Try a couple of things that you'd normally discard and see if a bad idea doesn't lead to a good one once you get a bit of user feedback.
I have some IaaS ideas I've been kicking around but they are all capital intensive. The only way I can make it financially feasible is to build my own servers. And with no SLA or warranties I would basically have to live at the datacenter.
Really, who needs online file synchronization when a USB key and rsync are perfectly fine!