Its more like a roadblock than a challenge, at the end of the day you figure out that you have to do 2 games in parallel, one in a cpu limited enviroment under DX and other for mantle where you can do a lot more.
How many devs whould choose for doing the game as DX primary and them add Mantle implementation where CPU requeriments will be lower but the endgame will the same? That to me is not a big deal after all. Now a game with mantle as primary could yield very good results, THATs good, but when you realise that you could only sell that game to GCN radeon users things fell apart.
There is the possibility as Paul98 wrote, that the consoles could drive new games and thereby we could end in a low api only solution for a game. Consoles and cgn is a huge market. And if you make a game that is just superior to the other games, people will buy it, and here the consoles plus cgn could be plenty. We have to see this not as a war between dx and mantle, but more as a way for the gaming market to outmaneuver the others and create compettitive advantages.
The question is, what game uses a low level api to the degree where dx is not viable. Well the obvious answer is ofcource rts and as said, rts is not a big console market.
But after Paul98 wrote it, it made me wonder if rts really is the only scenario where it could happen? Or if rts could never be a success at consoles? As it is now, yes, but what happens in a year or 2 with the new consoles. Looking at the oxide demo i think there is an opportunity for types of games we have not yet seen.
The situation is new. Consoles and PC is now very much alike unlike before. They can be treated as the same market. Normally when we have a new situation new solutions tend to arive after some time. We dont get more of the same, we get something new. But we will have to wait a year or two to see if something radically new pops up