One place where I worked studied a model with peak-load throttling. If you're trying to stream 4K video and the network isn't congested, you get full speed. If the network is congested, the subscribers who have been less active during previous peaks get priority and others get throttled. (Actually this was some years ago, before 4K video was a thing, but it's same idea)
Another model I remember studying involved charging for "time using the service" and not "number of bits transferred." This is similar to how voice plans have "anytime" and "off peak" allocations of "minutes."
So there are people thinking about this (or at least there were several years ago), but I'm a little surprised to not see any trials of alternative arrangements.
The Comcast cap has nothing to do with technical decisions, it is purely there to discourage video streaming in favor of Comcast's own options.
http://arstechnica.com/tech-policy/2010/03/fiber-its-not-all...
If you're not offering a competitive streaming option, you're not really eligible to receive that revenue anymore. You can't just invent a new way to monitize it with existing customers who have existing agreements for your services without offering them something to agree to, then allowing for their choice.
If corporate documents are instructing reps to tell customers it's about congestion when there is no congestion, that doesn't look good at all when the government starts asking about these plans and policies.
It's also probably not a good idea to be telling customers that the network you spend hundreds of millions of dollars marketing as super modern and awesome just can't handle their traffic.
I'm not a fan of caps, either, but aggregate home bandwidth usage is going to go up and up, while more and more customers are going to have everything they need from IP services and thus want to cancel traditional TV. Should you charge an IP customer who uses 4TB/month streaming 4K the same as one who uses 200GB? How else do you discriminate?