@Kaur:
STRIX has a terribly small footprint. I don't remember the correct data, but it's so small that a single target should better not move.
To aim at a road when a coluimn is passing might be promising, but some self-guided munitions have a tendency to not lead enough on fast moving targets afaik.

You have to be more mobile than the MBTs they are taking on, or else they can simply be evaded. I am also unsure of how useful the term "ambush" is. "Attack on a moving enemy" is useful, but lurking in a wood hoping the enemy is nice enough to drive by, is not the acme of tactical skill.
Hmm, actually that's one of the methods that produced the best kill ratios in the past afaik.
The art is to choose the right place and time, and to pull it off. It's nice to get away and repeat it elsewhere, of course.

The emphasis on mobility is a bit questionable imho. Getting away is important, but any emphasis on being faster is probably misleading.

The mobility-emphasizing tank destroyer concept of WW2 was not the expected success, whereas the StuG concept (always inferior in mobility to its major enemy T-34) that rested much of its AT tactics on old field artillery ambush tactics was a success (different environments and generally difficult to compare, but the opinions about the TD concept are afaik still rather negative).

The other reason is less military history than OR-like.
What does "more mobile" mean? It's terrain negotiability aspect is irrelevant in many terrains (not quite in mountains) because MBTs can already negotiate most terrains.
The speed aspect is the one that convinces me the least.
- not the tank's speed counts, but the tank unit speed
- 40-75 km/h depending on surface/type and depths of probably 5-15 km before the tanks do a lot of harm:
How much time does that give for leading an AT unit into a favourable position, probably more than once? Consider that the AT teams don't dictate the direction of the attack, they have time lags because they have to react.

A success in an AT mission requires imho either well-prepared ambushes, a lot of brute fireower or a combination of enemy mobility degradation and own mobility.
The MBT's speed is imho quite uninteresting. It's the unit's speed that counts, and that's under influence by some external factors (real and fake minefields, ECM against radio comm, intimidation by multi-spectral smoke walls - who wants to move into the unknown?).

Btw, what did you mean with "or else they can simply be evaded"?