If you look at the figures in this verge article it looks like the S3 costs were almost immaterial.
http://www.theverge.com/2013/11/5/5039216/everpix-life-and-d...
http://www.theverge.com/2013/11/5/5039216/everpix-life-and-d...
Even more amazing to me is why someone like Apple didn't purchase Everpix. They're in desperate need of competent web services, and god knows they have the cash. Of all possible futures for Everpix, I could never have seen outright shuttering as a possibility.
Tools for private photo collections are important, but don't have the same sort of explosive viral growth curve of, say, a public social network. Even private social networks have a more difficult time scaling up quickly compared to the more open but less privacy-respecting alternatives.
Maybe I'm still in mourning, but if I ran a VC fund, Everpix seems like a sure investment. You certainly had much more potential for growth -- and steady revenue -- than another Instagram clone or internet radio service. And once your growth eclipsed your fixed costs (which could have been reduced further by running your own servers when you reached an appropriate size), you'd be consistently profitable. And since Everpix provided such a valuable service, your customers would advertise on your behalf.
However, even I discovered Everpix only a few weeks ago, just before The Verge published a write-up about cloud photo backups -- and crowned Everpix the winner. You were just reaching a critical mass of followers... I can't help but think Everpix would have been a wild success if it stuck around for just another year.