I looked at the Zero W power components with a strong magnifying glass, and compared to the schematic, and a few measurements with multimeter.
If the USB PWR IN jack is down, the 2306KE chip is above it, that's the power chip. It has four capacitors in a line, left to right, which seem to be C8 (1.8V), C4 (5.2V), C3 (5.2V), and C7 (3.3V). Above C3/C7 is giant L1 inductor. Below C8/C4 is smaller L2 inductor. (Why is L1 different on Zero W compared to Zero non-W? Does wireless need more inductance there?) Above C4 is C2, then R1.
That makes sense to me, correct me if I am wrong. (I wish we could host pictures here, I'd post my diagram.)
I'm thinking now, this will be easier than I thought. I was thinking I'd have to desolder the +5V pin on the USB PWR IN jack, because I assumed it would rabbit right into the board. I now see it has a fine trace from the SMD pin to a power plane (basically where the label is PWR IN on the top of the board). So assuming there are indeed no rabbits, I can cut the trace right there!
I'll try this (probably tomorrow) and post my results. But assuming it's just that one trace, this is entirely possible and not too tricky at all. EDIT: Tried it on a spare Zero W. I believe I cut the trace from the PWR IN microusb +5V pin on the top of the board, and also the PP1 on the bottom (three tiny traces). I'm still getting power. Maybe I didn't get it cut enough, or maybe it does route elsewhere in the middle of the wafer. I may have to physically lift the pin.