The issue isn't that there isn't a lot of space up there (as the name would imply, there's a lot of room). The issue is that nothing up there stays still: every

The issue isn't that there isn't a lot of space up there (as the name would imply, there's a lot of room). The issue is that nothing up there stays still: everything is orbiting the earth, moving at velocities in excess of 7 km/sec (sometimes well in excess at very high altitudes). Different space missions have different orbit requirements, so there is a wide range of inclinations, altitudes, eccentricities, and ascending nodes. Take a look: http://stuffin.space/ http://stuffin.space/. It's a very dynamic system with a dizzying amount of pieces in it. The issue isn't so much mega-constellations in and of themselves, since presumably with good operations automation and standardized communications between operators we can thrust and avoid potential collisions as they arise, even with very large constellations. The issue is reliability: applying a small burn to nudge out of the way isn't an issue when one or both of the two spacecraft in a potential collision is operating and has propellant (although accidents still happen [1]), but if they're both dead, that's a recipe for a hypervelocity trainwreck. That's why it's important to ensure reliable fail-safe de-orbiting of end-of-life spacecraft. The FCC is moving in this direction, but unfortunately the rulemaking is still pending. [1] https://en.wikipedia.org/wiki/2009_satellite_collision https://en.wikipedia.org/wiki/2009_satellite_collision