Data curation has been a topic cropping up at conferences I have been to this past year. I've heard it mentioned in sessions at ACRL and ALA, mostly by librarians from the big ARLs.
"Sources at Google have disclosed that the humble domain, http://research.google.com, will soon provide a home for terabytes of open-source scientific datasets. The storage will be free to scientists and access to the data will be free for all."
"The storage would fill a major need for scientists who want to openly share their data, and would allow citizen scientists access to an unprecedented amount of data to explore."
I still have to wonder how this will be monetized. Or, will this project be underwritten by Google's main revenue stream? Guess those institutional repositories will still have some room in them after all.
Read the full story with links to more details at "Google to Host Terabytes of Open-Source Science Data" on Wired Science.