JaredFromSubway

The origin of "Jared from Subway" in crypto: March 2023

JaredFromSubway was deployed to Ethereum mainnet in March 2023 (ENS: jaredfromsubway.eth, also tagged jared_eth on Etherscan), around the same window as the SEC sued Coinbase and Binance for unregistered securities trading. The name is a nod to the satirical Twitter account "Jared from Subway", the public face of every sandwich-attack joke in DeFi — the moniker stuck and Jared from Subway crypto became shorthand on Twitter and crypto media for the highest-volume MEV operator on Ethereum. Within a few weeks the bot was consistently profitable; within three months it was the dominant MEV bot on Ethereum, regularly capturing more weekly profit than the rest of the top-10 combined.

The bot's rise coincided with a meme-coin frenzy on Uniswap (PEPE, MILADY, MOG, HARRY POTTER OBAMA SONIC 10 INU, and dozens of derivatives), which generated massive sandwich opportunities — tens of thousands of users buying low-cap tokens with wide slippage tolerances. JaredFromSubway extracted hundreds of ETH per week through that period and never lost its lead afterward.

The contract

The JaredFromSubway smart contract is deployed at jaredfromsubway.eth. The source code is not verified on Etherscan — a deliberate choice that keeps the exact attack logic confidential. The bytecode reveals the high-level structure: custom entry points for swap-only attacks, sandwich attacks, JIT liquidity, and combined swap-JIT bundles. The contract holds a small ETH float for gas and accumulates WETH + stablecoins from successful attacks.

Because the contract is on-chain and every transaction is publicly auditable, the bot's entire profit history is independently verifiable. Flashbots MEV-Inspect and direct Etherscan analytics both confirm the 5,900+ ETH net figure. For an independent breakdown see the Etherscan verification page.

5-layer and 7-layer sandwich attacks

A traditional sandwich attack is 3 transactions: front-run, victim, back-run. JaredFromSubway pioneered multi-layer sandwiches:

• 5-layer: front-run A → victim A → center swap → victim B → back-run. Two victims are trapped in one bundle, with the center swap moving price between them so victim B pays even more.
• 7-layer: three victims, two center swaps, one front-run, one back-run. The mathematics of slippage tolerance get significantly harder, but the profit per bundle scales with the number of victims.

Multi-layer sandwiches require precise sizing — each layer must respect every victim's slippage tolerance, or one of them reverts and the whole bundle fails. The off-chain solver that computes optimal layer sizes is the bot's most closely guarded piece of IP.

The swap-JIT combo

JaredFromSubway also pioneered combining sandwich attacks with JIT liquidity in a single bundle. Step 1: swap to move the price (front-run). Step 2: add concentrated liquidity to the pool via a direct Uniswap V3 pool call (bypassing the NonfungiblePositionManager NFT layer). Step 3: victim swap executes at the degraded rate, and the bot collects LP fees on the victim's trade. Step 4: remove liquidity and back-run the swap. Three revenue streams from one block: price manipulation + LP fees + arbitrage tail.

The 99.9% builder bribe rate

MEV bundles include a coinbase transfer — an ETH tip paid directly to the block builder that includes the bundle. Most competing MEV bots tip 90–97% of their expected profit; JaredFromSubway tips up to 99.9%, keeping only 0.1% per attack. At small per-bundle profit, that 0.1% sounds trivial. At scale — 40,000+ bundles since inception — it compounds to 5,900+ ETH.

The ultra-high bribe rate is deliberate strategy. By outbidding every competitor on tip percentage, JaredFromSubway guarantees inclusion priority across all five major builders — beaverbuild, Titan, rsync-builder, BuilderNet, and Quasar. The result is the highest block-presence rate of any MEV bot on Ethereum: over 60% of blocks include at least one JaredFromSubway bundle during peak activity.

MEV attribution and the obfuscation arms race

Third-party MEV attribution pipelines classify on-chain transactions as sandwiches, arbitrages, JIT liquidity, or liquidations and tag them to specific operators. JaredFromSubway's multi-layer sandwiches initially broke standard attribution heuristics: classifiers expected 3-transaction patterns, and the 5- and 7-layer structures didn't match. For several weeks the bot's profit was misattributed to multiple smaller operators, masking the true scale of the operation. Detection logic has since been updated across the open-source pipelines (most notably Flashbots MEV-Inspect), but newer techniques (swap-JIT, dual-purpose transactions) continue to obfuscate attribution at the margin.

How MevLex compares

Where jaredfromsubway.eth extracts value through multi-layer sandwich attacks, MevLex runs a different strategy today: live on-chain arbitrage — cross-DEX, triangular, and multi-hop, funded from its own capital or flash loans — with every trade settling on-chain and independently verifiable. A dedicated sandwich / front-running engine like Jared's is coming soon as a separate MevLex app. See the pricing page for plans and access options.

Further reading and references

• Flashbots MEV-Inspect (mev-inspect-py) — Open-source MEV classification pipeline maintained by Flashbots — the canonical reference for attributing JaredFromSubway sandwich activity on-chain.
• Bankless coverage of JaredFromSubway — Editorial coverage of the operator's emergence as the dominant Ethereum sandwich bot in 2023.
• Etherscan contract page — Direct on-chain record of every transaction the bot has submitted since March 2023.