Guide

Solana MEV explained

MEV — maximal extractable value — is profit someone captures by controlling where transactions land inside a block, not just whether they land. On Solana, that usually means bots watching the mempool (or private order flow) for profitable swaps, liquidations, and arbitrage, then racing to insert their own transactions ahead of or around yours. If you have ever swapped on a DEX and received noticeably worse prices than the quote showed, MEV is a likely suspect. This guide explains how MEV works on Solana, how it differs from Ethereum, what Jito bundles changed, and what traders and builders can do about it.

What MEV is in plain language

Blockchains process transactions in a specific order. Whoever chooses that order — historically validators, increasingly specialized block engines — can sometimes reorder trades to extract value. A searcher bot might see your pending swap on Raydium or Jupiter, buy the token just before you (pushing the price up), let your trade execute at a worse rate, then sell immediately after (the sandwich attack). The bot's profit is MEV; your loss is worse execution, often called slippage beyond your tolerance.

Not all MEV is harmful. Arbitrage — buying an asset cheap on one pool and selling dear on another until prices align — keeps decentralized markets efficient. Liquidations on lending protocols are MEV too: searchers race to repay underwater loans and claim collateral bonuses, which keeps protocols solvent. The debate is about extractive MEV (sandwiches, toxic backruns) versus productive MEV (arbitrage, liquidations).

On Solana, MEV economics tie directly into priority fees and validator tips. Searchers often pay validators (or Jito relays) a share of profit to guarantee their bundle lands in the block they want. Understanding that auction helps explain why fees spike during memecoin launches and why your swap might fail even with a generous tip.

Why Solana MEV is not Ethereum MEV

Ethereum MEV is tightly linked to Flashbots and builder markets around 12-second blocks. Solana produces a new leader slot roughly every 400 milliseconds, so opportunities appear and vanish faster. There is no single global mempool in the same sense — transactions propagate through gossip, and leaders see different subsets at different times.

Solana's account model also matters. Programs declare which accounts a transaction touches; the runtime can schedule non-conflicting transactions in parallel within a block. Two swaps on unrelated token pairs may execute simultaneously, while two swaps hitting the same liquidity pool must serialize. That parallelism reduces some contention but concentrates MEV on hot accounts — popular AMM pools, oracle feeds, and lending market reserves.

Compute budgets add another wrinkle. Heavy arbitrage routes burn compute units (CUs); searchers attach priority fees proportional to expected profit. During congestion, the same CU auction that delays your simple transfer also prices out smaller MEV bots, leaving only well-capitalized searchers. Our account model guide explains why hot accounts become serialization bottlenecks.

Common MEV strategies on Solana

DEX arbitrage

Automated market makers (AMMs) price tokens from pool reserves. When SOL/USDC trades at $148 on Orca and $149 on Phoenix, a bot buys low and sells high in one atomic transaction. This is generally beneficial: prices converge across venues. Users still pay indirectly through wider spreads during volatile periods when arbitrageurs demand higher edge to cover gas and failed-tx risk.

Sandwich attacks

A sandwich requires the attacker to land a frontrun buy before your swap and a backrun sell after. It works best when your slippage tolerance is loose — e.g. you accept up to 5% price movement — and the pool is shallow. The attacker does not need your private key; they only need to see your transaction before it executes and outbid you on priority fees for adjacent slots in the same block.

Symptoms: quoted price looked fine, executed price was worse, and Solscan shows two suspicious swaps from the same address immediately before and after yours. Our guide on reading transactions on Solscan helps you spot that pattern.

Lending liquidations

Protocols like Marginfi, Kamino, and Solend let users borrow against collateral. When collateral value falls below maintenance thresholds, anyone can liquidate the position and earn a bonus. Searchers monitor oracle prices and race to liquidate; this is MEV that keeps lending markets healthy, though it can spike during crashes when many positions go underwater at once.

NFT and token sniping

Bots watch for new pool creations, metadata updates, or Candy Machine mints, then buy within the first slots. Less classic MEV, but the same ordering game: first transaction wins. Launch designers sometimes use allowlists or commit-reveal schemes to reduce bot advantage.

