MEV Fundamentals

What Is Maximal Extractable Value?

Originally coined as “Miner Extractable Value” by Daian et al. in Flash Boys 2.0 (2019), MEV refers to the total value that can be extracted from block production beyond the standard block reward and gas fees. After Ethereum’s merge to proof-of-stake, the community adopted “Maximal” to reflect that validators, not miners, now hold this power --- and that the extraction extends well beyond the block producer to an entire supply chain of specialized actors.

Formally, for a block $B$ with transaction set $T$ and ordering $\sigma$:

$$\text{MEV}(B)=\underset{\sigma ,T^{\prime }\subseteq T\cup T_{\text{injected}}}{\max }\left[V(\sigma ,T^{\prime })-V({\sigma }_{\text{naive}},T)\right]$$

where $V(\sigma ,T^{\prime })$ is the value captured under ordering $\sigma$ with transaction set $T^{\prime }$, and ${\sigma }_{\text{naive}}$ is the default ordering (e.g., gas price priority). The key insight: the block producer can both reorder existing transactions and inject new ones.

An Engineering Analogy

If the blockchain mempool is a Kafka topic --- a stream of unordered pending transactions --- then the block producer is a consumer with a special privilege: it can reorder messages before committing them to the log. In a normal message queue, consumers process messages in arrival order. Here, the consumer gets to sort, filter, and insert its own synthetic messages before writing the batch. That reordering privilege is the source of MEV.

Historical Context

Daian et al. documented “priority gas auctions” (PGAs) on Ethereum in 2019, where bots competed by bidding up gas prices to front-run DEX trades. This created negative externalities: chain congestion, failed transactions consuming gas, and latency races that favored well-connected nodes. The paper established that MEV was not hypothetical --- it was already being extracted at scale.

Flashbots emerged in 2020 to move MEV extraction off-chain, creating a structured auction system rather than chaotic on-chain gas wars. On Solana, Jito Labs built analogous infrastructure, allowing searchers to submit transaction bundles that execute atomically --- crucial for MEV strategies that require multiple transactions to succeed or fail together.

Taxonomy of MEV Types

Arbitrage

The cleanest form of MEV. A price discrepancy exists between two venues (e.g., two AMM pools for the same pair), and a searcher submits transactions to buy on the cheap venue and sell on the expensive one within the same block.

Arbitrage is often considered “benign” MEV because it improves price consistency across venues --- the same role played by arbitrageurs in traditional market microstructure.

Liquidation

In lending protocols, positions that fall below collateral thresholds must be liquidated. Searchers compete to execute these liquidations and claim the liquidation bonus. This is structurally similar to how in traditional finance, margin calls create forced selling that intermediaries can profit from.

Sandwich Attacks

The predatory core of MEV. An attacker observes a pending trade, front-runs it to move the price, lets the victim trade at a worse price, then back-runs to capture the difference. Detailed mechanics are in Sandwich Attacks.

Just-In-Time (JIT) Liquidity

A sophisticated strategy where a searcher observes a large pending swap, mints a concentrated liquidity position around the expected execution price just before the swap, earns the trading fees, and removes the position immediately after. This extracts fees that would have gone to passive LPs.

JIT liquidity is a DeFi-native form of the latency arbitrage studied in traditional market microstructure --- the fast participant captures value from the slow one.

Backrunning

Placing a transaction immediately after a target transaction to capture a resulting opportunity. For example, after a large swap creates an arbitrage between pools, a backrunner corrects the price and pockets the difference. Less harmful than front-running, since the victim’s execution is unaffected.

Time-Bandit Attacks

A theoretical (and occasionally practical) attack where a validator with sufficient stake reorgs past blocks to capture MEV retroactively. This threatens blockchain finality itself. In practice, the economics of time-bandit attacks depend on whether the MEV in a past block exceeds the cost of reorging. Most chains have not seen sustained time-bandit activity, but the threat shapes protocol design.

The MEV Supply Chain

MEV extraction has industrialized into a layered supply chain:

Searchers monitor the mempool (or equivalent) for MEV opportunities. They run algorithms that detect arbitrage, liquidation, and sandwich opportunities, then construct transaction bundles. Searchers are the intelligence layer --- they find the value.

Builders aggregate bundles from multiple searchers into complete blocks, optimizing for total extracted value. On Ethereum, this role was formalized by proposer-builder separation (PBS). On Solana, Jito’s block engine serves an analogous function.

Validators (or proposers) select which builder’s block to propose. They capture MEV indirectly through tips paid by builders competing for block inclusion. Validators are the toll booth --- they don’t find MEV, but they control the bottleneck.

This separation creates a market structure:

$$\text{Searcher profit}=\text{Gross MEV}-\text{Builder tip}-\text{Gas cost}$$ $$\text{Builder profit}=\sum \text{Searcher tips}-\text{Validator tip}-\text{Opportunity cost}$$ $$\text{Validator profit}=\text{Block reward}+\text{Builder tip}$$

Competition among searchers drives most MEV value upstream to builders and validators. In equilibrium, searchers operate on thin margins while validators capture the bulk of the surplus.

MEV on Solana

Solana’s architecture --- 400ms block times, no public mempool by default, and leader-based block production --- creates a different MEV landscape than Ethereum. Without a persistent mempool, MEV extraction relies more on:

• Jito bundles: atomic transaction packages submitted to the Jito block engine, which validators running Jito’s modified client can include
• Leader proximity: low-latency connections to the current leader node
• Transaction spamming: submitting many copies of a transaction to increase inclusion probability

Estimates place Solana MEV in the hundreds of millions of dollars annually, driven substantially by memecoin trading on platforms like pump.fun and Raydium. The Jito tip mechanism alone has processed billions of dollars in cumulative tips, reflecting the scale of MEV competition on the network.

Socratic Questions

1. If MEV is the cost of decentralized transaction ordering, is there a theoretical minimum MEV for a given set of transactions? What would drive it to zero?
2. The MEV supply chain resembles a value chain with increasing market power at each layer. Is this consolidation inevitable, or could protocol changes distribute power more evenly?
3. Arbitrage MEV improves price consistency. Sandwich MEV degrades execution quality. How should a protocol designer distinguish “good” from “bad” MEV?
4. Solana’s lack of a public mempool was supposed to reduce MEV. Why didn’t it?
5. If validators capture most MEV in equilibrium, does MEV effectively become a security budget supplement? What are the implications for chains where MEV exceeds block rewards?

Further Reading

• Sandwich Attacks --- formal mechanics
• Protection Strategies --- defending against MEV
• Market Microstructure --- the TradFi foundations
• The Pump.fun Economy --- MEV extraction in Solana memecoin markets