> non-experimental depot could use electric propulsion to bring itself to satellites that need to be refueled. It's energetically expensive but likely still worth it to extend the life of expensive satellites.

I'd have to do the math, but I'd be curious about the confluence of launch market dynamics that would make pre-launched depots which need to change orbital planes cheaper than depots launched on demand into the right orbit from the start.

I don’t know how the math turns out, either, but because this thing exists, I would guess it’s so much cheaper to launch one huge depot than a hundred smaller ones (reasons could be higher cargo/total weight ratio and/or lower launch overhead) that it compensates for the extra maneuvering.

Of course, that’s gambling on those hundred customers to exist, on refueling to be reliable, etc.