Well, I don‘t have a crystal ball. As with all analysis of economics, it is a very complex dynamic system and I might be completely wrong.
That being said, most consumers have shown in the past, that they really don‘t care about value. Nvidia‘s marketshare in gaming has much more to do with mindshare then with value. AMD has been trying to gain marketshare with better price/performance for years and failed. Partly due to other issues like botched launches, but still.
So why not milk the shills? They will give their allocated GPU budget to nvidia anyway. The ones who don‘t want to spend obscene amounts of money on hardware have become harder and harder to retain as customers anyway. Those buyers realized there are tons of great games that run on a toaster. Look at the steam top 250 and count the games that need new hardware. Those buyers may also consider consoles or other SOCs as an alternative.
Most of my friends think spending 300€ on a GPU is too much, although they earn 80+k per year. Some have switched to consoles, some just enjoy older games. Some people in my wider circle are more hardcore gamers, but those bought GPUs even when the prices where ridiculously high. The gap between those two consumer groups seems to widen more and more. Raising prices for GPUs seems like a logical consequence.
So there is really no reason for GPU manufacturers to miss out on the next big run on general parallel compute. If the application will be mining or not is hard to tell.
Don't get me wrong, I'm think this sucks. I was in between enthusiast and low budget gaming. I liked the times you could get the cut down high end chip for ~400 bucks and get 90% performance for 50% of the price. But I bought new GPUs every 2-3 generations, often switching brands and mostly buying when some ridiculous value card came out (radeon 5700, 9800gts, r9 290, etc). I don't think companies earned that much with people like me, and I think my buying behavior is in the vast minority.