Cloud rendering is a fantasy for games on flat screens, because the latency is too high. To maintain 'presence' and not get sick in VR, the motion-to-photons latency has to be _consistently_ <20ms.
360 video for VR is certainly a bandwidth-hog, but I think that could well be offset by most VR content being game-like, where, though the game might weigh in at 40G, you download it once, and spend 40 hours in it, vs a 4k movie at the same size, which lasts 2 (and which you'd likely re-stream if you watched it again). In other words, widespread VR use, even with next-gen hardware, could actually lead to a reduced demand for bandwidth.
Anecdote: I used Erlang, over the course of several years, for very-high-level bot behavior and multi-player mission control, where performance was much less important than the things Erlang provides, but I still had issues with it. Nonetheless, I felt it had given me great leverage, and I thanked Joe Armstrong profusely, when I met him at a conference. Later at the con, I asked a panel whether there were active efforts at improving performance, perhaps with a JIT. They seemed to take it as an attack, and suggested that if I need performance, I should use another language. I mostly write in Clojure now.
I don't have time to read through the thread today, but I did before the Rift kickstarter was announced, and my impression has always been that the software-distortion-in-compensation-for-simpler-optics[1] idea was already out there when Carmack chimed in.
1. Allowing the screen to be much closer, drastically improving practically achievable FOV.