Yes, it appears I slipped in an errant "these" in my original question that did however reference ribbons. I've edited it for clarity, as I was speaking about risers in general, so sorry for the confusion. USB carries power obviously, so it's impossible to know if the 12v line is connected back without actually asking the question.
Irrelevant honestly. I've read those threads. I run GPUs which draw 350-400w per gpu on unpowered risers, some of them are 60cm long, all without issue. So do other folks with large farms in china and elsewhere.
The REAL problem is when your PCIE power rail is failing to produce enough current for the GPU's draw, so then it attempts to pull more from the PCIe socket. If the cables have varying quality, especially on the solder joints, you'll end up with the cable's 12v line acting like a fuse. If the motherboard is damaged from this, then the PSU is/was failing (likely the 12v rail), and/or it wasn't properly spec'd for the current draw.
Folks can't just buy a PSU that says X watts and presume it'll work. The important detail is the 12v rail *constant current drain* raiting. I have some which are rated for 27-29amps per GPU, and some which are 35amps per GPU. I match the appropriate rail with the appropriate GPU depending on current draw.
Coincidentally, it doesn't matter if these or any products "back feed"--that doesn't even make sense. DC flows in one direction, and there are plenty of diodes on GPU PCB's to prevent such a situation. You'd fry the tracer on the PCB before you fried a cable if there was a backflow issue. The tracer is thinner than the PCI riser cable. This is all supported by many of the photos, where the riser burns up at the slot--the current is drawing/trying to draw from the slot--likely, again, from a PSU lacking the ability to provide current. If it were a back-flow issue, the tracers would poof.
Not trying to hijack the OP's thread, but just provide clarity on a wives-tale.