article, if one parks paid for fab equipment in another building or delays its delivery, that fab is not technically counted in Intel’s capacity numbers. It may be done, equipment purchased, and ready, but if the tools are not installed, it isn’t counted in capacity numbers.
With math like that, capacity utilization is great. Officially the PC market is doing swimingly too, never mind things like this. Server marketshare has never been higher either, any losses not being counted sure helps there. So Intel is either selling more silicon area than ever before to imploding markets or something is amiss. Part of that something is the way utilization is calculated and part of it is a lack of messaging about fab utilization, something that was a common theme at analyst events but recently has been quite a bit lacking.
In light of Intel’s capacity utilization crisis we have a new breakthrough in Intel largess, they will be fabbing ARM based SoCs for 3rd party customers. If you read the presentation where this was first discussed, you can see a few interesting things. While the official spin on Intel picking up ARM’s Artisan IP as an offering is a minor dalliance for Intel, the reality is quite the opposite. Lets take a look at why this deal is anything but what it is said to be.
Note: The following is analysis for professional level subscribers only.
Disclosures: Charlie Demerjian and Stone Arch Networking Services, Inc. have no consulting relationships, investment relationships, or hold any investment positions with any of the companies mentioned in this report.
Latest posts by Charlie Demerjian (see all)
- HyperX ships it’s 60 millionth enthusiast memory module - Oct 15, 2018
- Bittware/Nallatech water cools 300W of Xilinx FPGA - Oct 12, 2018
- More on Intel’s 10nm process problems - Sep 17, 2018
- Intel puts out another 14nm 2020 server platform - Sep 11, 2018
- Why Can’t Intel Supply Enough 14nm Xeons? - Sep 10, 2018