I'm kinda thinking like this. Focus each one (little known fact, the focus is adjustable with a screwdriver) for the minimum dot size at 50 ft. I turn on one reference pointer and lay it and the next one on a flat surface, then apply UV sensitive epoxy. Tweak the position to make a perfect convergence at 50 ft and apply UV causing the epoxy to freeze in position. That stuff hardens in about 1 min of strong UV. Usually epoxies do not shrink during their cure and is inflexible so it should hold the alignment as well as possible. Then turn off that one but leave the reference beam on. Align the next one with the reference and repeat.
Make rows of 7, then perform a similar procedure to line up the row assemblies with the reference beam in the primary row.
I am not sure how these things are driven. I need to take one apart. It looks like the driver may have a photodiode inside the unit for feedback, I remember they were more complicated than a resistor ballast. And it's unlikely they built a circuit in there for something that is not indispensible. How they make these things for <$0.50 is beyond me.
And I am almost certain the diode body is the cathode and it's tied to the case. This sucks because that even if I wanted to bypass the driver, the case being ground would prevent being able to put them in series with the mouting scheme I just described. So at this point it looks like I need to bypass the switch and provide 4.5v to all the pointers, so now we're talking 10 amps at 4.5v total, which is harder to build a driver for and less efficient.
Now most of the laser pointers I've had are crap, and I've taken a few apart. One had some kind of metallized plastic making the connection between the laser diode case and the brass body, this would suck. I can't remember exactly what it was like but I remember that I decided there was no reasonable way to fix it. Others designs seemed to only have trouble with the button, which I won't be using.
I'm all over this idea now. It's just such a spectacular degree of stupidity, I can't imagine why it hasn't been done so far.