If developers charge what it actually costs to produce one parking spot, nobody would buy parking.
If this is true, then parking demand is zero. I think you're swinging too widely.
I've sold many sites that had high demand for parking in which the builder did not change plans to add.
How were they going to add parking, if they had a finished design, costed, to a certain height and depth and approval for same? Clearly, they would have to resubmit to the City and add an additional underground level. This has a wide range of significant cost impacts.
First, it delays any potential construction while resubmission/variance and re-design take place.
Second, unless you're certain you will sell the additional parking created by a full additional floor, you're now going to have surplus spaces you can't sell, which means you have to recover that cost elsewhere.
Third is the absolute cost of 'simply' adding one underground level to a design.
Fourth is the accretive or additional cost that may occur when you go deeper, there are implications for instance if you go below the water table in terms of not only waterproofing the garage/foundation, but also de-watering costs during construction.
There may be additional structural costs as well, as the new level now has to support one more level that the previous basement level did.
Had one developer add $50,000 to the cost of parking pushing it to $125 k just so nobody would purchase any additional.
This is not why they raised the price, they raised the price because people were willing to pay it, and it drove revenue to the topline and profit to the bottom line.
Clearly, in your example, the developer mis-estimated demand. But not so much that it made economic sense to delay the project while another level of parking went through design, and approval and added construction costs and delayed sales.
Sales of units were clearly going well enough that a shortage of parking was not material to the success of the project, it merely made the parking more profitable.