Well, as i wrote already - this video is just to show the routine moves for economic development. Its not some record run or something. Its not about game VS AI at all, you dont need such performance to win AI.
In vanilla moo2 there is only one demo race, what in theory can compete with good uni races - demolith. If demolith will get a good sector around, and uni will get bad one - it could win then. All others will lose even in this case. If you really like the way techraces works - you can try VDC mod, or some other ones. They all are trying to fix Uni's blatant imbalance, by making demo cheaper than uni as a pick, and by actually enabling morale in game (vanilla game have no viable moral techs, so, while Uni should eventually lose a pace due to demo's or dict morale boost later, it doesnt happen in vanilla due to awful morale tech placements, they are simply hoplessly too late). In mods techraces are very good in result of fix.
With maps - original map generator so random, so actually the most stable strategy involve skipping colonisation at all (returning to AI now). You can win many ways, but if you are planning to win from one planet - it means what it doesnt matter if you got small tundra or huge gaia. And the fastest and most reliable strategy VS AI doesnt involve any colonization (colonybases included), not labs-autofacts, theyll just slowing you down. Yes, AI is so bad, so if you will build research and build autofactories - you will just give him few more turns of life by doing it.
Depends what you mean by "good" here. All 7 AI's always could be killed pre-T100 from prewarp on huge with right race, no matter of map, even if you turn wormholes off. With wormholes - its about them, artifacts, splinters and leaders, average time is around T80 then, my best is T63 from outstandingly lucky map.
4-5 systems is not a real data without population, if you talk about economic development, not AI bashing. Uni will have about 10 systems and 100 pop around this time (if map will afford), and can finish all the techfields before T200, settling a whole galaxy by that time. Sadly i didnt attacked Antares for years, so not remember about it clearly, but if you have an option about taking Orion as viable - it could be done quite early, using Orion tech, if not - slightly later. So yes, if i got you right its not too good, but more than enough VS AI's for sure.
I think trading with AI is a cheap nonsence, its of use only when you're learning the game and want to see whats inside it. Later its just cheap and unfair, you just got tech for free this way. Better to learn to develop without it imho.
Uni's techrate at T100 is about 300-400 from prewarp, at t 150, dunno, few thousands, game is rarely go that far.