Sounds like a current issue. While heating, it draws more current, which causes the boot loop. Once it reaches set temp, it's not drawing much current, and no boot loop. The psu isn't current limited, current throttling and sensing happens in the firmware/microcontroller, and since the unit updated fine, displays fine, and holds temperature once it reaches set temp, the microcontroller is generally fine, but there may still be an issue. It seems like the GD32 current limit is being exceeded on start up. If you would solder something big that draws a lot of current, it would probably cause another boot loop. My guess is it's one of 3 things- possibly a stray ball of solder bridging some of the current pins on the microcontroller, an actual fault in that part of the microcontroller, or an issue with one of the capacitors on the power supply (a bad cap, or a bad joint). You're going to have to open up the back, pull out the power supply, inspect the joints (take pics and post them if unsure), and even make sure the capacitors are oriented the right way (see attached image). Make sure you discharge the caps first too! While the board is out, you can see the controller board through the back. Use a good light and magnification to inspect the GD32 (the large black square on the front control board) looking for stray solder balls bridging any pins. As shown in SteveyG's video, getting to the controller through the front is kind of a pain. You have to heat the front panel to warm up the double sided adhesive, then pry it off. That can complicate things as far as returns go, so it's best to go through the back for now.
And yes, if the big capacitor wasn't in the right way, you'd know immediately. Just beating people to the punch.