Maybe you can suck it and see. If the wire(s) prove to warm, either replace or strip insulation near the ends and solder a parallel wire.
But FYI, the following was my original ramble....
The TDA can do 4x85W = 340W into 2R (Ohm) speakers or 4x50W = 200W into 4R, so that's ~34A or 20A using the usual divide by 10 rule.
If you add a 65W NUC then add 6.5A.
Hence an expected max of ~40A else ~26A.
Unfortunately I never use ampacity charts other than f.ex Powerstream's AWG table for sanity checking. That table reckons 18AWG = 16A for chassis wiring or 2.3A for transmission (which I take to mean still-air suspended) so hence I'd suggest that 18AWG is too thin.
But if the length is short enough it may be ok - ie, fatter end conductors or connectors conduct enough heat from the cable. I'd say that PAC cable is too long, but a test and feel (for heat) might be worth a try.
I design by acceptable total voltage drop. EG - if a 2m run from the battery to the device and I allow a 1V drop, that's 0.5V for (say) 40A per meter.
V=IR hence R=V/I = 0.5V/40A = (max) 0.0125R per meter = 12.5R/km and hence 15G (10.44R/km) => 14G, or 16G (13.2R/km) if a bit more than 1V is acceptable.
Actually that's strange - normally my method never comes close to exceeding ampacity charts yet 14G & 16G are rated for 32A & 22A chassis wiring respectively so I reckon I've mad a mistake somewhere...
Coincidentally substituting a total 2V drop above @ 40A means 18G.
Maybe my error is that I usually deal with smaller voltage drops & bigger currents, and I allow for the GND wiring as well. (I usually only allow a 0.5V total drop for large loads like headlights and bigger...)
Of course tables are at a certain temperature and as temp increase so does resistance... And I've forgotten if those AWG tables are at 25C or 65C or 85C etc...