Caps suck. Always. No to this.
Comments like this want to make me punch babies.
Rant
Are you aware that almost every single commercially successful game includes caps in their games? Warcraft III has a limit to the number of troops you can command, a limit to the level of heroes, a limit to the number of heroes and a limit to the number of items they can use. Was this game successful? Did these caps help to balance the game? Hell yeah to both of these.
Another example is world of warcraft. There is an armor cap, there is a cap on the rank of skills you can obtain for each level, there is a level cap, there is a cap to the number of people you can bring into any given instance. Do all of these things contribute to the game being more balanced an fun? Of course they do! Even games that you might not normally consider to have caps, still have them.
Think about civilization IV: Your cities have a limit to the number of tiles they can work at a time and consequently have a maximum population. Cities have a maximum number of trade routes, and units have a maximum strength. Can you mod these things out? Sure you can, and that was what made the game awesome, but the caps in the original didn't make it any less fun. People probably never even thought twice about worked tiles as a limit, but it was one of the biggest contributing factors to properly pacing the game.
/End Rant
Most of your other comments are really odd: Different mana conversions? What?
Obviously both the high evocation bonuses and the high shard bonuses are contributing factors to why magic spells can get out of hand. This is largely in response to death shards being convertable and reducing the random chance of getting fuxxored on spells because of a poor start.
Unless increasing the mana required to cast spells affected the number of spells you could cast per combat or increased the activation time of those spells it wouldn't contribute to balancing the function of each 'action'. If it did, that might work out okay. It is much easier to strike a balance by making a spell do a static ammount of damage and simply allowing more of those actions.
Shard change may be a good idea, but should be considered carefully. It should be hard to do and have limitations, otherwise the risk of players having all shards of one type would be very high and unbalance the game.
Part of the problem is that the ability to do this already exists. If you play as Ceresa you will have an idea of how rediculous this can get.
For those of you that don't want to read my rant, or that still have adolescent fear from the words 'nerf' and 'cap'. Let me explain that some things lend themselves better to diminishing returns (soft caps) than others. The level of your character for example. Let's say you capped their levels at 15 in order to strike a balance with regular units. If that unit could still gain xp and required lets say 8 times as much to reach the next level, would it break the game? Probably not. Especially since xp will now be shared and having more champions at level 15 is far more effective than your one champion that might reach level 20+
However a soft cap (diminishing returns) might not work so well on shard bonus to damage because there is little reason to do any other action. That is to say if you could increase your blizzard damage by 20% more but you never used any of the other spells besides blizzard, it might get out of hand. Additionally having soft caps on both of these things might break the game.
That isn't to say I'm entirely against soft caps with any of those elements in the orginal post, just that there are consequences of letting values have no absolute maximum. I like the idea of a prismatic shard being at the start location of the soveriegn! I'll update the post with it as well as another idea that came to me.