100km. The article you linked used miles.

That one transfers the load between different sections of its length. The longest conveyer that does not transfers the load is 31km at Boddington Bauxite Mine in Australia.

> This is five times longer [...]

Does that actually introduce any new technical difficulties?

Suppose I needed a 2m x 1m table. I could get some 2m long boards and put enough in parallel to get my 1m width. I could put 1m boards perpendicular to those underneath at the ends and nail the 2m long boards to them. Then add legs at the 4 corners and I have a table.

Then I need a 5m x 1m table. I could take the same approach except using 5m boards instead of 2m boards. I might need to nail some more perpendiculars in the middle part of the table or maybe add some legs in the middle.

As I need longer and longer tables that approach keeps working up to maybe 10m. I don't think my local source of boards has boards longer than that.

So if I need a 30m x 1m table I need a new approach. I need to figure out something that doesn't require single boards that span the whole thing. It has to be limited to boards I can reasonably get. Let's say I figure out how to make that 30m x 1m table using no boards longer than 4m.

At that point I expect that I'd be able to build tables of any greater length by just doing whatever I did in the middle of the 30m table but doing more of that.

I'd expect that a similar thing happens when you build a 100km long conveyor. The specific project was "build a 100km long conveyor" but to solve it you figure out the solution to "build an arbitrarily long conveyor" and then apply that to "build a 100km long conveyor".