The rotating mass of the turbine spinning at ~150,000 rpm on a shaft is one heck of a gyro. So as the bike pivots along its axis where the tires meet the ground (Think Knee Dragging) this gyro is generating massive torsional forces onto the shaft as we force it thru an arc as the bike moves from side to side.
With the forces at play if we viewed it in action with ultra high speed cameras we would observe the turbine and shaft to be squirming around like a spaghetti noodle and the turbine blades flapping like a plastic pinwheel. Bottom line being that although we see metal as something very hard the forces in there make it a very violent place. Add in additional heat due to pressure over-spinning it and it may just be over the edge of its design metallurgically.
But it is the bolt for the lower sprocket that is coming out, not a bolt at the turbine or any other problem at the compressor.
The deformations occurring throughout will cause the bolts to come apart no matter what you do.
The second less obvious possibility is the gearing change itself, All mechanical gears create vibrations. These vibrations are mathematical in nature and must be taken into account as the spinning parts are already under huge stress. Changing the ratios may very well create some destructive harmonic forces within the assembly. I'm sure the engineers had all this accounted for in the design.
It's certainly a fact that overgearing increases loads on the sprockets if operation is identical in terms of engine rpm, though of course up to about 180 you don't need to be at the limits of rpm or anything like to have all the torque the front wheel (wheelie) and rear tire can handle, so you don't have to have that last 10-15% of possible rpm or even close to it. Some will shift late regardless, some will shift earlier. If earlier, then where's the extra stress?
In practice, one guy without gears and a derestricted bike may average 10-15% more rpm than another guy with a derestricted bike and the gears -- Turbo329 for example doesn't seem to particularly wind his bike out when no need to do so, at least in his videos, while some like to keep the engine singing above 8000 all the time, even above 10,000 most of the time in spirited riding. So I am not convinced that in all cases the compressor spins faster for a given rider and Bear gears (sprockets) than with another rider on a H2R or an unrestricted bike with stock sprockets, which don't seem to have the problem.
Slop in earlier gears: seems to be a fact and certainly a contributor. Using a different thread locker than Kawasaki uses with possible difference in resistance to oil seems another possible cause. At least the part of using a different thread locker is a fact in some cases. Overtorquing, also possible in some cases. As opinion, a thread locker requiring particularly high force to break when intact really should not be needed, rather a thread locker is needed that retains its strength in these conditions. Slop in current gears? Seems unknown if that's still a contributor if instead the cause might be different thread locker washing out.
The fact that Bear seems to think a different bolt is required (for which they omit the oil passage Kawasaki determined was worth providing) suggests to me they are not confident in their new sprockets with stock bolt, though -- or they don't really know what's happening. Not confidence-inspiring.