> Economic theory says that (assuming a competitive market) the selling price of the object should converge upon the incremental cost to produce the object.

This does not need to apply always. Does money have the value equal to production cost?

Very few things apply always. Obviously money doesn't, but money is not your garden-variety competitive market like carrots or coal. The only legal suppliers of money are the nation-states that issue it. If you tried to print "money" yourself, it'd most likely be worthless because either nobody would use it (ok, maybe some cryptocurrencies excepted, but for reasons that have more to do with anonymity and a desire to stash illicit savings with a potentially lower probability of forfeiture than anything else), or you'd get in trouble for counterfeiting.

My point is that this is not a hard rule, and using it to prove something about PoW is on shaky ground.

Ditto land and art.

You can't produce more land, so there is no marginal cost - and art is literally not competitive - what makes art "art" is that it's unique.

... or, more commonly, that art is protected by copyright.

What you said is not true in case of PoW, and we know it experimentally already - mining costs follow the price of bitcoin, not the other way around. You can look at the charts yourself.

It makes no sense to try to mine a crypto token if the cost to produce one is greater than its value. Advances in mining technology or the increasing value of crypto can help keep it profitable, but there are diminishing returns to investment in technology, and (one would hope) the value of Bitcoin will eventually stabilize.