I believe the proposal will be for the amount of $POKT that is needed to stake for each of the chains on which they are doing research in order to achieve the daily RPC throughput needed to conduct the research. While it would probably be structured as a grant since there is no funding category called “loan”, it could potentially be considered a loan - as in, the recipients can declare their intention to return the POKT to the DAO when they are finished doing the research, e.g., by the end of two years if staking via “pay as you go” model. Alternately , if RPC usage is only needed for a month or so, then a smaller grant can be made to cover RPC usage for just those couple months (in which case there would be no idea of returnin the tokens to the DAO)
If Arthur could enumerate the desired daily throughput per chain and how many chains an over what period of time, we can translate that to the right amount of POKT. I recommend that the DAO match the efficiency of effort which Arthur has brought this to the table by having GRIP get this proposal-ready within 48 hours and immediately to a vote thereafter, seeing that the vote itself needs to run 7 days. It is not a particularly complicated or controversial ask.