Pure speculation on my part but I've felt that the original sins were:
- Selling ARM-based Xscale in favor of trying to push x86 into phones.
- Combining TD with HVM.
The ARM part is maybe a more obvious choice, as we are long past the point where the company should have started attempting to disrupt itself, using ARM.
Now the company has countless competitors using ARM and there are NO customers who will stick with x86 going forward. The platform is burning.
TD produces most of the Foundry loss, because it is capital intensive and always will be. The rest of the fabs, formerly known as HVM, can get to profitability with minimal new external customers, and would attract customers as a standalone business.
Product groups need access to leading nodes but without giving so much margin to TSMC. x86 has lost its competitive moat and will never again be a price maker. It will have to now compete on price, so needs the fab margins to make it work.
The TD capital cycle needs those first-year (high ASP) profits to invest in the next node, and can sell technology transfers to an independent Foundry.
Product groups need to urgently stand up ARM development and start offering that as an alternative to the existing x86 product, or risk having revenues go into a steeper secular decline than it already happening.
Once customers leave x86, they don't come back, and x86 can at best try to slow that inexorable shift, but it can not be stopped.
The other thing is to replace the (miserable, incompetent) Board. They are a bunch of lifers, driving the company into the ditch.
If I see these things happen, it will be a sign that Intel may be ready for something more sustainable than is currently the case.
Might need the stock to get below $10 for a while to force the Board out and do the radical restructuring which is needed for the company to avoid drifting into obscurity.