This is done on some motors, but I would guess more for reducing brush overlap between comm segments and maybe slightly changing timing than increasing RPM, although all "could" be the result of trimming the brushes. With the physics involved though, just trimming the brushes increases the pressure on the brush face without increasing spring tension. Think of it this way, the pressure a spring exerts on the brush face where it contacts the comm would be expressed in units of force per square unit of measure. With the same spring force applied to a smaller area, the force per square unit of measure has increased. A practical example of all this is... say a woman steps on your foot with her heel while wearing a pair of sneakers... no big thing. Say the same woman steps on your foot with her heel while wearing very skinny high heels... ouch!
Anyway, reducing the brush face and increasing spring tension may not be such a good idea because the pressure on the brush would effectively be increasing a whole lot. There's a lot in play in all this. Generally, the size of the brush face is important compared to the size (diameter) of the commutator. Small commutator and big brushes = no bueno. Big commutator and small brushes = OK, but maybe not optimum. Small commutator and small brushes = bueno, big commutator and big brushes = bueno.
All this gets into the general subject of "commutation," which, to my mind, is probably the most important aspect of how and why a motor works and performs the way it does. The brushes (and the material they're made from), the size of both the brushes and the commutator, timing, how (and how well) the coil wires are connected to the comm tabs, spring tension and material... as well as the number of coils, the angle of the legs, even the metal the comms are made from (*most often copper, but tin-plated and even silver and I guess others), even the body of the comm (usually phenolic, but I remember that even anodized comms have been (or maybe are?) used... all sorts of minutiae have a really large effect on the motor's performance. I have a set of the old LaGanke "silver" brushes I never tossed just because they reminded me of all the old stuff we tried for that extra edge. The label says 93% silver and 7% graphite, and while both copper and silver are soft, malleable, ductile metals, I figure silver brushes are probably pretty hard and would take longer to break-in and might cause premature comm wear. Even among brushes of the same basic composition, there can be changes in performance characteristics that are huge because the sintering process used to make them can have some variations if not done well or something.
Anyway, this is one huge subject, when you look at everything involved with brushes and commutators. Anyone can wind an armature, but producing race-winning armatures is literally rocket science.
PS: Since all the current a motor "needs" must pass through only the brush contact area, it's possible that a motor with reduced contact area brushes would not be able to run optimally being "starved for current."