"Data smelter" != “climate computer"
What Simon advocates is efficient use of computing vs energy, and I would not call them data smelters. A smelter produces a useful product. What comes out of a lot of these energy-to-bits conversion facilities is crap.
Microsoft is building data centers near Washington's Grand Coulee Dam for their power. And just who wants to use their services? Not enough people to justify their construction. Microsoft is good at wasting the money it makes from Windows and Office. MSN hasn't made a dime, and it probably never will.
Buffett needs to get an accountant to look at the return on investment for these money pits. Once he does some simple accounting, he will quickly see that Web 2.0 and fluffy-minded computing is not a good investment. Google makes its money from advertising. Can Microsoft take a chunk out of that? I'm not sure they can.
Google came up with a great solution to a sticky problem. Microsoft has done well with "me, too" but they aren't finding any real solutions to serious problems. Microsoft does not have an internal culture of innovation. Far from it, the culture works to stifle innovation and discourage success. Good people are motivated into leaving Microsoft. This is why Ballmer is trying to buy innovation, even if it is failing innovation (like Yahoo!).
Do your homework, Mr. Buffett, and keep your money in your pocket.