Thank you @middleway.eth and @mfw78 for your comments!
@middleway.eth : you are fully correct! I was a bit sloppy in describing the proposal, but in practice it will work exactly the way you describe (and it is indeed DAI not USDC)
@mfw78 let me give you a bit of context (which was lacking in my initial message). We see this integration as a first step toward a potentially deep collaboration between us and Balancer. For example, see the result of this research grant (there is some additional internal research on how you could integrate a batch with an AMM that we will release soon). The Balancer team moved extremely rapidly precisely because they understood this as a first step to a longer journey.
You are correct that competition between solvers should be enough to eventually start using the Balancer’s discount. The problem on our end is speed: we would like them to start doing it quickly to move rapidly to the next collaboration stage. This is also why we did not calculate this reward based on costs: the rewards are not meant to provide the incentives (as you said, there is the competition for that); they are meant to be a nudge to do that quickly. Finally, we did not base our proposal on an estimated order flow to those specific pools.