While a certain part of me could never say no to any Moon-based enterprise, the Moon seems barely more hospitable to data centers than humans.<p>With no atmosphere, the only two choices for cooling are straight into the Lunar surface or radiating directly into space, either of which seems like a huge pain in the ass.<p>Solar power is in some respects as good as it gets on the Moon, <i>except</i> for the whole two weeks of darkness thing. Even building near a pole, it's going to take quite a bit of infrastructure to keep the lights on in the data center, either in the form of massive solar arrays circling the pole connected by hundreds of miles of power lines, or massive batteries of massive batteries.<p>And at the same time, the antennas for collecting all that data from satellites and problems also need to be globe spanning. While a dark-side-of-the-Moon observatory is great for observing whatever it happens to be facing at the moment (which is valuable in itself), it's going to be pointing the wrong way from whatever you're collecting data from for two weeks at a time. So you need some system for buffering transmissions for two weeks at a time, or for retransmitting from an array of satellites in orbit to antennas on the ground, or an array of antennas spanning the Lunar globe.<p>So while I would totally vote for a Lunar datacenter if I were in Congress spending billions of dollars of other people's money, it seems slightly less feasible than an undersea datacenter. (At least with the undersea datacenter, you can cool into the ocean, and run the entire thing on the temperature difference between geothermal heat and the icy depths.)
Reminds me a bit of Robert Heinlein's "The Moon is a Harsh Mistress"<p><a href="http://en.wikipedia.org/wiki/The_Moon_Is_a_Harsh_Mistress" rel="nofollow">http://en.wikipedia.org/wiki/The_Moon_Is_a_Harsh_Mistress</a><p>The colony is dug underground in the moon and various resources are shot back to earth with a large catapult. And a super computer plays a large role in the plot.<p>My hope is that a private company may find profit in something like this, even if it is subcontracting for NASA or as part of the efforts of getting people to Mars.
In addition to the many, many other problems with this, the moon is pretty well unprotected from cosmic rays and radiation. Hardening computers against that is, to put it lightly, uncheap.