This technology sounds like it would also allow high density optical computers with similar architectures (therefore well understood) to convetional computers. This would bridge the gap between the fastest conventional computers and quantum computers. This is important because capacitive and thermal losses are dominating silicon now and quantum computers (at least for the moment) are not good at solving conventional computing problems.
You could argue that optical PCB's are already available. It's that there aren't really any need for them. The whole point of an optical link is that it can go fast enought so that you can use serial transfer to communicate at high speeds. However, most chips don't work that way. If it made sense to build optical trancievers into chips, then optical PCB's would instantly become commonplace.. Chips are heading in the opposite direction. Jillions of IOs and interconnects. vertical integration is helping but an architectural paradigm change has to happen first. Perhaps this graphene technology can be an enabler. But way the heck down the road.
The article says "The researchers went on to pattern the materials into compact domains, which acted as terahertz frequency plasmonic waveguides with negligible cross-talk, even when packed as closely as 20 nanometers", if it can allow to travel two streams with 20nm distance it can be explored for optical PCB. By the way this is just an elementary thought, may seem impossible as on today.
Looking at the last sentence of this post "Funding for the project...by the National Science Foundation and the Air Force Office of Scientific Research", it could be guessed that this technology would find its application in the "robotic eye" for video surveylance...or who knows, this might take its place in the eyes of the next gen robots?