Thanks Arbitrum for the clarity. We’ll be going ahead with our vote FOR this proposal.
-
We are voting FOR this proposal. As Entropy pointed out, timeboost requires custom tooling to participate in ahead-of-time auctions which raises the overhead costs and barriers to entry to participate. This has meant that (i) concentrated the participation in timeboost to the point that 3 entities with larger balance sheets account for 97% of all auction wins, and (ii) the majority of non-CEX/DEX arbitrage txs bypass timeboost. We think that moving to a per-transaction priority-fee market should make priority access more competitive and easier for existing searchers, and new markets like propAMMs to integrate with.
-
In terms of concerns with frontrunning/MEV protection, as Offchain Labs pointed out, Arbitrum One’s sequencer continues to use a private mempool so this does not itself create a new public-mempool sandwich or frontrunning surface.
1 Like
Cornell Blockchain continues to support this proposal consolidated with the Fast Feed AIP in the single Constitutional on-chain vote.
Fast Feed Rational: [Constitutional] AIP: Transition Arbitrum One ordering policy to Priority Gas Auctions (PGA) - #14 by cornellbc.eth
I supported the Fast Feed at temperature check and I’m carrying that position into the on-chain phase. On PGA, what persuades me is the data on how Timeboost actually performed. It produced real revenue — ~$7.46M cumulative — but never became the default execution path: 87% of atomic arbitrage and ~98% of liquidation activity bypassed it, and three entities took ~97% of all express-lane wins. An ahead-of-time auction that required custom tooling concentrated participation rather than opening it up. Per-transaction priority fees are a more familiar mechanism with a lower barrier to entry, and competition is spread across every transaction rather than a periodic slot.
I won’t claim independent expertise on the MEV question, which is the one that matters most here. What I’ve leaned on is that ordering happens in the sequencer’s private mempool rather than a public one, and that L2BEAT’s research team and GMX’s governance committee both reviewed the design on exactly that point and were satisfied. The anti-starvation boost also looks like a sensible protection for users who never bid. Sunsetting Timeboost on Nova without adopting PGA there is consistent with the minimisation path the DAO already approved.
On Fast Feed my reasoning is unchanged from July: it’s a read layer on top of the sequencer after ordering is already resolved, with 97% of revenue
routed to the Treasury under the split we ratified for Timeboost and Trail of Bits audits on both contracts.
On a side note: Consolidating two separately temp-checked proposals into one atomic payload works here — both passed on their own merits — but I’d rather it not become the default.
I voted FOR this proposal.
I support the transition to PGA and the Fast Feed as a step toward a more open and competitive transaction ordering system, while also creating an additional revenue stream for the ArbitrumDAO.
I particularly appreciate that the DAO retains control over the proceeds and the ability to adjust the system as it evolves.
Looking forward to seeing how PGA performs in practice and how the additional economic activity it generates can contribute to Arbitrum’s long-term growth.
Voting FOR, my stance on both stays the same. PGA opens priority up to anyone, way healthier. Fast Feed is the DAO selling a product and earning revenue, thats the direction I want Arbitrum going.
Voting FOR. Same yes as both snapshot rounds… Timeboost was a good experiment that didnt get adoption, glad we can learn and pivot. And love seeing us actually make money with that 97% DAO split
We support the move from Timeboost to PGA and voted FOR the combined proposal.
We checked the Toggler contract and have one implementation question.
The two-year expiry is not self-enforcing: setCollectTips() checks that the contract has been activated, but does not check whether expiryTimestamp has passed. So after year two the manager can still toggle tip collection until someone calls the permissionless revoke() and removes the Toggler as chain owner.
Is this behavior deliberate?
Separately, the 28 July update committed to periodic accounting of PGA proceeds, public dashboards, and a separate DAO distribution vote every six months. We expect those commitments to carry through once PGA is activated.
Merlyn Labs is voting FOR this proposal.
Anything that engineers believe will make Arbitrum a more competitive technology among other blockchain options is always a yes from us.
The following reflects the views of L2BEAT’s governance team, composed of @krst and @Manugotsuka, and is based on their combined research, fact-checking, and discussion.
We voted FOR.
We supported both components during their respective offchain votes (PGA & Fast feed), and we continue to support them now that they have been combined into a single onchain proposal.
Our view remains that this direction can make Arbitrum’s transaction ordering market more accessible and easier to participate in, while still allowing the DAO to capture value from priority demand. The anti-starvation mechanism is also important, as a more competitive priority market should not come at the expense of regular users.
We also think the Fast Feed can turn economically valuable ordered transaction data into a transparent revenue product for the DAO, without changing transaction ordering, inclusion guarantees, or the fees paid by regular users.
Before the offchain votes, we asked our Research Team to review both proposals. They did not identify material issues with the PGA transition, and their view on the Fast Feed was that it may create some MEV opportunities, but not the harmful frontrunning or sandwiching behavior that would be more concerning.
For these reasons, we are reaffirming our support.