A Reddit user known as Mamolengo may have found a solution to the 12VHPWR melting issue. He created a custom monitoring program that triggers a command if the connector’s temperature exceeds a fail-safe limit. This would mean shutdown, killing a GPU-intensive app, etc. It could be achieved by combining his program with a dirt cheap ($8-10) 10k sensor. The 12VHPWR connector earned a bad reputation for melting after some users reported it. There is even a melting report on the PSU side, but it seems only a random incident. Although the actual failure rate is not high, PCI SIG will release a revised connector version. Welcome the 12V-2×6 connector. It is said that would resolve the known issues while being backward-compatible. This is something we shall see, of course.
Nvidia’s official statements for the melting issue are of user error. Even if that is true, it doesn’t explain everything. There are reports of melting even for Cablemod’s connector. Cablemod is one of the most respected manufacturers in its field. So, even if it is a user error (which might be the most reasonable assumption), that doesn’t mean much by itself. Regardless, the revised connector is here and hopefully will resolve this issue. Nvidia already adopted some modifications of the revised connector in its 4070 graphics cards. This might not be the best practice, though. Not-fully tested products might bring more problems than solutions.
I don’t know if Mamolengo’s solution is something we should trust. There were other suggestions for this issue, as well. Not every suggestion was legit, though. I believe the best practice is to see if the 12V-2×6 connector is the answer we should expect. I would feel in the comfy zone. Until then, stay safe, stay in the know.