I am only going to argue for the sake of arguing.
Typical 1.25" quality heim is rated at about 80,000 # of Radial Static Load. Centering the rear axle under a fully dressed 4 door JK, I doubt, could ever see anywhere near that type of load. The way that it is designed in the picture above will likely hold-up.
Not arguing, I figure you're always game for a reasonable discussion.
Agreed, they have a very high radial load rating that they will likely never see at the axle.
When a wishbone link is loaded laterially, the chassis joints are loaded radially, regardless of orientation (vertical or horizontal bolt). The axle link, assuming no deflection in the link itself, will be loaded horizontally with lateral loads on the axle.
With the single lateral constraining joint mounted with a horizontal bolt, all the lateral force is being applied to the joint 90* to the race. Arguable the worst possible direction to load the joint. Rotating the joint 90* with a vertical bolt would put the loading back on the race of the heim.
It would be interesting to know the load rating of the joint through the bolt (90* from the race/shank). I would suspect it is significanly lower, but quickly serching I can't find any mfg that gives a number.
For reference, I know 2002 GC's came with a wishbone 3 link rear. Even the clevite they used at the axle was mounted vertically.
I don't know if this crude pic makes more sense than my wordy explanation, but rotating the axle joint switches the load from axially through the ball 90* to the race, and puts it back on the race. Your theoretical point of rotation is the blue dot.