For those interested, I asked around and it looks like the entire Twitter collection is only going to be 5TB worth.
I calculate the tweets data must be around 3.5TB so that figure seems about right (if you throw in user and times tamp data)
I hope they do elect to store the actual URLs when URL shorteners are involved. Really, though it would eat a hole in the URL shorteners' business, I really think sites like Twitter should be dynamically 'fixing' those links before inserting them into the database.
I've been curious. Does this include the tweets of "private" accounts?
The article mentioned that only public accounts would be included.