The issue with Tokyo is that the density supports metro expansions.

Per distance costs should be theoretically comparable? (Regulation costs and labor coming to mind as drivers of cost differences)

I was talking about the demand side / use more than the supply side / cost. But after a cursory examination of the statistics, Tokyo doesn't seem to have a measurably greater density by any common metric vs other cities.

So I suppose my above was wrong. I'd guess the more reliable enablers are only sufficient density, coupled with good geology and a cooperating regulatory / permitting structure.