Hopefully you dont mean mantle that has 2 games and doesnt work for all amd's hardware even yet?
I'm not sure you even know what you're talking about. How many games used DX11... before DX11 was fully released? How long between DX11 actually becoming available and the first retail shipping game that used DX11?
Dx7 through dx11.2 have taken YEARS to gain widespread adoption, Mantle will be no different, the real question is why would you expect it to be different?
Game's development isn't a short process, neither is making an API, neither is getting games to use a brand new API, DX10 uptake was horrific for a hell of a long time, with the first several games barely being any different in the slightest, in fact almost all performing worse IIRC, than DX9.
Mantle categorically, according to every single tech site preview/review/article on it has a tiny fraction of the overhead DX does and simply is working better.
DX11 didn't work on all past hardware when it was launched, in fact working on very little. I honestly don't know why people make such comments. People act like in gaining momentum faster than any API before... it's failing. Dx9/10/11 didn't have two shipped games with support both before the beta program was even ready and only 5 months or so after consumers heard the first bit of information on it.
We've already heard about DX12(in as much as we heard about Mantle at first), and it's not due to be with us for another year, and adoption from that point onwards we'll have to see.
How many games use something, is irrelevant to how good something is.