Senior Google engineers are aware of SAFE and are interested in the engineering sense in that it may help them organise small parts of Google’s internal infrastructure more easily. Some of Google’s efforts with QUIC have strong parallels with RUDP in Maidsafe for example.
Google however have very little interest in a non-centralised resource based internet for obvious reasons, though I would doubt they would actively work against SAFE so long as they are allowed to index content. I know they think Safecoin “has a point” unlike Bitcoin which is seen as pointless and inevitably doomed in the long run. They also think that some form of token proof of networked resource is inevitable, but are not convinced that Safecoin is that. Basically, they’re taking a wait and see for the wider scheme of things, but doing R&D on many of the same low level issues as MaidSafe implemented years ago.
I might add that Google don’t store personal information (specifically info protected by law) in single or central locations, they chop it up into bits just as SAFE does, and only the bit necessary is available to the internal services which consume it. The access keys to see the entire database at once are available to a surprisingly few people in the company, I am told less than twenty. This strategy has to date meant Google has contained leaks of personal information better than most.