What is On-chain Arbitrage? Opportunities Beyond MEV
Many people assume on-chain arbitrage is a game reserved for professional quant teams—requiring coding skills, computational power, and gas bidding wars, leaving no room for retail investors. This is only half true. While some forms of arbitrage (like MEV sandwich attacks) have become too competitive for ordinary participants, on-chain arbitrage encompasses far more than just MEV. Some opportunities inherently don't require speed, and some don't even require coding. This article breaks it down into three layers: the current state of the most competitive MEV arbitrage, how to do arbitrage without speed, and new developments emerging in 2026.
A leading global cryptocurrency platform,suitable for both beginners and experienced traders.
New user benefit: 20% off trading fees upon registration!！
MEV Is Not the Whole Picture of On-chain Arbitrage
First, let's correct a common misconception. Many people conflate "on-chain arbitrage" with "MEV," but MEV is just one type of on-chain arbitrage, not the entirety.
MEV (Maximal Extractable Value) refers to the additional profit miners or validators can extract by reordering transactions within a block. Typical scenarios include sandwich attacks—placing buy and sell orders around a user's large transaction to profit from slippage. Arbitrage is the most "legitimate" form of MEV—identifying price differences between different DEXs and executing a single transaction to lock in profits.
However, MEV arbitrage has a fundamental problem: it's essentially a "zero-sum game." Your profit comes from the extra money users pay due to slippage. Competition in this arena has become brutally intense. In June 2026, Ethereum's most famous MEV bot, Jaredfromsubway.eth, was targeted in an attack, losing over $7.5 million. The attacker deployed 66 fake contracts disguised as mainstream assets weeks in advance, creating seemingly real arbitrage paths. When the bot automatically executed trades and completed approvals, the attacker transferred all assets out. Even top-tier "predators" can be preyed upon, proving this field is no longer accessible to retail investors.
Speed-based Arbitrage: Retail Traders Have Little Chance
Recent changes on the Bitcoin mainnet further illustrate a clear fact: the era of arbitrage based on competing for gas in the public mempool is ending.
The mechanism for on-chain arbitrage is simple: whoever pays the highest gas fee gets their transaction included first. But around June 2026, a new phenomenon emerged on the Bitcoin blockchain—the highest bidder doesn't always win. Some transactions with only 20 satoshis in fees (extremely low) won 207 times in a week; others consistently paid twenty-something satoshis but mostly failed.
The underlying reason: mining pools privately accept "off-chain transaction fees"—actual fees are paid off-chain directly to the pool, while the on-chain fee is symbolically set extremely low. This "dark pool" mechanism is similar to Ethereum's Flashbots but more covert on Bitcoin, with no public products and relying entirely on private relationships. As long as ordinary retail traders compete in the public mempool, they are playing a losing game against an invisible opponent.
The situation on Ethereum is similar. Academic research analyzing arbitrage data on Base chain from June 2025 to February 2026 found that probabilistic search (repeated submission attempts, relying on luck to catch opportunities) accounted for 23% of arbitrage activities but consumed 95% of on-chain spam transactions and 20% of Base network gas fees. After Base adjusted its configuration, protocol fee revenue shifted from "spam transactions" to "successful arbitrage," further compressing the survival space for probabilistic search bots.
In short, speed-based arbitrage has become a highly industrialized market—with private channels, institutional-grade infrastructure, and professional teams operating 24/7. Manual retail operations have virtually no chance of winning.
Arbitrage Without Speed: A Few Paths for Retail Traders
So, can ordinary people still do on-chain arbitrage? Yes, but only if they choose the right battlefield.
Path One: Cross-PERP DEX Spread Arbitrage
This is a direction that doesn't require speed. The core logic: the mark price or funding rate of the same asset on different decentralized perpetual exchanges occasionally deviates briefly. Traders can go long on the lower-priced platform and short on the higher-priced one, then close both positions when prices converge to lock in profits. This doesn't depend on market direction and is theoretically "market neutral."
Currently, the most arbitrage-friendly KYC-free perpetual DEXs include Hyperliquid, dYdX, and GMX. Hyperliquid uses an on-chain order book model with fast matching and decent depth; dYdX has a mature funding rate mechanism; GMX uses an oracle pricing model, and price deviations from order book platforms can be more significant in certain market conditions.
The core of this arbitrage isn't "who grabs it first" but "balancing deviation identification and execution speed." It typically requires a script to monitor spreads, but the barrier is far lower than MEV arbitrage—no need for private channels, mining pool connections, or competing for block header positions.
Path Two: Passive Arbitrage on Dynamic Weight AMMs
This is a relatively new direction. Some AMM (Automated Market Maker) pools adjust asset weights over time—essentially executing a "quantitative rebalancing strategy" automatically. When weights change, the pool's price temporarily deviates from the market price, creating arbitrage opportunities.
Academic research found that on a dynamic weight pool on Ethereum mainnet, each arbitrage extracted approximately $2.58 in July 2025, dropping to $0.28 by January 2026. Arbitrage frequency increased from once every 30 blocks to once every 7.5 blocks. Increased competition compressed arbitrage profits, but the pool itself achieved better execution efficiency.
More interesting is the situation on Base chain. A similar dynamic weight pool had a median profit per arbitrage of less than $0.001 in January 2026, with 98% of trades extracting less than 1 cent. These trades still occur because arbitrageurs aren't targeting this pool specifically—it's just one stop in their cross-multiple-DEX arbitrage paths, known as "incidental routing." This means manually arbitraging these pools yields little profit for retail traders, but if you understand cross-DEX routing, you can combine multiple trade paths to profit from overall spreads.
Path Three: Liquidation Arbitrage
In lending protocols, when a borrower's collateral ratio falls below a threshold, their position is liquidated. Liquidators can repay the borrower's debt and receive collateral as a reward. This type of arbitrage doesn't require speed; it requires "finding positions about to be liquidated" and "having sufficient funds to execute liquidation." Liquidation is crucial for the health of lending protocols and is one of the most "legitimate" revenue sources in the MEV ecosystem.
A leading global cryptocurrency platform,suitable for both beginners and experienced traders.
New user benefit: 20% off trading fees upon registration!！
Where Are the Risk Boundaries?
On-chain arbitrage is not risk-free. Cross-DEX arbitrage requires locking margin on both sides; if spreads don't converge quickly, capital is tied up, incurring opportunity costs. In extreme market conditions, a sharp move on one leg can lead to insufficient margin, even if the other leg has unrealized gains. It's advisable to leave ample extra margin buffer on both legs and avoid maxing out leverage.
Contract-level risks are also escalating. The attack on Jaredfromsubway.eth highlights a broader issue: automated arbitrage bots themselves can become targets. Attackers don't "hack" contracts but exploit bot behavior logic to set traps. For manual retail traders, this risk is relatively low, but if you start using scripts for automated arbitrage strategies, contract authorization management and security audits can no longer be ignored.
