This connector somehow has it's own Wikipedia page and most of it is about how bad it is. Look at the table at the end: https://en.wikipedia.org/wiki/16-pin_12V

This connector somehow has it's own Wikipedia page and most of it is about how bad it is. Look at the table at the end: https://en.wikipedia.org/wiki/16-pin_12VHPWR_connector#Reliability_and_design_changes https://en.wikipedia.org/wiki/16-pin_12VHPWR_connector#Relia... The typical way to use these is also inherently flawed. On the nVidia FE cards, they use a vertical connector which has a bus bar connecting all pins directly in the connector. Meanwhile, the adapter has a similar bus bar where all the incoming 12V wires are soldered on to. This means you have six pins per potential connecting two bus bars. Guess how this ensures relatively even current distribution? It doesn't, at all. It relies completely on just the contact resistance between pins to match. Contrast this with the old 8-pin design, where each pin would have it's own 2-3 ft wire to the PSU, which adds resistance in series which each pin. That in turn reduces the influence of contact resistance on current distribution. And all cards had separate shunts for metering and actively balancing current across the multiple 8-pin connectors used. The 12VHPWR cards don't do this and the FE cards can't do this for design reasons. They all have a single 12 V plane. Only one ultra-expensive custom ASUS layout is known to have per-pin current metering and shunts (but it still has a single 12 V plane, so it can't actively balance current), and it's not known whether it is even set up to shut down when it detects a gross imbalance indicating connector failure.