This article is for informational and educational purposes only. It does not constitute investment, legal, or financial advice, nor a recommendation or solicitation to buy, sell, or hold any digital asset, token, or financial instrument. Named protocols, projects, and tokens are referenced for technical illustration; the publisher holds no positions in and has received no compensation from any project, platform, or source named herein. Trading and on-chain automation involve substantial risk of loss, including from smart-contract bugs, gas spikes, latency, and capital lock-up. Certain extraction strategies described — including sandwich attacks — may constitute market manipulation under the Korean Virtual Asset User Protection Act and equivalent statutes elsewhere; this article describes observed behavior in the ecosystem and is not a technique the author endorses or recommends. Conduct your own research and consult a licensed advisor before making any decisions.

The Day I Discovered MEV

Someone's Front-Running Your Trade — And Making Millions

Imagine you're buying a house. You submit your offer, but before it goes through, someone who has access to the paperwork swoops in, buys the property first, then flips it to you at a higher price — all before you even realize what happened. You overpaid, they pocketed the difference, and nobody told you.

On Wall Street, this kind of thing gets you investigated by the SEC. On the blockchain, it happens every single second. And the people doing it are pulling in hundreds of millions of dollars a year.

Welcome to the world of MEV — a term I'd never heard of until I accidentally stumbled into one of the most competitive arenas in all of crypto.

So What Is MEV, Exactly?

MEV stands for Maximal Extractable Value. It was originally coined as "Miner Extractable Value" in the seminal 2019 paper Flash Boys 2.0 by Phil Daian and collaborators. The research community rebranded it to "Maximal" around 2021, recognizing that the concept applies to any block-building system, not just proof-of-work mining.

Here's the deal. On a blockchain, transactions get bundled into "blocks." The validators who create these blocks get to decide which transactions go in and in what order. That power — the ability to reorder, include, or exclude transactions — creates extractable value. That's MEV.

Think of it like a stock broker who gets to decide which customer orders hit the market first. If they could peek at every incoming trade, rearrange the queue to their advantage, and front-run the big ones — the extra profit they extract would be MEV. In traditional finance, that's illegal. In blockchain, it's baked into the architecture.

It's sometimes called the "invisible tax" of blockchain. When you swap tokens on a DEX and get a slightly worse price than expected, that gap might be someone's MEV profit. You paid a tax you never saw, to someone you'll never meet.

Wall Street Did This First

None of this is actually new. Wall Street figured it out decades ago.

High-frequency trading firms — the kind Michael Lewis wrote about in Flash Boys — lease server space literally in the same building as the NYSE. They spend hundreds of millions on proprietary fiber-optic cables between Chicago and New Jersey, shaving microseconds off their response times. Arbitraging tiny price differences between exchanges, front-running large institutional orders — all of this has been happening since before most people had smartphones.

Lewis's central claim — "The market is rigged" — was controversial when the book came out in 2014. Within the industry, it was more like a confirmation of what everyone already knew.

Blockchain MEV is essentially the crypto-native version of HFT. But there are two fundamental differences that make it an entirely different game.

What Makes Blockchain Different — Two Revolutions

Revolution #1: Radical Transparency

In traditional finance, HFT happens behind closed doors. Dark pools, private order flows, information asymmetry designed into the system. Retail investors never see what's coming.

Blockchain flips this on its head. On chains like Ethereum, pending transactions sit in a public queue called the "mempool." Anyone can see them. It's like filing your SEC paperwork in a glass case where every hedge fund can read it before your trade settles. (Solana actually works differently — it doesn't have a traditional mempool, which I'll get into in a later post.)

This transparency is a double-edged sword. Everyone can see the opportunities. But everyone can also try to exploit them. On Wall Street, you need millions in infrastructure, regulatory licenses, and industry connections to play the HFT game. On the blockchain, anyone who can write code can enter the arena. The barrier to entry is knowledge, not capital. At least in theory.

Revolution #2: Atomic Execution

This is the one that genuinely blew my mind. The concept of "all or nothing" execution.

In traditional finance, if you want to arbitrage a price difference between the NYSE and NASDAQ, you need to execute two separate trades — buy on one, sell on the other. If the price moves between those two trades, you eat the loss. That time gap between execution is pure risk. Any stock trader who's had a fill on one leg of a pairs trade but not the other knows this feeling.

On the blockchain, you can bundle both sides into a single transaction. Either both succeed, or both fail. There's no in-between. You buy and sell simultaneously, and if the sell side would result in a loss, the entire transaction — including the buy — automatically reverts as if it never happened.

It's like a deal where if any part falls through, the entire thing unwinds automatically — no lawyers, no disputes, no cleanup. In real estate, this would be a miracle. On the blockchain, it's just how things work.

On chains like Ethereum, this atomic execution enables something called flash loans — borrowing millions of dollars with zero collateral, using it for arbitrage, and repaying it all within the same transaction. Can't repay? The loan never happened. The entire transaction reverts as if you never borrowed a cent.

