In order for friction to destroy steel, it needs to actually wear it away one particle at a time.
Not entirely correct. While that may be the most common aging / failure method on a road car. On a race car, heat effects are what kills rotors -- of all construction. When you heat a steal (cast iron) rotor near (or past) the glass transition point (the point where it "melts", or transitions from solid to liquid) it will wear quickly and unevenly, begin to warp, and start developing cracks. Look at any used rotors from a race car; almost all of them have small, spider cracks in the braking surface from the repeated heat cycles. (heat causes metal to expand, but the heat isn't applied evenly over the entire disc, and it's a circle so the inside will expand more than the outside.)
But yes, in this equation, mass wins. Carbon fibre is great for many repeated, brief, super high heat cycles -- which is why F1 uses them. In this case, it's one HUGE prolonged dump of energy. That sort of thing will shatter a carbon rotor.