It's a fair comment that the price of XMOS is pretty irrelevant overall.
The chip itself, the sockets, support components etc may seem cheap, but when you start piling up all those "little" costs on top of each other...
But more importantly, and aside from components costs, added complexity means longer R&D and testing time (all done by an external firm on a consultant basis AFAIK, charging money for their time, unlike when, say, bPlan or Acube are doing it in-house for themselves) and potentially a more expensive PCB with more layers. This can be a major contributor to the end-user price tag, especially on extremely low volume projects, where *all* R&D and testing costs would have to be divided on just *a few hundreds* of units.
So I hope there is a plan for it, that there is an obvious point (at least obvious to *themselves*, since *nobody* outside seems to be able to come up with an idea where this makes sense), if not, then I would label it as a design flaw that played its part in putting the consumer price tag way out of reach for most potential customers...
