I dont see 1. working out. The HEDT line is dead-ish IMO, cant see where it fits in a world where 8 cores from the competition is mainstream.
Coffeelake is the cure and it needs to grow an additional pair in the meantime
.
HEDT platform will definitely not be dead. Intel can put 8-cores on two platforms, and the higher-end platform will have 10 & 12 core options as well, with a significant upgrade in prices and more PCI-e lanes and quad channel memory. There are enough people stupid enough to buy it make it worthwhile for them to do so - along with maintaining the halo effect.
The problem with them doing it anytime soon is that they would have to have this planned for years already for maximum benefit.
I think we'll see Intel only meeting AMD prices in the middle - but not head-on - in the retail market. OEMs will get massive discounts to "encourage" their loyalty.
Intel's HEDT platform has features AMD's AM4 platform simply doesn't have - it is irrelevant if those features are valuable for most users, they will be genuinely valuable for some people... and there are, as I said, enough idiots that will pay far more than they should just because it makes them feel good.
Assuming AMD pricing as follows and NO Intel performance benefits:
Ryzen 3: 2xxx x4: $130~$200, 4/4
Ryzen 5: 2xxx x4: $250~$350, 4/8, 6/12
Ryzen 7: 2xxx x8: $350~$500, 8/16
i3 9xxx x4: $150~$200, 4/4
i5 9xxx x4: $220~$300, 4/8
i7 9xxx x6: $350~$400, 6/12 (dual platform)
i7 9xxx x8: $550~$700, 8/16 (dual platform)
i7 9xxx x10: $999~$1200, 10/20 (HEDT only)
This is what Intel would do, IMHO, with equal performance, in regards to retail pricing to counter Ryzen. Pentium and Celeron would still be dual core, but Pentium will have hyper-threading.