It looks like the data was already in S3: [1]
The recovery probably would have been a lot faster had they had a fully provisioned EC2 image standing by... which I'd bet they will from now on.
[1] - https://news.ycombinator.com/item?id=32032316