I think it rather likely that a test fine enough to discriminate changes the rather low powers involved, will discover that on a road bike any kind of a flexible link (even the short flexijoint Slow Johnny posits) will lose enough power to cancel any gain all wheel drive offers, and then some. Also, the weight weenies won't be pleased by what a flexible coupling weighs, even if it has to transmit only 250W.
On the whole, the entire AWD bike, and this flexible coupling, is a solution in search of a problem,
***
More evidence, albeit only a thought experiment rather than empirical experience:
At first sight, it may seem as if a balloon-tyred bike could use all-wheel drive. But, in fact, my 60mm Big Apple shod bike has so much grip at speed on really twisty (as distinct from sweeping) turns on downhills that the problem is too much grip at both ends, causing safe understeer, rather than too little grip, causing fast oversteer which is dangerous but exhilarating for cyclists who can control it and aren't frightened of taking a spill. More power in the front wheel (in which there is presently none) will just aggravate the problem; what you want is less grip at the front so you can induce smooth push-oversteer from the rear wheel.
I was hoping that a bike which already corners exceptionally fast would be elevated to another class altogether when I switched from front hub-drive to centre motor/rear drive, but it turned out my fat tyres screwed me; it seems quite possible that to corner faster by means of all-wheel drive I'd have to go down to tyres so narrow that they would be dangerously unstable all the time, and godawfully uncomfortable besides, so there would be no gain in extreme conditions, and in normal conditions I'd lose efficiency, speed and handling (handling is what you do when you exceed the roadholding of your bike and it all starts to go wrong).
Interesting bike, Ridealot, in a Ripley's sort of way, but not very practical.
Andre Jute
Amused but not impressed