There's been a thread about the desirability of the common state in 4X games where once a faction gets big enough, no one can stand against them since they can simply field too large an army for anyone to survive against, in addition to other benefits like massive research fund advantage and so on.
Personally, the "steamroller effect", where once a player gets big enough the game is practically won (typically in the middle phases of a game) even though a boring grind to take the rest of the map remains, though sometimes it manifests as an AI opponent achieving such a dominant position and swamping the player under waves upon waves of armies, frequently aided by a higher unit-for-unit power level. I'm not sure what the developers' stance on the steamroller effect is, but assuming they want to combat it to some degree, I've decided to write up a list of game mechanics designed to at least weaken its hold over the game.
1. Have armies in enemy territory move slower (can't take full advantage of the infrastructure for logistical support), which helps the defenders in organising a defence in time. Somewhat relatedly, impose a larger upkeep cost on armies not in friendly cities and more so even in hostile territory, simulating the difficulty of long supply lines into enemy lands. Both of these methods would allow military adventurism just fine, but would make them less of a no-brainer.
2. Give armies defending their cities (a.k.a. homes) some bonuses to reflect the fact that they're fighting for something with personal importance to them. Again, wouldn't stop a determinate attacker but would help to give smaller defenders a chance against larger aggressors.
3. Somewhat connected to the first one, and possibly laborious to implement and, if handled badly, boring to play with: create a supply line from armies in hostile territory to home territory and allow that supply line to be raided (and defended!) just like trade lanes. Certainly a fairly involved option but might make for either interesting strategic choices if done right or boring micro-management if done wrong.
4. Have research efficiency depend on resource expenditure percentage rather than absolute terms. That way, barring bonuses to research, a large kingdom with 40% income spent in research and a small kingdom with 40% income spent in research would be on an even footing. Or maybe have some sort of compromise, where the capital gets a massive percentage bonus to any research there but the rest of the cities owned contribute normally, which would still give a large kingdom an edge but not as large a one as in a traditional linear model. These sort of mechanics would help combat the research side of the steamroller issue.
5. Stress the difficulty of holding a city conquered by force. The possibilities for this include culture difference penalties to revolt chance, hero units better able to either stave off revolt or convert culture, creation of partisan units around a conquered city and so on. This wouldn't prevent steamrolling, but would slow it down and make it less of a steady onrushing wave and more a matter of series of expansion waves, giving the defender time to react.
6. Reasonable diplomatic AI. When an AI player is clearly losing a war, they should be intelligent enough to cave in to demands of tribute in exchange of peace rather than the all-too-common state where they just get more and more determined to war as their lands are steadily taken away. This would also combat the tendency of 4X games to devolve into an eternal war starting from mid-game. Maybe provide this as an option at game start if a more traditional political AI is popular enough.
7. Related to number five, have special governor units (or hero units with a governor ability) that are better able to help hold down conquered populace. This could go a number of ways, depending on how the governor units are produced: if they're like in the Total War series where they're the offspring of a player's governor-generals, their rate of "production" would be fairly static, or they could be relatively expensive special units. They would simulate the difficulty in finding reliable and loyal people to give command of entire cities to, and would make it harder to conquer a great many cities quickly.
8. Merely giving a high economical weight to commerce might also help somewhat; if the price of warfare is losing a not-insignificant percentage of one's income as trade with the neighbour shuts down, it also makes war a less automatically best choice and more a matter of a strategic posture.
9. War weariness mechanics, where a long time of war starts to produce unhappiness in one's lands, though wars of aggression and wars of defence should certainly be distinguished from one another in this consideration, with the latter having far less effect.
10. Attacks having a (small) chance to create low-level hero units for the defender's side as a call-out to the common story element where a person flees from hostiles only to swear revenge and becomes a determined enemy of the attackers who killed his family/friends/village/etc.
Now, I realise I'm not in charge of the development process and don't know what kind of playstyle the devs want to promote, but I'd dearly like to see at least some of these mechanics (or their ilk) in the finished game. I admit, it's partly because I like to play a builder style so while some warfare is nice to have, I don't like it when it becomes both endless and the self-evidently optimal way to play. I'm also willing to discuss others' opinions on these of course as well as to see others' suggestions. I admit that having all of them in the game might produce a game that's too hostile towards an aggressive playstyle, of course, but I wanted to write up a fairly comprehensive set of possibilities.