Perhaps there could be some continuous rollover with all data older than five years being made available through the Internet Archive. I'm no longer affiliated with Twingly but of course know them very well. I can make a proposal! It would be a great idea and I guess for Twingly it could mean increased brand recognition.
I didn't realise you couldn't download and use the data from Internet Archive. If not, that's pretty silly to back up the feeds to them, and I'm a bit annoyed to have contributed. I'd like to make them available to everyone to download, analyse, plug into their reader etc etc etc...
For anything substantial (like say, their actual crawl), they'll only do it on a case by case basis with a rather restrictive license and you have to drive up there and plop down the machines to copy it onto.