Yes, agreed - missed connections don't cost Delta much other than CSAT. Which is why I have a bugaboo about it. I am pulling this number out of my rear end, but probably only 2-3% or so of flights (or about 10-15% of delayed flights) are going to be tagged as "Delta's fault" (maintenance, non-WX crew issues, etc) and so the vast majority of misconnects from delays are going to be at the passenger's expense, not Delta's.

Similarly, people who buy a 36 minute connecting ticket, and the inbound flight arrives exactly on time, and the passenger gets off the plane as quickly as possible, and gets to the gate as quickly as possible (brisk walk, no bathroom, etc), and still misconnects due to an unreasonably short MCT and early gate closure.. will be claimed not to be Delta's fault and therefore they have no obligation to do anything for the passenger.

Indeed, a strict reading of the contract of carriage would actually imply that Delta is not even obligated to rebook them on another flight because Delta has fulfilled its obligations and it is the passenger who "technically" failed to meet them. Obviously they will, but they will almost certainly not receive accommodations for overnight stays or other compensation even though Delta knew at the time they sold the ticket that the passenger would likely not make the connection.

Excessively short MCTs are arguably a form of bait and switch: selling tickets to passengers that Delta may reasonably know are not likely to be flyable by a passenger.

edited to add: when costs are switched from passengers to the airlines, there is evidence that airlines change MCT to be more realistic. MCT has gone up at multiple airports in Europe (I know for sure at MUC, LHR, AMS) since EC261 was implemented. While not conclusive, this is strongly indicative that excessively short MCTs may be exploitative in some cases.