1.d4 Nf6 2.c4 e6 3.Nc3 Bb4 4.a3 Bxc3+ 5.bxc3 c5 6.e3 Nc6 7.Bd3 d6 8.Ne2 b6 9.0–0 Qd7 10.e4 Ba6 11.Bg5

MNb wrote on 09/17/21 at 06:43:20:

White gets good play by sacrificing the c4-pawn, e.g. 13.d5! Ne5 (13...Na5 is similar) 14.a4 Bxc4 15.Bxc4 Nxc4 16.Nf4/Ng3 0–0–0 17.Nh5 Qe7 and then activate rooks+queen, nudge a couple of pawns forward, and play against the weaknesses on f6/d6. Black's king is also a bit vulnerable (to an a4-a5 push at the right moment, say).

Yeah, but unfortunately Black can improve as well by playing ...h6 a move earlier and inviting a transposition the Spassky game: 11...h6 12.Bxf6 gxf6 and now White has to do something better than 13.Ng3 and 13.Nf4.
Still thanks, after reading your comment it crossed my mind that White can postpone castling a bit and play 9.e4 first. That will keep me busy for a while.

A couple of months ago I bought GM Moskalenko's An Attacking Repoertoire for White. I have the same experience: 8...Qa5 and 9...Nf6 promises White nothing. Black will always take pawn c5, find good squares for the knights and start play against White's weaklings on the queenside.
The most interesting deviation imo is 8.Qd2 as in Neverov-Eggleston, Hasting 2008/2009, but I have my doubts. Black has the annoying threat ...f4.
So I looked at the old-fashioned 4.a3 Bxc3+ 5.bxc3 c5 6.e3. The Capablanca plan with ...Ne8 (eg the famous Geller-Euwe game) is not scary but interesting and equal (Geller-Broadbent from a year later is nice). The same applies to the Hübner plan with c5, d6 and e5 (Agrest-Lugovoi. St. Petersburg 1992 or Carlsen-Anand, internet 2020). Both sides can play for a win.
What bothers me at the moment (from White's perspective) is a plan with castling queenside, eg Geller-Spassky, USSRch 1955. Three years later GM Geller did much better by following Jon. Tait's advise. 

FreeRepublic wrote on 12/04/17 at 18:41:56:

I agree! However, in my games (and in the databases) 8...Qa5 has been way more popular than 8...f5.

MartinC wrote on 12/05/17 at 09:55:58:

I wonder what Anand was intending in game 9 of his 2013 World Campionship match against Carlsen. Alas, Carlsen played 7...exd5 (instead of 7...Nxd5), so we'll never know. Nevertheless it's interesting that Anand chose the f3 Nimzo in this "must-win" game.

I've looked at the 8.Qd3 lines - I wouldn't mind facing 8...cxd4 or 8...b6, but 8...0-0 bothers me, e.g. 9.e4 Ne7 10.f4 b6 11.Qe3 Ba6 12.Bxa6 Nxa6 13.Nf3 Qd7!? (both Chandler and Sokolov chose the more solid 13...cxd4 14.cxd4 Nc7 against Shirov). Black intends to place his queen aggressively on a4 (followed by centralising a rook or two), or else use the b4 square for his knight if White plays 14.a4, i.e. 14...cxd4 15.cxd4 Nb4 16.Qb3 a5 =+

Can't help you either I'm afraid. This is the line I play as Black against 4 f3 and haven't yet found a path to anything exciting for White

If the Nimzo-Indian equalises, then at least White can make things interesting by playing 4.f3!?. Or so I thought...
In the main line with 9...Nf6, I'm struggling to see the appeal from White's point of view. How would you handle this variation?