The ARM foundries are having to spend a huge percentage of their revenues on CapEx in the attempt to get to FinFET.
While I do think some R&D is captialized, most of the R&D to get to FinFET is...well, it's R&D, not capex :-)
the ARM foundries are having huge trouble catching up and are bankrupting themselves at the same time.
Neither Samsung nor TSMC are anywhere close to being "bankrupted". And, if you believe their announcements, FinFETs will be in HVM from the foundries in early to mid 2015 (not sure how much I believe 'em, though).
It's quite possible the CapEx requirements will drop TSMC out of the state-of-the-art fab category into second tier status leaving only Intel and Samsung. This would give Intel a huge competitive advantage in smartphones and mobility.
Ah yes, if the company that designs its own chips for its smartphones also has the world's leading edge semiconductor volumes, then this is a *good* thing for Intel...right.
Samsung is Intel's worst freaking nightmare.
The ARM foundries also have economic problems and margins problems at 20nm and will have the same problem with FinFET because theirs is a half-step.
Right. Samsung, which makes $8B+/quarter, is going to have "economic problems" at 20nm or beyond. No, I think with Intel's entire yearly net income earned per quarter, Samsung can afford to fight the good foundry fight.
Most of Intel's smartphone problems will be solved in 6-12 months.
12-18.
Most of ARM's fabrication problems won't ever be solved and they are getting worse each day. To solve them they need time, money and technology. They don't have enough of any of them.
Samsung has more money than Intel does and generates free cash flow that Intel could only ever dream of.