That's what I was talking about. But even making Franco go from seeing the beligerant axis as a threat to an opportunity would be too much to ask - remember the game has to constantly maintain 10's of parallel AIs
So? Computing power is hardly a limitation here. We're not playing chess against Kasparov, we're just solving primitive optimization problems. In fact, in my suggestion, Franco isn't even solving anything. He just compares his and his allies' forces to any possible opponents, and decides that it's time to execute a very simple behavior that was dormant before this. If the AI can't do that, it could hardly be expected to manage Army Groups?*
Besides, as you correctly pointed out, it does have to maintain 10's of parallel AI's. One for each country, but,
probably, also one for every formation on the map. That's actually 100's of AI's operating at the same time. The operational AI's, however, have a much more complex environment to analyze, compared to the diplomatic one that countries deal with - if computing power ever becomes a problem, it certainly won't be because of country AI's, but because of the proliferation of Corps, Army and other AI's.
*In other words, we'd be back where we started - a problem much greater than Franco ending up in the Axis. If Paradox's approach revolves around ordering formations around, which it certainly seems to, there have to be a lot of pretty complicated AI's.