What the user was actually doing on their home VPN be largely opaque to the ISP. They might be able to make educated guesses like "this much traffic over this period of time looks a lot like somebody streaming a movie off their NAS", but that's pretty crude. I'm not sure they could really monetize that sort of data.
On the other hand home VPNs being common would probably increase the how much uploading the average user did. These sort of residential connections are almost always asymmetric because ISPs are counting on people downloading far more than they upload. By empowering users with home VPNs, the balance of upload/download might be upset in a way that isn't profitable for the ISP.
As for upload amounts, you are probably uploading less data, not more, because you don’t have to upload 100% of the data to the cloud to access 1% of it on demand.
If a device manufacturer could count on you having VPN access to your home network they could make it trivial to set up and use. You’d open an app or visit a URL, log in, and you’d have access.
I think that any solution that includes a VPN is going to lose out in convenience even slightly against a solution that doesn't, and in a market where convenience and ease of use goes above all else, it would just not work.
No, it really isn't difficult without the cloud. All the cloud does is make it a bit more convenient.
It has good margins and could afford to market heavily and invest in feature development vs anything else. As a result an alternative might well be great for users, great for the world, and also an unsuccessful business.
Doubly so because it's essentially already solved. Setup a standard NVR and portmap it. The marginal gains are just in making it more secure (few care, obviously, or ring wouldn't exit) or a little easier to configure.
I'm not sure how meaningful that actually is. I know that it's reasonably easy because I've done it a number of times, but I have zero interest in making a productized solution.