May is not too late in my opinion. Vega is still to show its teeth and could perform anywhere. On top of that we have no idea on its pricing. We've had no rumours on that front. I personally doubt we'll see it competing against a Titan XP, but it will likely beat a 1080 and probably come close to a 1080ti in certain instances at a cheaper price. I can only really speak for myself here, but above 1080 performance at the right price is excellent. I know it will last longer than a 1080, but I also recognise it'll be a year late to the game. If the 1080 had originally been cheaper, it would have been an amazing purchase and most would have upgraded by now. But it was just too expensive and alienated a lot of potential customers, all of which are now waiting for Vega. Nvidia were taking the micky with it. They were selling a midrange chip at enthusiast prices just because it performed above expectation and had no competition. It's like they said to themselves, 'Hey, this is a great GPU! Lets sell it at an exorbitant price and make a killing.' That's all well and good in less significant areas of industry and economics, but not on the scale we see here. Nvidia charging this much is because they can, not because they should.
I don't feel the Pascal refresh will to be all that impressive. I imagine it'll just be a slight clock increase with lower power consumption and a cheaper price. The important element is cheaper price. Nvidia could quite easily reduce the price of the 1080 so far that it knocks the wind out of the sales (pun not intended) of Vega 10, just as the GTX 980 was still very popular due to its price drop despite being weaker than the R9 Fury. Even if Vega 10 is as low as $500 (doubtful), all Nvidia has to do is reduce the 1080 down to $450 and it's still a competitive GPU, and with this refresh it might be possible for Nvidia to continue to make a tidy profit.
