A USC grad student is proposing that we build a supercomputer at one of the lunar poles:
http://www.wired.com/wiredenterprise/2012/...rcomputer-moon/
The idea is to alleviate data overload issues with NASA Deep Space Network. It's a creative idea, although his proposed time-frame is too ambitious. Also, wouldn't it be cheaper to orbit an array of spaceborne antennas (at a Lagrange point) to act as a front-end for the DSN?