just don't make things too complicated the first time around. There can always be a third CW if we don't think the second is complex enough. But if the second CW is too complex, we'll never get it done. I don't like modded weapons personally, since enforcing it will be a logistical nightmare.
This isn't a 2nd Cold War. It's the 1st Galactic Conquest.
Siber has a strong point. We cannot easily design a weapon customization system that's balanced without taking a LOT of things into account and possibly making it more effort than it's worth.
Mjolnir; I had considered a tech 'theft' or salvage mechanic that would allow the victor of a battle to recover enemy tech from each destroyed ship. Each ship destroyed could yield a % of salvageable tech (dependant on how many of that weapon/module the ship had; the more it had, the greater the chance of salvage) which would work like a research topic; but once you reached 100% salvage, then it would be unlocked as a research topic to be properly researched and completed.
So for an example, you'd have a much better chance getting a Beamer by destroying a Cronus than getting a Gat Blaster by destroying a Hestia.
Each salvaged part might yield x% towards the total (dependant on the part - some would be easier to reverse engineer than others) so salvage could be a fairly slow-moving mechanic.
I also think you've listed way too many "pools", to the point of maybe one weapon or module per pool. As it is, we have maybe two dozen weapons and a dozen modules.
I suggest boiling it down to larger pools like;
Torpedoes & Missiles
Active Defenses (PD)
Passive Defenses (Deflecs, Aegis, etc.)
Exotic Defenses (Impeder, etc.)
Deployment Options (allows Deployers w/ respective tech areas)
Experimental Weapons (Frosch, Lancet)
This list is nearly as long, but it combines some of those areas fairly well. If we up the "take" count to 4 just for starting tech, and allow people to capture tech from enemies, or acquire it through certain sectors, or just research more of it, it's far more focused than your list without packing too much into each pool or -under- packing pools to make them useless.
We can still have pre-requisites for pools. For example, if you want to start with point beams, you need Beam Weapons & Active Defenses. If you want a Platform Deployer, you need Blaster Weapons & Deployment Options, and so on and so forth.
As it is, repair mechanics will not be in place due to the overhead required for combat. This is final, and anyone arguing can go stuff their face up a drainpipe for all I care, because there's no way in hell we're going to model and keep track of damage a ship has taken. Ships will be automagically repaired assuming they survive combat, but any dead ships stay dead.
The combat system is still a bit shaky. I'm putting forward an AI vs AI system run by the 'staff' to ensure a balanced outcome - in most cases, the player can't really lose to the AI. Whilst AI vs AI isn't as "interesting", it's far more balanced and ensures that 1) nobody fudges the results in their favour, 2) we don't need to run tiebreaker fights and 3) we know the outcome will be fair & balanced, unless the staff member -running- it rigs it, in which case they'll be out of their position faster than you can say "that's not even funny".