Overall, I love the idea!
Cheapening Worlds - Just about every sport (pro, college, HS...) has its varying levels of "championships" and it doesn't do anything to take away from the "big" championship. For some athletes, winning a championship in minor league baseball may be the end goal, where for others it is an intermediate goal. Both instances are OK. There are some cheer gyms who for varying reasons may never become a big powerhouse / mega gym capable of flying in athletes, offering scholarships, etc. and that's fine. Now that the Summit is over and done, I don't think many would ever say it took away from Worlds. In fact, for my CP it made the desire to get to Worlds even greater.
Criteria - Definitely open to more than one. I like @CGAcheer's suggestion of a combination of gym size, number of locations, and number of worlds teams.
Gym size - just level 5 athletes, percentage of level 5 to lower level athletes, entire gym, with / without prep...that part I don't know. I think using last year (or more) to test out the theories may help refine the criteria.
Number of locations - no gyms with multiple locations in D2
Number of worlds teams - a smaller gym that can field more than 1 worlds gym should be D1. (This would prevent larger gyms from spinning off a "Level 5 entity" of some sort.)
Worlds / Number of bids - Financially, does USASF "need" to give out the at-large bids to help fund the event? (This is my assumption, but you know what they say...) Based on this theory, simply reducing the number of bids to what would almost be the equivalent of D1 wouldn't work. I think we sort of need the D2 teams.
International teams - I know the ages for the divisions don't really mesh, but I feel some international teams could find a home in D2 somehow. It would be great to see some more international teams compete with similarly skilled teams.
Smaller programs who want to compete "up" - I am not against teams who want a shot at the "big dogs." Maybe requiring a few competition scores higher than some threshold percent of perfection to "appeal in" to D1? Though I don't know how much of an issue this would be?
Better to be 80% right than 100% wrong - I would love to see something put out with the expectation that it may / will need to be refined after trying it out, rather than it never happening because 100% agreement could never be reached (it won't.)