This line looks unsound.
White's plan seems to be to threaten to win back the piece by 10...dxc4 11. Re1 (pinning the knight), teasing Black into weakening his king's position further by 11...f5 to defend e4, which then leads to a very complicated situation that White hopes to manage better.
However, Black is fine if he just lets go of the knight, since he's still a pawn up. Obviously, his kingside is very weak, but he can move out his queen with tempo (by attacking e4) and castle long.
11...Be6 12. Rxe4 Qd5 13. Re1/Re2/Qe2 etc. 0-0-0
The black king is safe and c4 is actually well protected. White doesn't have enough compensation left for that pawn.
To back this up, according to the Lichess online database, White has a win rate of 62% if Black plays 11...f5 (71 games), but only one of 27%/38% respectively if the answer is 11...Qd5 or 11...Be6 (11/8 games). 11...Bf5 also seems to work (35%, 20 games).
That's just not enough to make the variation appealing for grandmasters (who probably consider the refutation too obvious), and if none of them play a line it will hardly become very popular. It's a nice Blitz surprise weapon though, as you can see the majority of players from a database leaning towards fast games will choose 11...f5.