I agree with the point behind your question, but pDAO will never own 100% of RPL supply - that would require pDAO to be the only entity purchasing RPL, all existing NOs to have sold their entire RPL stack, and all sell-side demand on all CEXs/DEXs to have been exhausted.
But I do agree, there should be a limit to how much RPL pDAO owns. Too little supply on the market causes liquidity issues. If pDAO owns “too much” RPL, pDAO can propose to sell some on the market to increase supply.
This is one of the reasons we oppose UARS: At maturity RPL would be competitive with rETH. They both produce ETH yield so investors would pick the one with the best ROI. At maturity, the LST governance token competes with the LST itself. This does not seem like a desirable outcome or sensible product strategy.
Under our proposal the reason for owning RPL at maturity would be continuing RPL yield. This is a much clearer value proposition, and means our governance token is not directly competing with our product.
Over time the system eventually reaches equilibrium (insofar as anything actually reaches and maintains equilibrium in crypto markets). pDAO will only purchase enough RPL to ensure the node is collateralised. As soon as the minimum level is reached the automatic purchases stop.
Sure. I’m copying from a response I made to Jasper previously, please lmk if you have more questions.
The RPL requirement protects against centralised takeover by making it prohibitively expensive to become the majority of the validator set. If a centralised entity, or group of, control too large a share of the validator set they are able to make viable threats to the protocol or have outsized impact on it (“nice TVL you got there, shame if a huge number of NOs exited simultaneously”).
Let’s imagine the RPL requirement did not exist today. NOs can earn 14% commission on borrowed ETH and there is no need for additional collateral on top of the ETH bond. How might a centralised entity take advantage of that?
One option would be to empty the RP deposit pool using Lido ETH and have centralised NOs run smartnodes. Thanks to the 14% commission Lido could pay NOs their usual rate and the DAO would profit at the expense of rETH holders. Or they could split the extra profit so everybody wins - except rETH holders and those who care about decentralisation.
The RPL requirement is the only thing that prevented that scenario from occurring in the past, and the rework removes it without providing any replacement protection. RPIP-59 has some mechanisms that would slow it down but these could be avoided with sockpuppetry.
Ofc under the rework things would be different due to changes in commission, but without RPL collateral the attack vector remains. This attack would be viable if no_share is high enough to make it profitable.
FWIW, this issue came up before in NodeSet’s review of the tokenomics and was never addressed by the proposal authors. Some seem to think that because ETH-only NOs do not have voting power in pDAO there is no risk, but as Wander explained this is incorrect:
The centralization concern isn’t for governance, it’s for key-man risk and consequently rETH safety. If Coinbase has 90% of validators, then rETH is essentially cbETH with some extra complexity (risk). Because RP is permissionless, there’s no way to prevent this outcome unless the portion of ETH-only minipools is limited (Lido is going with 10% for these reasons).
There is nothing in the rework proposal that protects against this outcome. The only solution offered by Samus and Val was:
Obviously this would leave to a mass exodus of NOs - home stakers included - and is not a viable solution to the problem. But it is the only one offered by the rework proposal.