http://facweb.knowlton.ohio-state.edu/pviton/courses2/crp776...
Skip to page 4, where it covers ESAL. A standard trailer is expected to do around 800-1000 times the damage of a standard passenger car. The conclusion:
"When discussing road wear cars don't matter: road damage is effectively caused by trucks."
So, if road use is charged by a steeply progressive schedule, and driver cost is not an issue, will we see fleets of light autonomous trucks replace the heavyweights?
That's possible too, but would be tremendously less efficient as you'd have the same air resistence (or at least a substantial amount) but without the same cargo capacity.
Railroads tend not to develop potholes.
Also, remember that roads aren't made out of head-hardened solid steel - if they were, they'd probably nearly forever.
Rail itself may not get quickly damaged (though they do) however rail wheels, which are made of relatively softer metal, develop all kinds of damage, require the machining away of the damaged surface periodically (as in, wheels shrink over time as mateiral is cut off it to get to new metal below) and only last a matter of years for heavy haul environments: http://www.asa.transport.nsw.gov.au/sites/default/files/asa/...
Search for spalling, shelling and rolling contact fatigue as examples.
(There might be a paper in figuring out the sweet spot given hypothetical costs ...)