This is not a gaping hole in opening theory. Looking through my reference database, White has an excellent score after 10.Re1.
Black needs to reply with 10...d5 not to be clearly worse, because of the weakness on e6(White can just crash through if Black is careless). After this there aren't many games, but White's score after 11.c4 is 73.3% in 15 games. The computer evaluates the position as approximately +0.20 for White, which indicates that White has full compensation for the pawn.
This isn't hard to imagine. White has a development lead, has a weakness on e6 to work on, open lines for bishops and rooks etc. It is essentially a very sound gambit. The only reason why the computer gives White just +0.20 eval is because computers do not tend to make mistakes that humans make. They can handle scary looking positions in another way than humans can, and they don't make tactical errors. If you look through the games played after 11.c4, you will notice that virtually every black player has played a game-losing move at move 13 in this line (after 11... bxc3 12.Nxc3, Nxc3 13.bxc3, Be7?? and now White obtains a winning advantage after 14.Nf5!).
This means that the position is too expensive for Black to enter just for a pawn. It's totally not worth the risk, in my opinion.