If Wikipedia were to go under and nothing similar were to replace it (unlikely, but who knows) people might go back to using encyclopedias on dead wood or walled garden online encyclopedias both of which would lead to people spending less time searching.
I'm not sure how much of the money will go towards full time "wikipedia editors", for example.
So, if we believe the wikipedia marketing that most of the editors are "largely anonymous internet users who write without pay", what exactly are they spending the money on ?
If it's hosting and such, I'd rather see them not manage their own infrastructure and try to move towards the cloud (no matter how corny it sounds). Putting Wikipedia on EC2 or AppEngine should be transparent for users and might allow them to focus on their core competency (which would be what, btw -- crowd sourcing management ?)
Not trying to be snarky but I fail to see how this money is helping that much.
My remark was mostly triggered by the fact that here you had Google with a big cluster, a cloud offering and tons of bandwidth paying money to Wikimedia so they could buy their own hardware and bandwidth at a worse price probably. A direct deal would have looked better.
I suspect a pretty hefty sum; I've no idea what they spend on hosting right now but I bet it wouldn't be a drastic cost cut.
Edit: Someone posted in the original article about similar grants - http://en.wikipedia.org/wiki/Wikimedia_Foundation#Grants , including a recent $2 million potential "grant" funding from the Omidyar Network.