It seems that many of the performance issues could be processor related.
However the last device was, as you said yourself, supposed to be a second supplier for the Superhub 1, and a device that in practice, would perform the same as the original Superhub.
This new "Superhub 2" sounds like it's actually based on newer-generation hardware, and vastly improved over the outdated legacy components in the Superhub 1. That aside, the current Superhub would be hard pressed to hit 100mbps over wireless, let alone 200, and I believe even hardwired it won't do 400.
In my opinion, after 200mbps VM would have to replace the Superhub anyway, and as far as I'm concerned the sooner they do it the better.
Interestingly, Broadcom claim the BCM3383 is 50% faster than what came before, yet only runs at 500Mhz, suggesting the Superhub 1's CPU runs at 333Mhz. I've seen routers with 384Mhz MIPS processors become CPU bottlenecked when transmitting over wireless, in just single-band configurations.
Back of the envelope calculations suggest that an 800Mhz-1Ghz MIPS CPU to be the required minimum to efficiently operate concurrent dual-band wireless without the processor being a bottleneck. A multi-threaded/multi-core 500Mhz like the BCM3383 would obviously also do the trick. I do know however 684Mhz ARM processors are too slow to run dual-band concurrent wireless properly. The 333Mhz MIPS in the Superhub is thus in my opinion is barely sufficient to handle single-band wireless-N.
I do wonder then, if the Superhub's lack of dual-band concurrent wireless was impacted by the fact that VM put a processor in the Superhub that is simply too slow to handle it, similarly, its performance even single band could have been impacted by having a processor just barely fast enough to deal with it.