GURPS Magic has a few paragraphs about the cost of magic items, which make a number of assumptions about magic, including a very dubious assertion that powerstones aren't used for quick and dirty enchanting. As GURPS Magic was never intended as an economic simulation, it's easy to see why this decision was made, but it's not very plausible.
A very simple counterexample will prove the falsehood of this assertion. Let's say two enchanters get together to create a powerstone production company. By working together on enchantments, they can perform N enchantments per month. Now, if our enchanters go out and buy 2 10-point powerstones, three times per month, rather than the two of them working together on a powerstone, they can each do a solo enchantment. This, production went up to N+3 enchantments per month, giving them an additional income of $60 per month or $720 per year. As those two powerstones cost a total of $3,800, that works out to a return on investment of 19% per year, which is a very nice rate of return.
Furthermore, we can do better. A one college powerstone, enchanted into cheap materials, is 48 energy per casting. That's small enough to be practical for quick and dirty enchanting, and will on average require 12 castings per powerstone, so each stone costs only $576. With two stones like that, the rate of return on a powerstone is 62.5% per year, which is enough that even TL 3 economics should notice the profit.
The cost of a cheap powerstone is simple: it's 100% labor. However, determining labor cost for stones is not all that easy. The obvious method of computing total labor for a successful stone and dividing by success chance produces excessively high numbers, because we don't continue actually working on stones that are destroyed or flawed past usability. The actual required attempts is equal to the sum of a Geometric Series, where a is the constant 216/206 and r is is equal to 1/(chance we continue work on a stone); this has a value that ranges between 216/212 (all quirked stones are kept) and 216/206 (no quirked stones are kept), with a reasonable central value of 216/209 (stones with minor quirks are kept). Using that central value, castings per stone are as follows:
Now, if we keep the $1 per energy for a singly enchanted item, we get the following costs:
One-College Stones
General Purpose Stones
Of course, what we really need to know is how much it costs to use the stone, which we will call rental cost, though it's possible enchanters will actually own their stones. Since stones mostly don't lose value, this depends on our expected usage rate and ROI. I'm using a rate of 0.03% of stone cost per energy, which works out to a ROI of 10.96% for a stone with 100% usage rate, and then adding $0.2 for warehousing costs (efficient warehousing is still 16 square feet per stone; $73 per year is not an unrealistic value).
Now, we'd like a final cost that works out to $1 per energy up to 100 energy. Assuming mages optimize for the most money per mage, the optimal circle sizes and powerstones up to 100 energy as follows:
Those are pretty nice wages (Average income in 2 casts per day); 4 per day to give everyone Comfortable wealth, 5 per day gives the circle leader +25% per circle member (which makes him Wealthy on a 6 man circle), 6 per day hires everyone a Struggling assistant. On the other hand, given the difficulty of qualifying for enchantment, I don't have a problem with those wages.
Now, larger enchantments start requiring more time per mage, which really should boost wages, and eventually more expensive stones. Up to 300 energy, it seems fair to just multiply wages by time in hours, which gives us the following:
Finally, for truly large items, we have the problem that mages can only really cast once per day, and the total number of stones needed to saturate your mages is impractical, so we'll just set price to about 2x the rental cost, with a minimum of $100 per mage. This gives us:
Beyond this point, we're better off with paut; figure +$50 per energy (from increased wages; unlike powerstones, the lead time on paut is modest), so at 800 energy we are at $18,800. That's about the upper limit of Q&D.