It was an incomplete firmware flash and corrupted/incomplete parameter store. I reverted the fw to 2.0.3, then back to 2.0.7 multiple times. Nothing to do with the card(s).
I had tried 1 CF card, both CF cards, no CF cards. Same power consumption, same battery drain in sleep. The firmware was keeping 1) the processor throttled or 2) preventing a process/component from reaching baseline and shutting down.
It had nothing to do with the card, but the 1Dx is very sensitive to card types. A 600x 64GB UDMA6 card that I have takes ~3 seconds to spool up and allowing the camera to be responsive. The same card in my 1Ds3 takes < .5 seconds. However, the UDMA7 1066x card works fine in both and allows resume from sleep in < .5s. Probably has to do with the clock generator for bus frequency on the data line.
The power tap I was monitoring the power consumption of the camera showed that the current consumption was drawing the same in sleep mode as when on. +/- a few milliamps (probably the LCDs shutting down).
BTW, all the newer cards are clocking 142-148mb/s...I don't own any SD cards except for a 32GB extreme in my 1Ds3, which is only there for backup purposes.
Of course, since the problems seems to be occuring with all these newer bodies, I wonder if the hardware is advancing but their firmware team is still stuck in the stone age. RtOS's aren't that hard to write for, but simple mistakes and assumptions can be catastrophic (look at the Mars Rover's problem with VxWorks). I appreciate that Canon's locked the firmware so that people can't screw with it - there's something to be said about NOT having everybody's fingers in firmware development. However, it would also be nice to be able to see the code and possibly figure out what the cause is.
Well, hell, it would be nice if they just gave us a "factory-hard-reset" button somewhere inside the battery compartment that we could poke (or at least let us clear everything from EOS Utility)!