This assumes that "communication overhead" is proportional to the number of dyads; I can certainly see why one might suspect that to be the case, but I certainly don't see that it is implausible that that overhead would expand faster than that (I don't see an obvious theoretical foundation for literally "exponential" growth, but I wouldn't reject it out of hand, either. Of course, you'd probably want a clearer operationalizations of "overhead" to answer that question.)