Thread regarding Intel Corp. layoffs

To answer the question IoT, Modems etc.. all don't required leading edge silicon the die's are small have lots of analog and the value proposition is likely best on one to three generation lagging nodes.

Thus beyond memory, x86, Aps processor they are all very capable and value proposition on older nodes.

http://www.eetimes.com/author.asp?section_id=36&doc_id=1330366

Each new node requires huge volumes to get value. You get performance die scaling, and power / value but only for some. The design, mask, prototyping, qual and ramp can run easily hundreds of millions of dollars to billions, so you need huge scale or high ASP.

Thus leading edge node adopters are far and few: Intel x86 ( 150-thousand ) ASP and couple hundred million units, Nvidia / AMD GPU ( 100s of dollar ASP, huge die ) then the leasding edge smartphone for Samsung and Apple's APS processor for the high end smartphone, 30-50 bucks best case into a platform that sells for 500 bucks, FPGA ( Xlinx and Altera now Intel )

That is it small list of companies indeed, the cost to port designs with just a few wafers isn't not a good business proposition. Thus why Intel foundry has few customers and TSMC / Samsung has many as the have huge factories and can amortize the development across many customers including Apple that allowed them to catch up..

PSO's mistake to not accept Apple was epic beyond belief. That enabled Apple to bankroll both TSMC and Samsung R&D with a guranteed leading edge customer to catch up to Intel's technology the past 4 years. Before that nobody could invest nor have guranteed volume at leading edge, even Nvidia and AMD who always want to be on leading edge were too small to fund it.

Virtuous cycle has moved to Samsung and TSMC with the smartphone. BK's was too late in getting ARM at 10nm as the lead is already lost. Only folks that will come to Intel are the ones that are turned down by Samsung and TSMC.