Walk into JP Morgan and try explaining that concept. "I'd like to borrow $50 million. No collateral. I'll pay it back in 12 seconds. If I can't, just pretend it never happened." They'd call security. On the blockchain, this happens thousands of times a day. Code replaces trust.

The MEV Playbook

There are several ways MEV gets extracted. Understanding them gives you a map of how this world actually operates.

Arbitrage: Token X is cheap on DEX A and expensive on DEX B. Buy on one, sell on the other, pocket the difference. It's the crypto version of noticing gas is $3.50 at one station and $3.80 across the street — except a bot does the round trip in milliseconds with much lower (though not zero) risk of getting stuck holding inventory, because the trades execute atomically. Everything happens atomically.

This is exactly what I'm trying to build a bot for.

Sandwich attacks: A bot spots a large swap queued up, buys the token right before the large trade (pushing the price up), then sells right after (profiting from the price movement). Imagine someone at a real estate auction who bids up a property right before your turn, then immediately flips it to you at a markup. You overpay, they profit, and it happens faster than you can react.

This strategy is ethically controversial because it directly harms regular users. It's so problematic that Jito, Solana's dominant MEV infrastructure, shut down its own mempool feature in March 2024 specifically because sandwich bots were abusing it.

Backrunning: Instead of jumping ahead, this strategy drafts right behind a large trade — like a NASCAR driver tucking in behind the leader to save fuel and wait for the right moment to make a move. When a big swap pushes prices out of balance across DEXes, a backrunner steps in to correct the imbalance and captures the spread. Less predatory, more opportunistic. It actually makes markets more efficient.

Liquidations: In DeFi lending protocols, when a borrower's collateral drops below a certain threshold, liquidation bots race to close the position and collect a reward. This is MEV that actually serves a critical purpose — it keeps lending protocols solvent. Without liquidation bots, bad debt would pile up and threaten the entire system. They're simultaneously profit-seeking and systemically important. Strange combination.

An Invisible Economy Worth Hundreds of Millions

Let me put some numbers on this to make it concrete.

On Solana alone, industry estimates from on-chain research outfits have put annual MEV-related revenue in the hundreds of millions of dollars as of 2025; methodologies and totals vary across sources and have not been independently verified by the publisher. Ethereum's reported figures are larger. Combined across all chains, we're looking at a multi-billion dollar economy that most people have never heard of.

Why haven't they heard of it? Because MEV is invisible by design. When you swap a token and get a slightly worse price than expected — maybe 0.1% worse — that fractional difference is someone else's revenue. Multiply that across millions of swaps per day, and it adds up fast.

Jito, the dominant MEV infrastructure on Solana, runs the block engine that the majority of validators use. On peak days, it processes billions of dollars in transaction value. This isn't a niche sideshow — it's deeply embedded in how the network actually functions.

Here's what makes it interesting: not all MEV is harmful. Arbitrageurs equalize prices across DEXes, making markets more efficient. Liquidation bots keep lending protocols healthy. But sandwich attacks directly extract value from regular users. There's a moral spectrum within MEV, and where you sit on it matters.

Why Solana?

There are plenty of blockchains out there. Ethereum, Arbitrum, Base, Polygon. Why am I starting with Solana?

A few reasons.

Speed. Solana produces a new block roughly every 400 milliseconds. Ethereum takes 12 seconds. In arbitrage, speed is opportunity. The faster opportunities appear and disappear, the more an advantage in reaction time matters.

Fees. Solana's transaction fees are extremely low. On Ethereum, a single swap can cost tens of dollars during high congestion — and arbitrage margins are typically razor-thin. If the fee exceeds the profit, the trade is worse than useless. On Solana, fees are essentially a rounding error.

Different rules. Solana doesn't have a public mempool like Ethereum. Transactions go directly to validators via a mechanism called Gulf Stream. This makes Ethereum-style front-running structurally difficult. Instead, MEV on Solana works through Jito's bundle auction system — searchers submit transaction bundles that compete based on tip efficiency — the more you're willing to pay per unit of compute, the higher your priority. Different battlefield, different tactics.

That "different" part is both a challenge and an opportunity. The playbooks that work on Ethereum don't directly translate. But it also means I'm not competing head-to-head with established Ethereum MEV operations that have years of infrastructure built up. A new arena might mean new openings.

Why I'm Jumping In

I had no idea any of this existed until very recently. I came in thinking, "I'll build a bot that finds price differences between DEXes and trades automatically. Easy money, right?"

A market worth hundreds of millions of dollars. What I naively thought was risk-free arbitrage made possible by atomic execution. Just write good code and let it rip — how hard could it be?

The more I learn, the more I realize how naive that sounds. This is a battlefield where hundreds of bots compete in millisecond-scale races. It's a game requiring infrastructure, strategy, and capital — all at once. A handful of well-resourced operations dominate the space, and the gap between a solo developer working from a regular server and a funded team co-located next to validators is enormous.

But I'm going to try anyway. This is a real-time record of that attempt. What I learn, where I get stuck, how I work through it — success or failure, documented honestly as it happens.

Here goes nothing. Let's see how far this takes me.