What's the cost per kg to get a satellite into space and how do you solve that problem, knowing there's no financial ROI ?
How do you solve this problem? You work on the difference in engineering cost between $21/lb and $5k/lb, say by simplifying systems, building them more robustly so the $30k paper trail that guarantees that the otherwise $700 part will perform exactly as advertised goes away, or aiming for reusability.
Of course, if you did that on a large enough scale, you'd be Elon Musk.