Ancona, you're incorrect about that. Terminal velocity is a function of wind resistance, density and other things, as well as time of fall.
Gravity accelerates something at ~32 feet per second -- per second, not just 32 ft/sec (you're leaving out the last term there - seconds is squared) from infinite height - the speed keeps going up till wind resistance limits it.
Else you could jump off the Empire state building without a chute and survive - that's only about 21.8 mph at 32 ft/second. It would hurt, but not kill you.
The sectional density of a tungsten (or tungsten carbide, which is what they were talking about) is much higher than a human body, the latter reaching about 120 mph or so terminal velocity at density roughly one, and wind cross section about a square foot (head on, you can slow to about 70 or less, much less with a "squirrel suit").
Tungsten is a bit denser...about 12x, and would be assumed to be less than a square foot across if ballistically designed (like a big bullet).
The link you gave doesn't allow drag coeff as good as this would have, but at its limit it gets 15k feet per second from orbit. That's one fast bullet. I don't trust their math, having worked the much more complex stuff for hypersonic projectiles they left out, but at any rate - it's fast as heck.
Lets assume we have a tungsten rod - say 10,000 pounds or so (about the size of a phone pole I'd guess). At that weight, the old Space Shuttle could carry 6 of them, and some retro rockets to de-orbit them. Now we want to lift it to low earth orbit. That's going to take 10k foot pounds per foot we lift it (and we get that back when we drop it!).
Ok, LEO is about 90 miles. ~475200 feet. Times 10k. Or, 4.752e+9 foot pounds of energy - on each one. (For reference, a .223 bullet is about 1100, a .308 about 2500, and a .50 BMG about 20,000 fp).
This is neglecting the other energy you have to put into it to keep it orbiting, which is also huge - there's a big difference between reaching space and getting a parking orbit. Depending on how shallow a re-entry we can tolerate, we can keep a bunch of that energy on our projectile too, when we drop it, as sideways motion.
A foot pound is
1.355781 Joule, or watt-second. Or in more interesting terms,
2.939635 *10-10 metric ton of TNT. So this multiplies out to about 1 ton of TNT.
But unlike TNT, which sends energy in all directions - very little goes into say a 1 degree solid angle, this is all going one way, and it's going much faster than the speed of sound in rock, which makes the rock look like a liquid to it. Try this on your shooting range - any bullet over about 2800 fps goes though steel - even a plastic tipped varmint bullet, and leaves it splashed up on *both* sides like that stop motion picture of a milk drop we've all seen, whereas you can shoot one the same foot pounds, but slower, and all it does is dent it (Say, .223 45 gr varmint vs 454 casull heavy bullet, and 3/8" thick steel). it turns out the that average speed of an iron atom in molten iron is just 2780 fps...
So over TNT we get a huge concentration of energy all in one direction - hundreds to one. So the target sees hundreds of tons of TNT energy per thingie...that's a lot.
So taking all these effects into account, yes, a tungsten phone pole dropped from space will indeed penetrate 100's of meters into rock, "melting" it as it goes, even if the projectile itself is destroyed in the process - no fragile warhead required. And the resulting shock wave will pulverize any open cavity underneath. This has been tested now...but we've given up the capability to put them up there at this point, though you could just use the Shuttle engines without the Shuttle, as NASA has proposed for other missions.
I guess they were afraid of what would happen if they lost control of one out there in space...could make quite a mess, in a very small area.
NASA does understand hypervelocity stuff and the nonlinear effects of a fast projectile though. You muzzle velocity lovers, read and weep.
http://www.nasa.gov/centers/wstf/laboratories/hypervelocity/gasguns.html