To a great extent, component compatibility is unpredictable, and the only way to really know is to actually try different combinations. However, as a general rule, I have found that the idea of mixing and matching so that the strengths and weaknesses of one are "balanced" by contervailing strengths and weaknesses of another hardly ever works out; one tends to lose the good qualities of each as a result rather than attaining balance.
I have tried solid state linestages with tube amps, as well as the other way around, and while some mixes worked out reasonably well, for me all tube worked the best. Some of the worst combinations I tried were tube linestages feeding solid state.
I've only heard the Bel Canto Class D amp and a very expensive Rowland (@ $14k). A bunch of friends had fun trying the Rowland in their systems. The Rowland worked best in a system where it was fed by a balanced Hovland preamp. While "not bad" in this system, it did not sound great either. The top end seemed shut down and microdynamics were slightly blunted.
In the other friend's system, which is a very sensitive and very picky custom horn system (fed by a custom tube linestage), the Rowland sounded much worse. It had an odd, unnatural sound that I could not exactly place what was wrong (not an obvious tonal problem). Again, the top end was almost completely missing, but, more troubling was a loss of the sense of notes naturally blooming into space and decaying naturally. Musically, the system lacked "soul" and sounded lifeless compared to what it sounded like with very small triode amps. It may be the case that the amp would be at its best delivering much more power as these speakers require very little power (they are about 106 db/w efficient).
Both of these systems have been carefully constructed over the course of many years, so a quick drop-in test hardly says anything about the inherent qualities of the Rowland amp. But, it does demonstrate that they are hardly the panacea for any ALL systems.