The problem with having so many ships isn't just in the specific battles in local space, but more to do with how they exist on all the planets at once. We're simulating the whole galaxy in realtime, which I don't think is really required. Whatever you are interacting with directly needs to be simulated in realtime, sure, but more shortcutted sim logic should be possible on planets that aren't involving players.
We already do some of that anyway (using coarser simulations on planets where there are no scouts, etc), but things like the players putting scouts everywhere and keeping them there can still muck with things.
Shifting something like that to multiple cores only really works if the clashing results that would give don't matter or can be reconciled in some way. And if that is either magically deterministic somehow (which is hard, but doable I suppose), or transmitted from the host to the other players. For instance, the AI in SBR is really fast because we run it all simultaneously on all the various threads that you can muster (if you have 16 cores, it will use them all), and then the results are all reconciled at once after that.
There are a lot of fundamental things that would have to change about AI War in order for that sort of thing to work, though. It would be a ground-up recoding, and there are certain gameplay tradeoffs that would have to be made, while at the same time other gameplay possibilities would open up.
I'm not a fan of just doing the same thing repeatedly, so if we took on a project like that, I'd want to go for something where we can hit a high framerate and more sophisticated ship physics (ala TLF) without things just going nuts from being too slow. And then from there see what can be done with scale, and how AI-War-like it can be made to feel. It's part of why I wouldn't really want to call it a sequel, potentially.
The other thing is that, past a certain point, the number of ships becomes kind of redundant. We wind up having to combine the icons in far zoom so you can even see what the icons are. When you are zoomed in, the battlefield is so wide that there's no point in zooming in, so you zoom out to the point that all you ever see are the icons.
On the one hand we could go to a smaller set of ships per battle so that we don't need the far zoom icons, but instead you're seeing things kind of like zoomin out on TLF. But with a more spread-out battlefield. That has its downsides.
On the other, we could just cut out the zoomed-in ship art and go to an all-icon approach (why try to do nice ship art if nobody actually plays while seeing it? Resources could be better spent elsewhere, like making those icons and their environment look as awesome as possible). Even if we did that, trying to change things so that so much icon combining is not needed would be really nice. Then again, how many icons need to be combined is largely a matter of how zoomed out you are, so there is that.
Anyway, the main point is that there are a lot of factors. I'd be interested in doing another RTS someday, and certainly I'd want to take a lot of the best lessons from AI War. But I'd want to try to solve some of the various problems it has in the interface and performance department that are caused by sheer scale of certain parts of the game.
THAT said, if certain pieces of scale were solved, then other parts could be ramped up. For instance, if the planets weren't all simulated all the time, then we could have a lot more of them. The galaxy map could be something that you pan, zoom, etc, on.
It's something I've been turning over in my head with various permutations for half a decade now, really.
I don't have anything I'm super excited about doing with it yet, so there is that. But there are a variety of things that I could see becoming excited about doing if I have another flash of insight related to one of those things.
Plus, the longer I wait the better computer hardware becomes, and the less of an issue the performance is, heh.