Could you point out what in any of these services actually depends on the cloud?
Email runs perfectly fine on single servers (Gmail). Network file storage is a solved problem (Dropbox/Google Drive/YouTube/various photo buckets). What these things are missing are dynamic web 2.0 GUIs, not backends. And the GUI doesn't have anything to do with the cloud.
The cloud is a huge innovation, but it isn't an innovation that provides anything for us, it's an innovation for centralized businesses. The cloud has absolutely no value until you're trying to serve millions of users on one service. Federated systems like email are perfectly capable of handling everything that the cloud does for you and I.
We don't need the cloud.
I'd love to have a comparable platform that I could run from my closet, but that's still in the distance. For now, cloud services are usually the best way to minimize cost.
This is a perfect example of what I'm saying: this has nothing to do with the cloud, and everything to do with the tooling Google Apps provides. There's nothing preventing the open source community from implementing equivalent tooling which targets federated servers.
Google's services, at least, are written assuming they are running on a million machines inside of a Google datacenter, with all sorts of dependencies on the infrastructure; worse (for the purposes of converting them to easy-to-install packages), they are written assuming the existence of Google SREs.
Everyone else is in the same boat, with software meant to be run on a datacenter and not a MacMini. You'd really need to start over from scratch, and first with a platform that can run on one to a million machines, then rewriting everything to run on top of it.
It's not impossible, but it's a lot of work.
[1] Such a movement probably already exists, in which case you should join it.