Hey everyone,
Ross here from a16z. First, we are excited to see the introduction of this proposal. For nearly all EVM chains, migrating underneath Ethereum offers strong upside incentives with little downside given the current technical progress of Layer 2 scaling infrastructure. We think this is the right big-picture decision for Celo and we’ve backed that belief by casting a substantial yes vote on-chain for Proposal 116. Celo now stands at the forefront of L2 decentralization. If carried out effectively, Celo’s L2 transition will shape the way all subsequent L2s think about decentralization and incentive alignment.
The migration to a Layer 2 (L2) design comes with a multitude of benefits. The security offered by anchoring state to Ethereum mainnet, coupled with connectivity to the broader Ethereum ecosystem, provides massive incentives for this migration. We anticipate that other EVM chains will follow this trend with similar proposals in the foreseeable future. Assuming proper implementation, the Optimistic Rollup (ORU) and off-chain Data Availability (DA) layer design choice will provide robust security guarantees while maintaining low transaction fees for end users.
While this proposal serves as an excellent introduction to the shift towards Ethereum, it is important to note that there is still a significant amount of research and development to be conducted before the transition can occur. During this period, the community could consider some foundational questions that all L2s are grappling with. These questions primarily revolve around token utility, revenue generation, and revenue allocation:
- Revenue Generation
- How will revenue be generated in the L2 stack? In this context, revenue is loosely defined as “economic value captured at the protocol level.” We see two primary potential sources as 1) fees derived from user transactions and 2) internalized MEV. Source #1 is an umbrella category covering any value capture that could broadly be defined as “the purchase of block space.” Nested within revenue source #1 is developers building L3s on top of Celo’s L2 and having the L2 provide DA, security, and interoperability. Are we missing anything here?
- Revenue Accrual
- Where does revenue accrue to in the L2 stack?
- Does the off-chain DA layer design complicate this question beyond that of a rollup with on-chain DA? i.e. does the DA layer add another potentially independent entity to the stack.
- Sequencer Design Choices
- How will Celo’s sequencer set integrate with the broader L2 ecosystem, given the plan to transition Celo’s current consensus protocol and validator set over?
- Will Celo lack shared sequencing with other major L2s and thus lack the desirable attributes imbued by shared sequencing?
- If so, are there any potential benefits to this?
- MEV Capture
- Who will capture MEV on L2s long-term and how prevalent will it be vs. L1?
- Can this MEV be captured and allocated back to Celo in accordance with its goals? Should it be?
- All decisions related to the distribution of revenue are inherently political decisions. Who precisely within the Celo ecosystem should MEV be allocated to? e.g. users, developers, DAOs, CELO holders, sequencers, foundations, etc.
- How should those allocation decisions be made?
- How different does MEV look for Celo if the rollup lacks shared sequencing? For off-chain vs. on-chain DA?
All of the above questions drive towards a core issue facing all L2s: The long-term incentive alignment of the L2 stack is unclear. While many teams and individuals in the sequencer space are making strides in addressing these questions, we believe that they remain largely unresolved at present. Wrestling with these questions now matters, because the implementation choices that different L2s make based on their answers to these questions could lead to materially different outcomes as more economic value accrues to fully decentralized L2s. Celo’s potential position as the first L2 to decentralize its sequencer set provides an opportunity to get these questions right and define a path forward for other L2s to follow.
If you have ideas on the above questions or want to chat individually on L2 incentive alignment, I’m always available in the forums here or on Twitter @rossshuel.