This keeps coming up and the explanation for why it hasn't happened is really pretty simple.
In order to do an "upgrade" (by which one presumes a CPU upgrade, probably to be 64-bit capable), a couple of criteria would have to be met. The first is finding a cheap enough SoC that is compatible with other Pi SoCs. The second constraint is that it has to be of a size and packaging that it will work with PoP RAM modules. Until those two all-important criteria are met, all discussion of upgrades to the Pi0/Pi0W are moot, and especially any discussion about what features could or could not be added.
Now if some "angel" with bucket loads of money suddenly decided to spend several million dollars to have Broadcom develop and put into production an SoC that meets those criteria (and whatever else Dr. Upton thinks would be beneficial) and can be mass produced for no more than the BCM2835, then all bets are off. Do I see any volunteers to do that?