The problem is the presumption that the data doesn't have a physical location when you are dealing with a cloud. You may not directly know where a given hunk of data is physically stored at, but such storage is still a requirement for current computing practices. It can be destroyed, confiscated, lost, or even simply scrambled where you have no control over what happens. It can also be copied and distributed to places which may not be in a place you want it at (like a competitor or somebody who intends to do you harm).
Keeping data in a cloud is fine for temporary stuff or for data that is of a transitory nature that might be discarded a day or two later. Also if the data is of a nature that if it is published on the front page of a newspaper or on Wikipedia, nobody would care.... you generally don't have a problem. If you really want to keep the data for any length of time... due to legal requirements or even something that is vital to the mission success of your company or organization, it is really idiotic to rely upon 3rd parties who don't have a vested interested in your success to be keeping that data.