Maximum Extractable Value (MEV) is a complex concept whose effects affect all DeFi users.
MEV corresponds to the optimal arrangement of transactions within a block in order to extract the maximum value. It represents, in most cases, a loss for DeFi applications and their users, to the benefit of arbitrage bots, block creators and blockchain validators.
VEM nevertheless has several advantages: it helps to balance prices between the various DEXs and CEXs, and allows liquidations in the various lending applications to function smoothly.
We will first present a few examples of MEV, then describe how it is organised on Ethereum - the network with the most mature activity - and then on its layers 2 and Solana, before looking at what solutions are being put in place.
Some examples of MEV cases MEV encompasses many activities. Here is a non-exhaustive list:
The sandwich attack: a bot detects your buy order, executes its own before yours, then sells immediately after your purchase. Arbitrages between the various AMMs (Automated Market Makers) and centralised trading platforms. More specifically, CEX-DEX arbitrages often involve exploiting an AMM displaying an obsolete price - due to the update delay associated with the speed of block production - compared with a centralised exchange where the price moves with a much lower latency. This phenomenon, known as Loss-Versus-Rebalance (LVR), extracts value from liquidity providers. Oracle Extractable Value (OEV): this exploits the price discounting delay provided by oracles, enabling lending platform users to liquidate their loans in the way that is least advantageous for them (more information here) . Below is an estimate of MEV activity on Ethereum.
Source: Sorella Labs The MEV value chain on Ethereum When a transaction is carried out on Ethereum, it passes through several stages before being executed within a block. There are three main players in the MEV value chain: researchers, block builders and proposers.
Researchers are sophisticated, often automated, players who constantly scan blockchain activity for MEV opportunities, such as those mentioned above. In particular, they exploit the public mempool, where transactions are waiting to be included in a block, to preempt some of them by paying higher priority fees.
Block builders , also sophisticated entities, collect transactions from the public mempool and private mempools to organise them within a block. Their aim is to maximise the fees associated with the transactions they include. In this way, researchers compete with each other to have their transactions selected by the block builders.
Proposers are the validators of the Ethereum network. Every 12 seconds, a validator is randomly selected to propose the next block to be transmitted to the rest of the network. The proposer receives the blocks from the builders and chooses to propose the one that is most profitable for it, forcing the builders to share their profits with it.
Source: Flashbots We can see that there are several levels of competition within the value chain, and that only the proposer has a kind of monopoly, since there is only one at a time and it has the last word. The proposer is therefore the one who collects the lion's share of the profits associated with MEV, which ultimately helps to increase the rewards of ETH staking.
Some researchers in the Ethereum community want to put an end to this proposer monopoly. For example, Max Resnick is pushing for the implementation of Multi Concurrent Proposers (MCP) . The idea is to strengthen the Ethereum network's resistance to censorship in order to guarantee compliance with transaction organisation rules and therefore enable better predictability.
VEM on other blockchains On layers 2, VEM takes different forms. This is mainly due to the fact that they currently have a single sequencer that hosts the mempool privately and acts as both block builder and proposer. This situation limits sandwich attacks in particular, since only the sequencer is aware of the transactions. It also means that you have to trust the project team operating the sequencer.
Chains based on OP Stack organise their transactions in the same way as Ethereum: users all pay a common gas fee (base fee) to which may be added priority fees that allow their transaction to be placed higher in the block.
The rules for organising transactions may also differ. For example, on Arbitrum, transactions are simply ordered according to their order of arrival on the sequencer, on a "first come, first served" basis.
Solana, for its part, does not have a mempool and transactions are sent directly to the leaders who produce the blocks.
Not having a mempool or relying solely on a private mempool does not eliminate the MEV. The research by Flashbot thus indicates that these situations drive MEV strategies based on transaction spam in the hope of seizing the various arbitrage opportunities. This leads to higher transaction costs, network congestion and the cancellation of certain transactions. For example, on Solana, the transaction failure rate is around 40%.
The different solutions Rollup-Boost, developed in collaboration by Flashbots and Uniswap, enables layer 2s to order transactions fairly by following predefined rules while decentralising the sequencer. The idea is to use a trusted execution environment (TEE) that encrypts the mempool and guarantees that the transactions sent to it are executed fairly.
For more details on the various encryption methods and their limitations, you can read our analysis of them .
Source : Flashbots Guaranteeing the arrangement of transactions based on priority charges also allows applications to recoup the majority of the MEV generated by their users by implementing a "MEV tax" .
For its part, Arbitrum is considering the implementation of Timeboost, introducing two distinct transaction paths:
Normal path: transactions experience a short delay (default 200ms), but their processing remains unchanged - they are executed in the order they arrive. Express lane: transactions in this lane are not subject to any delay. Sophisticated players wishing to carry out arbitrage participate in minute auctions to obtain the exclusive right to use the express lane for the next minute. The costs of accessing this lane are paid into the DAO's treasury. This model seems to benefit the channel more than its applications.
On Solana, Jito has developed its own client for network validators as well as its liquid staking token, JitoSOL. Jito organises block space auctions to guarantee researchers the inclusion of batches of their transactions, preventing them from spamming the network in the hope that all their transactions will be executed.
On Ethereum, Sorella Labs is developing Angstrong, a DEX based on Uniswap V4, which uses the concept of Application Specific Sequencing (ASS). ASS enables an application to organise the order of its transactions within a batch. If this order is changed by the block builder, all the transactions are cancelled. This forces the builder to maintain the correct order of transactions, or risk losing revenue, and allows the application to internalise most of the MEV it generates.
The main disadvantage of SSA is the partial loss of composability between applications, since the transactions that concern them are no longer ordered according to a common logic (for more details, see this article) .
Conclusion We have found that MEV is an impossible phenomenon to suppress within blockchains, as these are opportunities for gains that will always be taken in one way or another. It is preferable for there to be a value chain with a clear organisation, because otherwise we see transaction spam and the business becomes more opaque.
One of the major objectives of a blockchain is to achieve a fair organisation of transactions thanks to predefined rules and strong resistance to censorship that guarantee the application of these rules.
VEM opportunities are all created by DeFi users, whether through swaps, the provision of liquidity within DEXs or loans made on lending platforms. It therefore seems logical that developments in the sector are tending towards a situation where most of the MEV can be redistributed to these same users.
Applications have every interest in internalising as much as possible of the MEV generated by their users in order to redistribute it to them if they want to remain attractive and develop their business. The challenge is to find the right balance between composability and internalisation of MEV, which varies according to the type of application.
In any case, it seems that MEV will constitute an increasingly small part of the revenues of the stakers of the different blockchains.
