Last post, I talked about the idea of designing dungeons under constraints. I noted that designing dungeons using an in-world budget would be kind of a pain, but sleeping on it, ACKS does give us the tools to do that - there's the table of structural feature prices (shared with OSE), putting a 10x10 square of dungeon corridor at 500gp, and ACKS' heuristic that a character with n XP has earned about 80% of that from GP recovered, plus the section in Domains at War: Campaigns on using magic to assist construction projects. Though looking over the guidelines there, it seems like using magic for underground construction would be tricky - Transmute Rock to Mud carries a collapse risk, Move Earth only works on earth and not stone, and Wall of Stone is dispellable. Most of these spells only increase the construction rate of manual laborers, rather than contributing value directly. Honestly Disintegrate might be the choice here for carving out dungeon cubes - but if you have a wizard with Disintegrate, pricing dungeon corridors gets weird, because they're free in money but not free in time.
In any case, to conclude that tangent, magic construction techniques might not be that useful for dungeons, but we can still get a decent estimate on how much dungeon a wizard could build based on their lifetime earnings, at 80% of their earned XP in GP. Sadly this breaks down in ACKS specifically at high levels because of the domain XP threshold rule, where you can earn GP from your domains but don't get XP for it. Fortunately, wizards earn the vast majority of their domain XP from spending money on magic research projects, so for now I'm not going to worry about this.
Let's consider two edge cases: the newly-fledged 9th-level wizard, and the biggest, baggest archmage who ever archmaged, 14th level. 310k XP and 1.06 million XP, respectively, for lifetime earnings of 248 kGP and right around 800 kGP. Presumably some of that will probably have been spent on towers, libraries, workshops, henchmen, research projects, etc - 20-25% on a dungeon seems reasonable. Let's call it 50k GP dungeon budget for the newly-minted wizard and 200k for the archmage. What does that get you?
Well, at 500 gp per 10' cube, about 100 10' cubes for the wizard, assuming nothing else. A 30x30 room is 9 cubes, so that's about 11 such rooms assuming no hallways, maybe 10 if you leave some budget for hallways and doors and such. If 30% of dungeon rooms contain monsters, then we should expect about three monster rooms. Given typical % in lair chances, we might see one lair or we might not. Going down to 30x20 rooms as your standard size gets you 15-16 rooms instead with some slop for hallways, which is still about 5 monster rooms, one of which is a lair. So that is not a big dungeon.
And then in the archmage case, you're working with quadruple the budget, so something on the order of 40-60 rooms, with 13-20 monster rooms of which around 3-7 are lairs?
Huh. So I guess if you take "dungeons come from wizards" seriously, dungeons probably shouldn't be enormous, for any model of wizards where they're secretive rather than cooperative with other wizards. Which seems a bit obvious in retrospect, but it's interesting to see just how small is really reasonable.
At the very high end, where you have a max-XP archmage who has spent 75% of his lifetime earnings on dungeon-building rather than 25%, you triple that again, up to 120-180 rooms, 40-60 monsters rooms including 10-20 lairs.
Looking back over price lists, one amusing consequence I could see coming out of this is in stairwells - a 10' wide flight of wooden stairs costs 60gp, while a 10' wide flight of stone stairs costs 180gp. So a wizard cheaping out on stairs might use wood instead of stone, and that creates some amusing potential interactions with fire (and certain jelly-type monsters, I suppose). Hey, you're a pro, you can Levitate, right? Stairs are for chumps.