if it is designed to breakup. And not if it isn't.
>no nuclear material involved
that is the beauty. No contamination.
>that's not going to have a huge impact.
in my comment i already specified the TNT equivalent of such an impact.
>there's a 0.7 probability
It isn't a matter of probability. You can deorbit with high precision, and pretty much hit any desired target on the ground if you have thousands of objects in space on a bunch of various orbits.
>Should we ever get to a point where a country is considering shooting down space datacentres, considerations about the impact on Earth is unlikely to stop them.
13 ton GBU-57 reaching M 2-3 gets 200 feet deep. De-orbitted 1-2 ton steel rods will have about the same effect - ie. you can hit many strategic objects of your attacker. And having in orbit, just in case, a ball or rod of 30-50 tons will get you a small tactical nuke equivalent.
"In the case of the system mentioned in the 2003 Air Force report above, a 6.1 by 0.3 metres (20 ft × 1 ft) tungsten cylinder impacting at Mach 10 (11,200 ft/s; 3,400 m/s) has kinetic energy equivalent to approximately 11.5 tons of TNT (48 GJ)."
I still believe that it is possible to do noticeably better, yet it would take some technological effort - somewhat similar to how Russian supercavitating torpedo https://en.wikipedia.org/wiki/VA-111_Shkval does or like some hypersonic weapons do by creating a plasma cloud ahead.
Even more - we can create plasma channel by using tandem schema. The first, smaller rod flies down heating and ionizing the air thus "paving" the way for the second, main, much larger rod. The delay between them is such that the main rod passes almost immediately right behind the second in the a-max (maximum deceleration) layer between 3 and 5 km altitude. If they enter 5km at 20 Mach, the first will slow down to almost 10 Mach at 3km, so the main should enter the 5km altitude 0.2 seconds behind the first. The wind speed at those altitudes is 10-20m/s, thus the channel has to be only about 2-4m wide. I actually have been thinking about this from the other side - a large gun for space launch.
The orbital kinetic strike weapons that have been proposed in the past are usually 2 ton titanium rods that would hit at about Mach 10, and even with that level of force they've been dismissed as less useful than ballistic warheads. Things falling from space just aren't as dangerous as people tend to assume.