AMD’s Q2 2015 analyst call had a few interesting bits in it, lets take a look at the technical side of it. There was also some financial news but we will ignore that because, well, SemiAccurate is a tech site.
The first thing that caught our ear was a $33M “technology node transition charge” that AMD is taking. Translated into non-financial doublespeak what this means is that a few designs that were going to be released on the 20nm node are now being re-done on the 14/16nm node. This is double plus good. Oh, no doublespeak. In short the 20nm node for TSMC and Samsung/GF is both unsuited for anything but borderline trivial low power parts, and it yields double plus badly too. I am trying to avoid the word ‘sucks’ because that does not befit the site’s stature.
So AMD is effectively re-laying out the parts, our hazy memory has two re-designs in it but we can’t say where we got that number. This would put the cost of re-doing a chip from 20nm to 14nm in the $15-20M range, so not trivial by any means. It also explains Zen’s slip from 2015 to 2016, schedule slips take time to redress unless you are a certain company, then you just shed design leads and keep a schedule on paper. The former way works better.
That brings us to the financial news, not that we care about the numbers, just what it represents. AMD is unlikely to be profitable in 2H like they forecast earlier because of a weak PC market. I know you are thinking, “do tell, the PC market is weak?”. Some people are stunned by this and they all seem to work for companies with names like Intel, Microsoft, AMD, Gartner, Forrester, and IDC. Everyone else with more than two brain cells to rub together for warmth understands that the PC market is withering faster than it can be spun as a temporary blip.
Days ago both IDC and Gartner pointed out that PC shipments nosedived by about 10% in Q2. Again. Luckily both have a chirpy outlook that calms twitchy investors and the companies whose stock they own, and who likely commissioned said reports and undoubtedly buy many copies of said white papers if they didn’t commission directly pay. Gosh, what a coincidence.
So the PC market is in a temporary blip that will be all better next year. This temporary blip has been blipping for how many years now? Luckily the analyst firms are dead on consistent in their outlooks, they have been saying this is all temporary and things will improve next year for several years now. As a counterpoint SemiAccurate has been consistently saying the desktop market is dead, not coming back, and the fundamental problems that killed it are quite irreversible at this point. Any company that is seriously banking on the PC revival is done for, it won’t happen. Any analyst firm claiming the same, well, they are probably well paid for their opinions.
The last point of note is about the ramp of the AMD ARM CPUs, specifically Seattle. If anyone thinks the first generation of ARM server CPUs is going to be a sales leader, they simply don’t understand the market, how it works, and why it takes the time it does to ramp. It isn’t complex but is out of the scope of this article, the long version is here if you want to know.
In short Seattle was never meant to be a volume part, and it is going to live up to those expectations. This is not a bad thing though, it is good and means the CPU will meet it’s intended goals, from what we hear it is doing quite well at the moment.
So in the end the news from AMD is bad on some fronts but fills in some gaps on others. The PC market is cratering, that is bad, AMD didn’t seem to expect it, and that is worse. Intel didn’t either so that tempers things a little but not much. The node transition explains the delay in Zen and will end up with a lot better product for the short delay. AMD’s ARM servers are doing what they should, just not what people wrongly expect them to. Six of one, not much of the other but so it goes.S|A
Latest posts by Charlie Demerjian (see all)
- Qualcomm ups Spectra ISP to 2.0 - Aug 15, 2017
- Coffee Lake is going to impact Intel’s margins - Aug 10, 2017
- SemiAccurate digs up Intel Coffee Lake specs - Aug 9, 2017
- Everspin hits the 1Gb milestone with new 28nm MRAMs - Aug 8, 2017
- Intel Xpoint/Optane still doesn’t work right - Aug 7, 2017