I think they'll find

that if they open just the datasets then the Open Source developer community will likely do many of the queries/mashups for them.

Or are they saying they haven't got any boxes capable of the actual workload? I bet Google would loan them a containerful of optimised data-munching servers at a fair price and even promise not to suck at the info out of it.

It does seem to be the Will rather than the Way that's lacking...