One of the pertinent comments on this I saw pointed out that all the torque is being transferred via 2 teeth.
Seems a recipe for broken teeth or jumping teeth.
People keep trying to re invent the bicycle transmission but the donkeys year old design still keeps going.
Does anyone remember the one on tomorrow's world years ago? It still used a chain and sprocket but the rear cassete sprocket had a segment that pivoted in or out causing the chain to move to a different sprocket. It did away with the derailleur and was supposed to make gear change smoother.
That's been doing the rounds on cycling social media. It's a novel idea but still in its infancy. There's no give in a rigid drive like that so what happens when any grit gets entrained?!
It looks great but I sincerely doubt the 99% efficiency. I wouldn't expect it to be anything like as efficient as a roller chain. For a start, this design has to scuff an arc at the points of contact since the roller can only rotate in one axis. It must therefore have some slop built into the system to accomodate that movement.
Here are two additional roller bearings upporting each end of the drive shaft, each of them is being forced at 90 degrees to their axis of rotation. The front bearing will be being pushed up while the rear is driven down. Since that force is not contributing to the intended Axis ofrotation, it is merely being resisted by the bearings (and the frame). On a motorcycle, this causes the bike to lift under acceleration as the suspension is pulled apart by the gear climbing up the pinion.
Even if it was as efficient as a roller, it doesn't look at all robust either.
Im not sure what problem this design is attempting to solve. I dont believe the efficiency claim (at all) so I am assuming its just a clever gimic.