To estimate the humongous number $N$ of satellites to be seeded in a cloud on area basis.. we can assume hexagonal packing of nodes in the geodesic dome of Buckminster Fuller's designation $np,$ the number of subdivisions of a spherical triangle side of the basis icosahedron.
![enter image description here](https://cdn.statically.io/img/i.sstatic.net/2VPJ0.png)
$y= 1$ is the blue satellite separation distance between the centers of equilateral triangles in a hexagonal close-packed triangular array of each side red length $a=\sqrt3 y $. With sphere radius $R= x/2+z = 8000/2+100 =4100 $ miles, and assuming uniform distribution valid,
$$ N= \frac{4\pi R^2}{\sqrt{3}a^2/4}=\frac{16 \pi}{3\sqrt3}(R/y)^2 \approx 9.674 (R/y)^2 \approx 1.626\; 10^8 $$
Square root of this is around $12,750$ satellites spread over so many miles occupying almost a hemisphere cloud shell.
But they won't stay put at one place even if injected at same escape velocity in different directions, rather they may swarm around a bright street lamp ( the Earth) like so many insects in low earth orbits. They may need to have omni-directional antenna receiver-transmitters to avoid collisions in a sort of Brownanian motion.. better to address a Physics website about what might possibly happen.