Thanks @slatestarcodex, this is one of our (or at least my) key reasons for writing the scenario.
One push: the concentration risk you flag isn't only European. A world where one US government controls every frontier lab should worry Americans too as it's a checks-and-balances problem. Distributed power across democracies seems a better post-AGI future (though I agree it could make coordination on safety and slow down harder in some scenarios).
I agree the latter is 100x less important than the first, but I don't think it's zero important. The more dominant and powerful AI is, the more that a country should care about having it. I think the people behind https://europe2031.ai/ are Europeans who have noticed that if we get an RSI ASI singleton, their whole continent will become irrelevant. Maybe in theory everyone should be so united under the banner of all humanity that this shouldn't matter, but I think it's unrealistic to expect European governments to ignore their own potential obsolescence.
This matters less to me as an American, but I still think there's this question of - sure, spend 99% of our effort making sure we don't die, but if we don't die then the world where only one entity controls superintelligent AI - whether that's one lab, or a US government still in control of all frontier labs - has new opportunities for dictatorship. I'd be less worried about that if Mistral was still in the running (although correspondingly more worried about other things, like whether we could build an international coalition to slow down AI in time).