Can an AWG 24 wire conduct 1000000000 Amperes?
This is where ChatGPT falls short. The question is not "if", but "for how long?"
Certainly the wire will vaporize, but in order to vaporize, it had to be carrying that current for some minimum time like a nanosecond before that.
Actually, I would dispute that this is possible at all: consider the dI/dt required to reach such a current even in such time scale where the wire
might possibly not immediately vaporize. Likely the voltage drop is sufficient to break down in air at the very least, and most likely cause field emission in vacuum too. The wire (and any insulation covering it) would quickly be consumed by the resulting plasma, and, I suppose, fully ionized in the process; the voltage drop might actually rise because no more free electrons are available. Eventually pair production would occur, stabilizing the voltage drop; or nuclear reactions.
For reference, the Z-machine does ca. 26MA, triggering nuclear reactions. The focus is, I think, a bit smaller than a 24AWG wire, but depending on length in question, perhaps it is comparable.
So, we can, with a little confidence, rule out currents quite this high, I think.
Continuing with the jocular tone, I say it’s not “for how long” but “if”. With gigaamperes we may need to treat a wire as a transmission line just to account for it evaporating before it starts to fully conduct.
Worse, not only would TLs be required to even begin to attempt such an experiment, nuclear physics gets involved, too.
...
As an exploding bridgewire, I think on the order of ~10kA and <1us, skin effect dominates, concentrating the power dissipation in the surface, and ~instant vaporization can occur; at lower dI/dt, probably this current is survivable, up to, oh I don't know, maybe 100s µs, 1 ms or so? Beyond which, (bulk) melting and vaporization occur rapidly.
These are fairly -- well, not
common, but plausible, let's say -- induced-lightning surge currents, within the realm of commercial or industrial practice. Somewhere around here might be considered the ultimate limit for such wire.
...
Anyway, yeah. For continuous or intermittent use, you can jam practically as much current as you want, until the wire heats up too much. The real question is, how hot is acceptable? How much voltage drop? Wiring tables assume some condition (limiting temperature at worst-case power dissipation capacity, e.g. wires in conduit in insulated wall space), so tend to be highly conservative. Chassis wiring is typically short and has good air circulation, so that more current can be passed for the same temp rise -- note that short wiring length, besides keeping the voltage drop down, also can provide some heat sinking out the ends of the wire, heat flow is not strictly radial (out of the wire surface) but lengthwise as well and therefore different assumptions apply.
Which is similarly relevant for PCB trace width calculations: it's practically irrelevant if the length is short or the aspect ratio isn't monstrous, and mostly matters on long narrow routes where heat flow is lateral.
Connectors add a whole 'nother wrinkle, as they might be rated more-or-less what the wire is, but the contact point(s), pressure, plating, temperature range, temp cycling, and mechanical vibration, all play a role in their endurance. Failure typically occurs when contact resistance increases too far, and the pin burns up. This can take merely hundreds of mechanical cycles (for poor quality or mismatched plating), and may be exacerbated by current flow (heating, electromigration). High-current connectors typically have multiple contact points -- contacts designed with multiple leaf springs, usually with a bump projecting from the surface into the terminal/post to concentrate contact pressure, and this takes up both mechanical tolerances and helps ensure consistent contact under motion. That said, I've seen ratings from 2 to 5A for headers in the style of Molex KK 0.1" series -- the upper range I believe by Sullins, though I don't recall their terminals looking any different from Molex's.
Tim