for taking out bunkers as deep as NORAD or as strong as Mt. Weather (as Bush has in mind when he came up with the idea of nuclear bunker busters), a nuclear weapon is needed. A space-based railgun would be expensive, but it can be useful. If the bunker is near a populated area, the plasma produced by the vaporized tungsten rods would produce a shockwave, although how powerful, I do not know. A cheaper alternative is a ICBM with the nuclear warhead traded for a kinetic penetrator or even the DDX's Railgun whose rounds will also fly up into space before dropping onto their targets.

The biggest problem in converting ICBMs into conventional strike weapons will be convincing everyone with the ability to see your launch plume (via spysats) and flight tragectory (OTH radars) that you are indeed launching a conventional ballistic missile and not a nuclear one. Why not further develop the MOAB? In WW2, the British had the massive Grand Slam 22,000lb bomb that could burrow 100 ft down before detonating, creating an ideal bunker-collapsing tremor. Why not figure out a carriage for a penetrator MOAB in a B-52? (I don't believe the B-1 or B-2 bays are long enough. Am I wrong?) Dropping a 20-25K lb penetrator bomb from 40-50K ft altitude should sink it into the ground nicely, reinforced concrete or not..

The energy of these falling objects is overrated. Projectiles dropped from orbit or suborbital trajectories will not enter the atmoshphere vertically. They will come in at shallow grazing angles forcing them to traverse many times more air than if they came in vertically. Assuming they retained any significant amount of their reentry velocity, they would also hit the ground at a shallow angle, so a super deep penetrator may go no deeper than a conventional deep penetrating bomb. Dropping "stones" from geosynchronous orbit is not very time sensitive. If you arrested the projectile's 6,000 mph orbital velocity, it would take 10 to 12 hours to reach the surface. Send a bomber. It's faster.

Hold on a second. Geo sync orbits are at 23,000 and some odd miles, IIRC. Dividing by 6 k gives me about 4 hours allowing for accel at 1 g. That really is not the show stopper. Money is. There are a lot of technical tricks the US military can do that aren't done becasue they are to expensive. Take metal storm. That would have been a great invention 400 years ago. But in the age of chain guns it is a waste of money. No, there is nothing an Orbital rail gun can do that can't be done cheaper by something else. Iraq has shown that the looney left is going to whine no matter what is done to limit warfare. And yet the need for war hasn't diminished. So eventually the Looney left will have to be captured and turned over to their heros. Remember the squeaky wheeel get greased, but only up to a point. Then it gets replaced. If Osama was to make Michael Moore his bitch, then neither one would have the time to raise hell with the USA. Two problems, one solution.

http://www.popsci.com/popsci/aviation/article/0,12543,636378,00.html A railgun would not be practical as a satellite for reasons others have mentioned. Lacking a large mass (like a railgun on the moon would have) it would suffer from reactive force for each round fired. Furthermore, when shooting projectiles from orbit to the planet surface, one needs to merely rely on gravity. I don't see how the difficulties of reentry would be a significant barrier to kinetic energy tungsten rods. They would be fairly streamlined, and the only part that needs significant shielding would be the guidance electronics. I recall reading about the "orbital satellite with rods" in a sf collection edited by Jerry Pournelle called "There Will Be War" when I was a kid. It was known as "project high frontier" and an essay about it was co-authored by Robert Heinlein & Lt. General Daniel Graham...