I've always found that making up new letters is a fine direction to take, if you put in serious effort to each letter.This sort of proliferation of class options has an upside and a downside. One of the trends I've seen toward an edition's lifecycle is that there's a sort of "Design Rules Funnel" where the bandwidth that the designers have within which to introduce genuinely NEW ideas/rules into classes gets narrower and narrower.
The rules framework laid out for 5e's classes is already pretty narrow in the grand scheme. You can see that in the design language for the few non-combat-focused abilities.
To use a really bad analogy (but it least it's a clear one), if we commit to a design principle of "each power is one letter of the English alphabet", eventually we run out of letters.
But also, we don't have to talk in maximilist terms. We don't need 30 classes; we could still cap at 12, or even a softer cap at 16-20 classes, with most classes being Subclass-based, a handful Sidekick, and a slightly bigger handful Menu. Furthermore, this design opens up homebrew and 3PP a lot in terms of what can be created, and allows the base system to be adopted on an even wider basis for a greater pool of supplementary material to draw on. IMO, from the standpoint of trying to continuously become the biggest and best game in the scene, it'd make sense for WotC to create this type of framework just to maximize their indirect influence via 3PP. But also selfishly, it'd help all of us who do 3PP too!
But even beyond that, 16-20 classes that use this supported framework and that theoretically being ALL YOU CAN USE still creates a more robust first party D&D that better fits the reality of players having different mechanical proficiencies as players. There's really no significant downsides worth talking about when it comes to a moderate expansion of this size, but there certainly are if we assume first party alone is responsible for expanding the game.