I don't know if you're posting these statements to provoke arguments, but just in case you're ignorant of the actual world market for GPUs, I'll state the following.
The under $300 segment is about 80% of total sales with the under $200 segment a major part of that 80%.
If you want to increase market share and revenue, don't you think it wise to target those segments?
For the short term, yes. But those largest segments of the dGPU market are the segments that have been steadily declining, probably due to quality iGPU options that are starting to overlap those lowest-end tiers and should be creeping into the midrange tiers over the next 2 years with Zen's APUs.
The only segment that is actually growing within the dGPU market (A market that, overall, is steadily shrinking) is the enthusiast/high end market. This is pretty much owned by nVidia right now because I think the Firepro class cards while great, don't represent a significant enough amount of units in that pro designer segment to make a real dent in the diversity that nVidia offers there.
AMD still has very far to climb. THe numbers a few weeks ago about them gaining ~10% in dGPUs is promising, but that is in the largest, yet shrinking segments of the market, on top of those numbers probably being attributed to nVidia simply shipping half the units they shipped compared to the previous quarter. I think to difference is something like AMD shipping an extra 100k units vs nVidia shipping 2-3mill less units than before.
That's a sobering reality.
This is why Vega is so important. I think there is still time for them to be competitive in that enthusiast class and gain some real market & revenue share and establish a nice chunk out of the only segment that will matter 4-5 years from now--
but Vega has to deliver. I don't think they have any room to maneuver if they can't gain any performance wins or real performance value, but as the nVidia partisans have said over and over again, Vega has to gain margin wins for AMD. Don't expect this to be some $300-350 minimum tier pricing if it is a great card that AMD can establish market share with. They will have to price it higher, and it has to be worth that price, imo. The good thing is that nVIdia has currently given them a ton of wiggle room to undercut and still grab some nice profits....but nVidia will have been there so long by the time Vega approaches that they just need to cut 1070/1080 price by ~$100 each and then toss the 1080ti into the field.