I wonder if we'd go back to non-threaded processors for a bit when this starts panning out...
I wonder also...
Will they not be trying match it against silicon plans, i.e same logic gates? They may not be taking this traditionally with the semi-conductor but it's still a base 2 system. Hard to say, I'm not a physicist or engineer so I suppose I wouldn't really have any idea what I'm talking about :(
Technology to build a 400 GHz transistor has been widely known for a very long time. GaAs transistors easily operate at those frequencies. The problem is how to make VLSI ICs of them at low cost as it can be done for CMOS. The problem (cost) is even bigger for graphene, so I really don't see any breakthrough.
// edit: I was wrong. GaAs transistors can operate at THz ranges... ;)
// edit 2: also there seem to be only little correlation between a maximum frequency of a single discrete transistor and of the whole IC created from millions of transistors. The basic limit with CMOS is *not* the maximum frequency of a silicone transistor (they are quite capable of running at several tens of GHz), but how many transistors you can put on the chip without the whole thing melting down due to excess heat produced by too frequent switching. Another problem is that to operate at higher speeds, the whole chip has to be smaller because of latencies to get signal from one point to another.
Therefore I'm not impressed by the 400 GHz value at all. I'd be impressed if those transistors could handle those frequencies at 1/10 of power used by a similar MOS transistor. Unfortunately it is still not the case.