We know the upper bound for most of those numbers. SpaceX already achieves internal marginal launch costs of ~$1000/kg, for instance. We know their rough costs

We know the upper bound for most of those numbers. SpaceX already achieves internal marginal launch costs of ~$1000/kg, for instance. We know their rough costs per satellite. In contrast, we know little to nothing about the inputs to the Drake equation. The numbers don't quite work out in favor of orbital datacenters at the current values. But we can tell from analyses like this what has to change to get there.