As the comment below notes, even if you ignore capital expenditures, there are only 3 profitable trains in the world. It's operating costs that make it unprofitable. The increase in value on the land may cover your construction costs, but operating it is a recurring expense. You'd have to realize extreme gains on the land to fund that.
What if you use the land price increase at the other end of a (quite) fast train to finance building the commute link?