Actually, you should only have 3 hp left, which is actually a bit worse than 4.
I'm looking at a build of what you're describing here, and it seems to me that you may have painted yourself into a corner, so to speak. The problem stems from a couple of things:
1) The Moog synths should probably stay in their own cabs. That way, you have them under their own power supplies, and they have their own audio outs (and can be used for the Rackbrute's audio out, in theory). This would then free up 120 hp in the Arturia cab. Plus, the DFAM + M32 combo in Moog's double-cab stand looks ass-kickin'.
2) You seem to be falling into the 'sexy module trap'. Yeah, some of these might look really killer and seem to pack loads of function if you base that concept on module size. However, this is a real problem, because both assumptions are often pretty wrong. Now, I will note that in the case of the Maths, it's a correct assumption...but maybe not so much with the Chord. While it looks like a must-have, it's actually more of a specialized device. It's also 28 hp, which is a bad idea for inclusion in a limited space rack. If this were something designed specifically for polyphonic use, which is what Qu-bit's quad stuff excels at, that would be a good choice...but then the Moogs would be a lousy choice. Or if you were doing a lot of sequencer-based generative stuff, the Chord would be awesome, but that doesn't seem to be the point of this cab.
The truth is, you need to be looking at smaller modules if you're going to go for maximum function in a 2 x 88hp cab. And you also want to look at what I call 'boring' modules...because in of themselves, they might be boring, but in tandem with one or two other devices, they open up those devices' potential massively. And again, these same 'boring' modules can also open up the Moog devices' potential.
Case in point: a comparator. Ladik makes one that fits in 4 hp, looks boring. Just emits a gate in a few ways when a voltage threshold gets crossed. Ho-hum.
Now, take that same boring comparator and feed a weird CV curve into it off of the Maths. Send the gate to an EG that treats an incoming gate as indicating a 'hold' state. Now, feed that envelope to a linear VCA that has a CV out from a sequencer passing thru it so that the envelope controls the VCA's passing CV level. Now, send that resulting voltage on to a VCO, but gate the VCO so that audio doesn't pass when the comparator's level is 'low'. OK...now jam all of that activity into only 16 hp.
Impossible? Well, I'm looking at that very thing in this other browser window over here...
Yeah, some modules look dull, some seem like they're too small to do the job, and so on. Not true, though. This is just a basic example of what presumably 'underpowered' modules can do, in conjunction with other sleepy-looking stuff.
My suggestion...like I said, re-cab the Moogs and free up your space. Keep the two modules that you do have, and figure out how to anchor an auxiliary cab around them, one that compliments the functions of the Moogs. But before doing that, spend some (a lot) of time on MG studying others' builds, and spend a lot of time with the Moogs to discover their strengths and weaknesses, and how each could be made better by subsections in the Arturia cab build, keeping your module sizes small to jam maximum function in minimum space. In the end, you'll wind up with a much more powerful system with a lot more complexity and capability.