I figured that the tech auction system would result in the price finding a natural level. If it's too expensive at $15B total invested per turn, people stop until it's at $5B(or whatever). That said, I don't think it's really worked like that - it seems to be more of a fluff investment for most people, not a game-optimal one. Thinking about it, I think I might have a better system:
Tech is divided into three fields - Miniaturization, Strengthening, and Advancement. Miniaturization is heavy on mech-sized weapons, but also includes things like mobile HPGs. Strengthening is big on armour and increased scale of construction. Advancement is blue-sky tech, or substantial jumps to existing tech that aren't really smaller or stronger. (Yes, these fields are a bit vague. That's intentional - it lets me keep them fairly balanced, and it means it's not totally obvious which tech is in which category, to discourage excessive targeting.)
Each player's research budget is divided up between the three fields. Each turn there will be a roll in each field for each player. If a tech is found in that field, your research budget there is reset and you get the tech. If it's not found, the research budget carries over to the next turn. Yes, this means you can discover up to three techs a turn(though I expect an average of 1-2 instead). I want to get to all the cool 2700s techs before we all get bored or too busy to continue, and that might help.
Each player researches independently. However, to allow for spread, the base cost of a tech is reduced by 10% per turn since discovery, plus 20% if a neighbour has it and a further 30% if you've salvaged it from them in battle. There's still no espionage (consider any spying efforts to be part of your research budget), but for really important techs, you might wind up launching an attack for the sole purpose of trying to blow up something that carries it.
I referred to the "base cost" of a tech, not the "cost". It's not a cost system per se - it remains a raffle system. If the base cost of a tech is $2,000m, that basically means that 2,000 raffle ballots saying "You lose!" are put into the draw. If you invest $1,000m into that field, that's 1,000 ballots with "You win!" on them, which means a 1/3 chance of winning. If you don't get it, then those 1000 ballots carry over, plus any additional that you invest on the next turn. If you invest another billion, you now have 2000 ballots, for a 50/50 chance of winning(or better, if your neighbour discovered it since then)
So, a few questions for the crowd:
1) Is this more interesting?
2) Am I worrying too much about obfuscation? I want to avoid "Oh, the next turn is improved armor, I'll snap that up for sure!", but I'm not sure secrecy is the right play there.
3) Would you feel like I was taking options away from you by essentially forcing you to buy research every turn?
4) Is the faster tech pace something you'd like to see?