Again, don't make the perfect the enemy of the good. You are acting as though there are only two options: "anything goes" and "absolutely perfect ironclad balance." It is possible for there to be more effective balance; the fact that absolutely perfect ironclad balance is unachievable and even potentially undesirable has nothing whatsoever to do with whether some more balance is (a) achievable, and (b) desirable.
Unless and until you actually take seriously the argument that, yes, perfect absolute balance is not worth it but partial, better-made balance may be, you're not taking the rebuttal seriously. You're dismissing an easy strawman ("perfect balance is unachievable, and undesirable even if it could be achieved!") rather than the actual thing being said ("the balance we have could be improved to a worthwhile degree.")
Let me return your question with another: Are the pillars actually what they're presented as being?
We've been told, many times over the years since the D&D Next Playtest, that these are the three fundamental things the game is about. That while it is possible to construct an artificial situation where one or more of them is irrelevant, in practice no game can avoid interacting with them except by actively trying to do so. If that is the case--if these are the three things the game is designed to focus on, even if some groups choose not to interact with one portion or another--then yes, these contributions have to include out-of-combat utility.
It is unfair and unreasonable to provide class options that have no meaningful way to interact with the things the game is designed to focus on. This does not mean that every class needs diamond-perfection equality in everything. But it does mean that some meaningful, class-specific contributions should be found in every class. Again, it does not matter if some subset of players choose not to engage with things the game was specifically designed to do. What matters is that the developers have told us that that's what the game was designed to do explicitly, and have implied, over and over and over, that each of the classes is meant to have a fair shot at playing the game as designed. You cannot have a fair shot at playing the game as designed if you literally don't have tools to do so when others do. As stated, it is not unreasonable to consider the impact of things available to every character; pretending like such things don't exist at all would be foolish design. But pretending like such things are sufficient to balance the classes is not acceptable; the whole point is that any of those things that a Fighter (etc.; the class is emblematic) can allegedly do, a Wizard (etc.; again, emblematic) can do too, while the latter also has a whole bunch of other things.
Unless and until the books explicitly tell us, "The classes are not designed to be peers; some classes are simply better-equipped than others," I cannot and will not accept any argument of the form "but Fighters (etc.) can do [thing every character regardless of class can do]" or "but Fighters (etc.) can have/acquire [items or features that anyone can have/acquire, especially if said things depend on DM fiat.]"
(It's worth noting, here, that some things can look like these arguments without actually being them, e.g. "Fighters get more feats," which can allow them to have a feat--in the frustratingly-unlikely event that DMs actually allow feats--when someone else might not. The main problem is that, for the vast majority of the game, the Fighter is at most a single feat ahead of anyone else, and while feats in 5e are much better on average than they were in 3e, just one of them is nowhere near enough to bridge the gap....especially when those feats are competing with the almost always more powerful, and infinitely more boring, ASIs.)
If the game of D&D has been designed to be about combat, exploration, and socialization as the core, fundamental game experiences, then yes, either every class should have meaningful contributions to offer in all three, or the game should be open and honest about the fact that some classes are simply better at playing the game than others. I grow fatigued and weary with the song and dance. And since I am quite confident that neither WotC nor any company worthy of carrying the brand would ever contemplate straight-up telling Fighter fans, "Sorry Fighters, your class was designed to be less powerful than Wizards (etc.), if you want a more powerful class consider Paladin instead," because they would be quite keenly aware of the dramatic and deserved backlash such honesty would generate, I would be extremely surprised if they chose that option. That leaves only "maintain the charade of commensurability," or "actually make the classes the peers they've been implied to be for decades."
TL;DR: Stop pretending the classes are (more or less) "equal" if they aren't. Either make them actual peers, or admit that they aren't--and since admitting that would be brand suicide, it seems the only rational option is to make them peers.
Edit:
Or, if you like it structured in the form of a question...
Why is it that all classes do have to have combat components (e.g. cantrips/Sneak Attack/Extra Attack) but there is no analogous need for all classes to have support for non-combat utility functions? If all three pillars are meant to be about equally important from a design perspective, why is only one pillar something everyone gets a guaranteed provision of class-specific competence, while the rest of the game is "eh, some classes get diddly-squat, others get phenomenal cosmic power"?
Why do Wizards get cantrips and sleep and fireball, all the combat options they could want unless they actively opt out? If it's supposed to be that some classes just don't interact with some pillars in any meaningful way (and no, I don't consider the hilariously-named Remarkable Athlete to interact with other pillars in a meaningful way), why is it that there are no pure-utility, zero-combat classes, ones that can't even choose to interact with combat even if they want to, other than through absolutely generic resources?