From the time of Xscale – Intel has always put its mobile ICs on trailing-edge processes.
As indeed Intel did in its various, failed, attempts at the ASIC business.
And as it has done with these Moorestowns and Medfields which were supposed to get Intel into mobile but didn’t.
And what of Silvermont’s successor – Airmont – due to run on Intel’s 14nm process which is slated for production later this year? A little dicky bird tells me the final netlist for Airmont is not completed. So Airmont won't at the front of the 14nm node - just as Airmont's mobile predecessors were never at the front of whatever was the leading node of their day.
Intel is doing what it has always done – confined mobile ICs (or any other non-PC or non-server type of chip) to a less than leading edge process.
Either this means there are forces within Intel who will stop Intel’s mobile business gaining traction over their dead bodies.
Or it means that Intel thinks it can get a successful result from repeating actions which have resulted in failure in the past.
Which, of course, is one of the definitions of being potty.