Well, a lot of it depends on how you plan to run the game. I was toying with the idea of a (very large) hexagonal map consisting mostly of empty space to distance inhabitable solar systems and make transport tech actually worth something. The idea is to start the game with players somewhat isolated to system 'clusters', and then being able to explore deeper into space with technological progress (this partly assumes that like IU, space is actually explored).
I thought of splitting combat into two aspects: whether or not the battle is won, and (if using physical armies) casualty tallies. Instead of the usual utterly-annihilate-the-defenders, forces could retreat to friendly territory. It also opens up the possibility of players having some say in default RNG by choosing from a list of attack strategies: aggressive attacks would leave a higher bodycount, but carefully-planned attacks would give a slight bonus to combat odds, even if more of the enemy forces escape.
One element to IU I suspected would become debilitating in the long run was the five-system cap on stability. I devised regional magistrates in place of civics research; one could hire an unlimited number of them, but they would be extremely expensive and incur a persistent maintenance cost. Civics levels could then reduce their prices and upkeep.
Those are my three big ideas, at least, and might not be compatible with whatever you're striving towards, especially if you're adverse to finicky economics details.
.
