This is a matter most probably full of subtleties and if there is one company that absolutely cannot afford to lose face, that's Intel, so expect them to spin the news in their favour for as long as possible. And in that regard, at the end of August we had the news of GF announcing that they would be stopping their 7nm development
, despite having kept saying how it was on track (despite some capacity problems) up until three months before
. Intel is a much bigger player, so expect much bigger moves.
Anyway, for further context, on Sept 17, SemiAccurate posted news
that Intel had downgraded their 10nm process, but that the downsides of removing a certain technology were not as bad as previously thought.
However, there comes a point in time where you have to question the end goal of it all, and if this latest news is true, that's probably what they did. We had already discussed based on some slides that the first iteration of 10nm wouldn't be as good as 14nm++ on all metrics and that we'd need to wait for 10nm+ to see some real improvements, so if the news of the downgrade is also true, then there comes a point where that altered 10nm is basically pointless compared to 14nm++, not to mention the yield issues that they are having.
So, to save face, they will postpone whatever bad news they have, especially as we're on the verge of holiday season with the brand new Coffee Lake refresh CPUs out there, so any updates to the roadmap will probably only be disclosed early next year. And that's not to mention that their roadmap is already full of interim 14nm products, also for the server space.
But even then, they can still keep 10nm releases scheduled for holiday season next year as they said, in very low volume just to not be called out on it, isn't that what they've been doing so far anyway - selling a 10nm low clocked mobile dual core CPU with the IGP disabled to a single laptop maker just to keep appearances? So, of course even if they have substantially revised their plans, like making a decision to spend even more on EUV and jumping to 7nm, that will probably only bear fruit in 2020 or 2021, so they have to keep the straight face for months now, all the way into late next year.