Look how long it's taken the general public just to get used to the idea of 64bit processing. The hardware technology has been available for many years now but even today with both Vista and Windows 7 offering a 64bit OS, most people are still using 32 bit systems. Just because a technology is feasible and available doesn't mean that the paying public is going to embrace it. And frankly I just don't see that happening with cloud gaming. People are still going to want their own games on their own systems, they won't just jump over to the cloud platforms simply because they're available. I can certainly see it being somewhat profitable down the road, but I don't believe it will ever replace current methods. For instance, I still have a copy of Baldur's Gate 2 that I intend to get back to some day, but that simply wouldn't be possible if I had to relay on a cloud based system. They'll only provide the latest games, and once they're deemed no longer popular enough they'll be purged from the system. All you're going to end up with on systems like that are the "game of the month" type titles, that people play for awhile then forget about a short while later. And they're generally of much poorer quality than what Skyrim is looking to be like.
Your example provides a reason to go to the cloud. One of the many reasons why 64 bit architecture hasn't been adopted is legacy applications needing to be run in 32 bit. There are a number of reasons, but that's just one.
There are plenty of technologies that are "feasible"...but cloud computing is already something that is already becoming the "next" move for many organizatoins, private or public. The reasons to move are just too many, and the problems it creates, in many cases, aren't big enough to stop the move. And that's the key.
Many people said DLC wouldn't catch on because people like to have physical media...yet DLC and media-less gaming is skyrocketing. Further, cloud-based gaming doesn't offer older games today, but one of the hallmarks of the cloud is rapid scalability and agility. If the demand is there and the licensing is possible, it will be made availble.
Cloud-based solutions are profitable today, and companies are scrambling to provide services. There are plenty of problems (why pay for a game I already own, bandwidth, handling user-created Mods, etc) but there are too many reasons to move this way for it not to become a major choice, if not the defacto option imo.