DARPA, I think, have a pretty good track record of funding intersting and useful stuff, but also early stage, low Technology Readiness Level research. And a pretty good track record of their funding turning in to new products eventually.

Faster cheaper computers would be good in all sorts of ways.