The more classes you add, the more spells you add, the more material you add, the more complex the game becomes. That's unavoidable. If WotC started banging out splats filled with crunch, we're back into 3e land where you have a character created using material from three or four different books, often times creating balance issues. All of that results in Mathfinder.
I'm very, very glad that they're refusing to go down that road.
This is not actually inherently correct, and 5e's basic design in fact vastly mitigates this. Why?
Because it expands in ways that are generally sandboxed from each other.
Multiclassing aside (it's both totally an optional rule and generally more trouble than it's worth), you get one class, ever. You get one subclass, ever (and even if you multiclass, the subclasses of each of your main classes are cordoned off - you can't multiclass Evoker/Bladesinger, and if you multiclass Bladesinger/Dragon Sorcerer your Dragon Sorcerer class abilities generally don't apply to your Bladesinger spells). So each new subclass is a whole new set of concepts, which largely don't interact with other subclasses or classes. And new subclasses are the primary way 5e expands.
Spells in 5e are easy to see interactions between - due to concentration, there generally
aren't interactions between spells, and what interactions there are usually need multiple spellcasters to set up.
Feats are the only area I see ever becoming really problematic - they've become problematic in both previous editions that have included them. They
started problematic in both previous editions that included them. But you'll notice we've gotten precisely zero new feats, and I expect new feats to be rare even if they kick up the rate of introducing new mechanics.
So in 5e, the math doesn't get more complicated as you expand the list of options. It simply opens up new concepts for people to use as characters.