I can verify for a fact that Locus received bug reports from all over the world. There was a team in LA set up to handled them.  True, the customers were reported to be more academic than commercial, but there were commercial customers..  For instance, I was told that my current employer ran the simulation for then proposed/in-development 80386 on a 370-PS/2 cluster under iTCF  FWIW: I was under the impression they used TCF to migrate the very long running simulations from two different processors (I don't know that for sure, but I have been told that by two different people, so I'm comfortable repeating it).

I was also told, for an IBM customer to try to get the product was quite difficult.  IBM sales basically tried really hard to not provide it and you had to know about it and know you to ask to get it.  Charlie may know more, but he was working on AIX/RS-6000 which as has been pointed out was a separate code base.

As for finding an old copy of AIX/370 or AIX/386 (which are the same except for specific code that cared), I wish you good luck.  IBM kept the IP locked up.   We were firewalled at Locus and the code was only on a couple of very special systems and access was tightly controlled.   I was not on that team, so I never saw their raw IP ( so I would not be contaminated as I was one of the TNC architects).  But I was allowed to talk with Bruce and Greg who were the TCF architects.   We did all talk about common issues; but Greg and Bruce stayed out the TNC IP (until after the IBM contract ended) and myself and Roman stayed out of TCF.

Clem