Do we need more original research into proprietary hardware?
The profit margins of a high margin hardware vendor over a white box shifter depend in the medium term on the continuing demand for hardware innovation.
High demand, big market for cutting edge gear that white box shifters can't match, big profits to plow back into research to continue the cycle. As demand drops off, however, the white box vendors' ability to perfect last generations stuff at much lower costs becomes relatively more attractive and R and D spending tails off.
It's the same curve that any technology goes through, and smartphones are rapidly moving to peak innovation too as low end devices start to be able to do most of the things anyone could want and new features from Samsung and Apple look increasingly less compelling.
My take is that software is the new hardware, and we're seeing a rapidly broadening arms race in terms of differentiating features on top of increasingly commodity hardware which is driving software R and D spend through the roof as hardware R and D spend declines.