Intel Corp (INTC)

[Unkwn]

Intel has always been able to fill the fabs (high utilization) with x86 with a blended mix resulting in 60+% gross margins.

Again,what you're saying is that there isn't enough x86 volume to amortize the fab capacity.
Foundry always has lower margin vs. x86 - i.e. - the opportunity cost is high to trade x86 capacity for foundry capacity.

I'm interpreting the foundry move as an implicit admission of PC/server/x86 - even mobile - stagnation. Why do foundry if internal x86 mobile volumes can fill up the fabs? Answer - they can't.

I am pretty sure Intel can't fill those 14nm fabs with x86 alone, even if they would succeed in mobile. In the past they could, I know that. Already starting with 22nm and slumping PC sales, they had trouble to do so and they mothballed a planned fab, remember. Things change, and x86, at least the PC business, isn't enough anymore to fill those fabs.

Margins: You seem to ignore two of my statements: Gross margins of existing products increase when fabs are paid by more chips due to higher utilization. Second, Intel can charge a nice premium to the already high +50% gross margin of TSMC, as long as they have a process lead and they can keep Moore's Law alive. With 14nm, they claim to be able to. If Intel captures a nice chunk of the foundry business at 14nm and 10nm and Moore's Law is still well and alive, they can charge a higher gross margin than that of TSMC at a larger node. Sounds clear to me. I'd expect at least 60%, where you end up with what you have today (high 60ies), taking into account the improved margins of Intel's high end x86 business. And that with a much higher revenue -> much higher profit. The numbers make a lot of sense to me.