Soldato
so here is how i see it.
nvidia need few thing starting with :
1-claim back fastest single card title ( wont be easy at all, if they fail thats a probleme, so canceling it is still wiser imo)
2-avoid watercooling, to not be perceived as copying of AMD ( not being able to keep frequency stable, temp and noise in check, with ppl focusing on these issues, most titanz sites will warm up the cards before benching unless bribed i guess xD)
3-adjust the price to avoid the counter AMD PR with the hashtag #2betterthan1, so the price need to drop far enough to make it difficult for AMD to hold the Ratio 1/2 to avoid the AMD counter PR on Titan Z launch.
honestly i think canceling it would be wiser, live to fight another day, wouldn't feel as bad as if they launch it and fails against the 295.
Isn't this being targeted as a compute card though, for designers etc that need compute that can also run games but more attractive than typical Tesla card?
The way I see it is if it doesn't beat AMD's dual card on gaming but comes close using only air and running on less power it's not a defeat as Nvidia can simply say it's a compute orientated card, which it is. Then later if they really wanted can launch a dual gaming GPU to target 290x2's gaming performance.
Either way it's no bother to Nvidia they will most likey shift Titan Z and a more typical gaming dual card as well. These cards are the last hurrah before moving onto 20nm. I doubt AMD / Nvidia will sell many anyway tbh. Any they can sell is icing on the cake at the end of 28nm...
£1200 for the AMD card is also a ridiculous price tag. $3000 for the Nvidia card more so. I really doubt many of either will be sold.