I'd probably need 200GB ($200/year). A more reasonable person would probably be fine with 20 or 50GB ($20/50 per year).
I'm not sure I understand the business model and the hype behind cloud syncing of music though. What is the point? If you're going to copy onto an MP3 player you'll need a local copy. So the use cases are:
1. On the Web;
2. On networked mobiles; and
3. As a form of backup.
Well (1) is covered quite well with Grooveshark. I can find most things I want there and it's an awful lot cheaper (up to free).
(2) I don't think makes a lot of sense given the high cost of mobile data. Maybe in the future mobile data will be an awful lot cheaper but there are fundamental limitations with wireless bandwidth that I think will make that very difficult.
There are many solutions for (3). In terms of raw storage, Amazon's prices are pretty cheap. But backup misses the point entirely I think.
When I buy digital content of any sort I don't want to back it up. I want to be able to recover it easily and simply. iTunes for example only allows downloading movies once (is that right?). If so, I'm just never going to buy movies that way. If I pay for it and can watch what I download any number of times, why can't I download it again if I accidentally lose it?
So iTunes and Amazon MP3s need this feature: log onto my account and click a link that says "download all purchased tracks".
At that point I don't need backup of any kind (for my digital content).
I think the only business model that makes sense is flat-rate subscriptions. You don't store your own music. The provider simply has all the music. This solves a lot of storage problems for the provider (meaning 1000 people share the same copy of the song rather than each uploading and storing it individually).
I can see how they'll get some scale here by having duplicates of some songs (particularly iTunes and Amazon bought MP3s). It'd be interesting to know how much saved space they have from deduplication.
Anyway, am I missing something here?