For reference:

" The Australian BHP Iron Ore is the longest train ever recorded in history at approximately 4.6 miles (7.353 km). In the Pilbara region of Western Australia, BHP owns and runs the Mount Newman railway. This is a private rail network designed to transport iron ore "

In the U.S. they often double stack containers--so that would be half the distance if they could:

"Invented in the United States in 1984, it is now being used for nearly seventy percent of United States intermodal shipments. Using double stack technology, a freight train of a given length can carry roughly twice as many containers, sharply reducing transport costs per container."

https://en.wikipedia.org/wiki/Double-stack_rail_transport

But generally going around the southern tip of South America could well be cheaper--although it may take longer.

You're right about that. However, this probably only works with diesel-powered locomotives and not with electrified railway systems.

why wouldn't it apply to electrified railways? I mean it's all energy in the end. I'm sure they make electric motors powerful enough and you can pull your own batteries as part of the load.

The clearance above the train. You have to rebuild every passenger transportation train to work with higher clearance or you build a dedicated railway system only for transportation. Neither financial nor ecologically sustainable.

I mean it would seem to be an innovation but "invention" seems like a stretch as there is nothing new about stacking stuff. I mean it should be a staightforward idea/extrapolation for any "good" mechanical engineer familiar with railroads and trains?

I wander if there are any publically available topographical maps along the roughly estimated region available to find an optimal path, even if it's a "as the crow flies" estimate.