Also, deliveries to Earth really aren't the point. You're building this to accelerate orbital manufacturing. It's a lot easier to build a 3000-ton space station when the steel is produced outside of Earth's gravity well in the first place, rather than trying to lift that mass from ground zero to orbit. The rare metals like platinum may be dropped in part to Earth, but that's a secondary factor to the main thrust of the operation. It's how you sell the concept to investors, but it's not why you're
doing it.
The one neat thing is that scaling operations up is
much faster
and cheaper. Once you've got a system to mine ore, refine it to metals, and shape metals into parts, you can build more equipment in situ. I'm presuming metal 3D-printing equipment at least to start (you wouldn't have standardized mass-production in the early days, you'd produce what was needed in the moment), but that's pretty flexible already. Even if it doesn't work in vacuum, sealing the process up in an atmosphere isn't that hard, especially not if you don't need that atmo to be breathable. You can extend that to creating local habitation for on-site staff, if remote operations aren't feasible in the long term, too; about all you'd need to transport in might be atmosphere, water, and foodstuffs, but you could set things up to have a hydroponic farm that maintains all that once it's in place. Energy wouldn't be a real issue since you'll have ready access to solar, even if you
don't decide to build something nuclear on-site.