Sovereignty necessitates choice
Sovereignty is a fundamental principle of the Cosmos design and its ideology. What this means in practice is developers have access to the full protocol design space and chains have control over their internal economy. Sovereignty ensures communities can exercise choice.
The beauty of Cosmos is that chains have the ability to design unique application experiences that would not be possible without the ability to customize every element of the stack. There is no “one-size-fits-all” solution that can serve all app-chains, rather each application must have the ability to choose its own path.
In our view, how a chain creates blocks is no different. We call this sovereign MEV. This post will describe why Skip and Osmosis believe the next generation of sovereign MEV will bring block construction on-chain.
Protocol/builder Separation (PBS) gives up sovereignty
Proposer/builder separation (PBS) was Ethereum’s solution for ensuring hundreds of thousands of commodity validators could simultaneously run nodes. However, this design is not suitable for Cosmos, which promotes full-stack custom application design and ecosystem choice.
Blockspace preferences such as toxic MEV prevention, payment splits, variable-length blockspace auctions, and custom fee markets are too granular for any generalized system to handle. Moreover, external, off-chain parties building blocks cannot be guaranteed to act in service of the chain long-term, or build blocks to exact protocol specification. This includes enshrined block APIs like PBS, or even Skip in its current form.
MEV has strong centralizing effects. If validators are given unequal access to MEV recapture, a small number of sophisticated players can bribe delegators for more stake. Both Skip and Flashbots initially focused on this problem by ensuring all validators participated equally in MEV. To further decentralize, Ethereum has moved to the PBS model, where validators can choose from a wide variety of builders. However, PBS still has multiple problems:
- Centralization of builders: as soon as block-building is outsourced out-of-protocol, the best builders are determined by access to private orderflow, and have a flywheel effect once they attain it to further entrench dominance.
- Unchanged power dynamics: the overarching goal of PBS is to pay validators as much money as possible. This is not conducive in the long run to using MEV to reward other parties, or to be invested back into the protocol.
- Loss of functionality: app-chain developers consider their mempool and block building a fundamental part of their design. By outsourcing building, they lose the right to have blocks built in the way that best suits their protocol, and off-chain parties cannot probably build exactly to their specification.
Thus far MEV solutions in Cosmos have mostly copied Ethereum’s prescribed path, starting with centralized auctions, but moving to Protocol/Builder-Separation (PBS) and a variety of order flow mechanisms. But it’s time for Cosmos to find its own path. Rather than replicating PBS, we believe we can do so much more. A truly Cosmos-native solution should maximize sovereignty and decentralization. This is our objective with the Protocol-Owned Builder (POB).
Protocol-owned builders put applications in full control
Recent consensus developments give applications access to new functionality that lets app-chains retain control of MEV. Both the ABCI++ capabilities included in CometBFT 0.34, and the Vote Extension feature coming in a future release, open a new design space for protocol-owned mempools and block building.
The Protocol-Owned Builder solution builds on learnings from the Skip x Osmosis ProtoRev collaboration, which performs a small portion of block construction, inserting cyclic arbitrage transactions to rebalance LP pools.
Properties of a Protocol-Owned builder
- Blocks are built communally by every validator, with results enforced by consensus
There are no off-chain builders, all building logic is trustless and transparent
- Governance has full control over block-building and payment distribution
The chain can vote how MEV is distributed among network participants, with payments enforced in consensus
- Searchers can still submit bundles privately for auction
Searchers see little to no change in how they would submit bundles compared to Skip’s builder today
- Equivalent, or superior, profitability to centralized solutions
Has maximally efficient liquidity - by in-housing building, some strategies can make use of infinite, flash-minted liquidity, and therefore can instantly bid up strategies to 100%. Profitable off-chain builders can be additionally compensated to improve the design of in-protocol builders for the community’s benefit
- Support for cross-chain MEV
Chains with a POB implementation can selectively allow coordination with other chains running POB (more on this design coming soon)
Building the future of Cosmos MEV together, starting on Osmosis
Moving MEV auctions on-chain will require innovation in credible auction design and mempool privacy. To get there we will need to introduce novel primitives via a new POB module. We plan to develop the Protocol-Owned Builder in stages, beginning with a transparent on-chain top-of block auction, which though imperfect, will allow us to iterate and eventually incorporate innovations including the https://github.com/anoma/ferveo threshold encryption scheme developed jointly by Osmosis and Anoma.
In the coming months Skip and Osmosis will design and implement the first POB on-chain. We are excited to build in the open and chart a new trajectory in on-chain, trustless MEV.
As we’ve begun to see in Ethereum, proposer-builder separation risks user-harm and off-chain coercion that threatens long-term protocol stability. Protocol-owned buildings are therefore essential for preserving the sovereignty of Cosmos.
We invite other Cosmos chains to collaborate in exploring the frontier of in-protocol MEV solutions with us