I did a dumb thing today to a RPi 3B, applying +12V directly from a 14 Ah battery to the +5V / GND GPIO pins (and in fact it was through a good soldered connection, not just spring contact wires). Unsurprisingly there was small jet of flame from the PCB, as this battery can probably deliver 50 A in a short burst. https://photos.app.goo.gl/PD5EG7X9pQoN9CQa6
The board is (literally) toast, and bad-smelling toast at that. But quite surprisingly, the SD card the RPi was running from at the time, still works when I tried it in a different Pi. Likewise the USB Audio adapter I had plugged in at the time also still works. I'm puzzled by how that's even possible, as I would expect putting a high-current +12V source directly on the +5V rail would blow every "silicon fuse" much faster than any physical fuse or pcb trace melting. At any rate, congratulations to the PCB designers who managed this feat!