Jito bundles and block engines

Jito operates a network of validators running modified client software that accepts bundles — ordered groups of transactions that must execute atomically (all succeed or all fail). Searchers submit bundles to Jito's block engine with a tip paid to validators. If the bundle is profitable, validators include it; tips are shared with stakers who delegate to Jito-enabled validators.

Bundles changed Solana MEV in two ways. First, they professionalized extraction: instead of chaotic spam, large searchers negotiate directly with block producers. Second, they created a partial private order flow channel — not every pending swap is visible to every bot, but bundles that pay well get priority.

For stakers, Jito-enabled validators often earn higher yields because tip revenue supplements inflation rewards. Our staking guide covers how commission and APY quotes differ when validators participate in MEV sharing. "Real" APY is not just inflation — it includes tips minus operator cut.

Jito is not the only path. Some applications route swaps through private relays or integrate MEV-protection modes that send transactions directly to cooperating validators, skipping public gossip where sandwich bots listen. Trade-offs include slightly higher latency and trust in the relay operator.

Priority fees, simulation, and the MEV auction

When you set a priority fee in Phantom or Jupiter, you are bidding for scarce block space and leader attention — the same resource MEV bots want. A swap during a hyped token launch might need a large tip just to land; that tip does not protect you from sandwiches unless the aggregator uses a protected route.

Always simulate before signing. Wallets and aggregators call simulateTransaction to estimate compute usage and catch program errors. Simulation does not show you future ordering risk — a tx that simulates cleanly can still be sandwiched — but it prevents signing doomed transactions. See our transaction simulation guide for RPC details and log reading.

Failed MEV bot transactions are common: another searcher won the race, the opportunity disappeared, or slippage checks reverted the bundle. Those failures still burn base fees and sometimes priority fees on individual legs. High failure rates during congestion are a sign you are competing in an MEV-heavy environment.

How to reduce harmful MEV as a user

Tighten slippage — Use the minimum tolerance that still succeeds. 0.5% on liquid pairs, higher only when you accept the risk. Loose slippage is an open invitation to sandwiches.
Prefer aggregators with protection modes — Jupiter and others offer routes that minimize sandwich surface; read the UI labels.
Split large trades — Smaller clips across time or pools reduce price impact and attacker profit, though you pay more fees.
Avoid peak congestion — Major mints and macro events flood the network; if timing is flexible, wait for calmer slots.
Check execution on explorers — Compare quoted vs executed amounts; persistent gaps mean you need tighter settings or a different route.
Use limit orders where available — On-chain limit books (e.g. Phoenix, OpenBook) let you define exact prices instead of market swaps.

No setting eliminates MEV on public mempools. The goal is to make attacks unprofitable relative to your trade size so bots target someone else.

What builders and protocol designers should know

If you operate a DEX, launchpad, or lending market, MEV affects user trust and retention. Mitigations include:

Commit-reveal or batch auctions for launches so buy order is not determined by raw speed alone.
Slippage-aware routing on the client that rejects routes with known toxic flow.
Private transaction endpoints for integrators willing to trust your infrastructure.
On-chain randomness or time windows for fair distribution (harder to implement without UX cost).
Monitoring — Track sandwich rates and abnormal slippage on your pools; spikes indicate you need architectural changes, not just higher default tips.

Program developers should minimize unnecessary account writes and keep swap interfaces idempotent where possible. Complex CPI chains burn more CUs, raising the bar for arbitrage but also user costs.

Key takeaways

MEV is value captured by controlling transaction order — sandwiches hurt users; arbitrage and liquidations often help market function.
Solana's fast slots and parallel execution change MEV dynamics versus Ethereum; hot AMM accounts are the main battleground.
Jito bundles channel professional MEV through tip auctions shared with stakers on participating validators.
Priority fees buy inclusion, not sandwich protection — use tight slippage, protected routes, and simulation together.
Builders reduce harm with fair launch mechanics, routing choices, and monitoring — not by pretending public swaps are MEV-free.