Having more than one camera record at once requires more RAM and even if you shot a video or a picture with only one camera they do a lot of processing and take multiple shots to combine them together.
My guess is the 3D sensing and Virtual Reality features, rendering realistic VR at high resolution is resource-intensive.
The three cameras also shooting at several exposures at the same time before and after triggering the button... and combining them, that’s got to be a lot of memory.
Didn't think of that, but makes sense. Thanks!