Sandwich attacks are a symptom. The $1.5B+ extracted from users since 2020 is a measurable proxy for a larger, unquantified problem: the inherently adversarial execution environment of public mempools. This environment is the root cause.
mev-the-hidden-tax-of-crypto
Blog
Why Quantifying Sandwich Attacks Is Just the Tip of the Iceberg
Focusing on DEX sandwiches misses the larger, more complex MEV from liquidations, oracle manipulation, and cross-chain arbitrage. This is a deep dive into the hidden tax of crypto.
introduction
THE DATA
Introduction
Sandwich attack losses are a quantifiable symptom of a deeper, systemic design flaw in on-chain trading.
The real cost is opportunity cost. Quantified losses ignore the systemic inefficiency that forces protocols like Uniswap and 1inch to build complex, reactive mitigations. The true cost is the innovation and capital locked out by this fundamental flaw.
Data reveals the attack surface. Analyzing sandwich patterns exposes the predictable latency arbitrage that MEV searchers exploit. This is the same vulnerability that intent-based architectures like UniswapX and CowSwap are designed to eliminate by shifting to a declarative model.
Evidence: The Flashbots SUAVE initiative and protocols like Across, which use a commit-reveal scheme, are direct architectural responses to the adversarial execution problem that sandwich data makes visible.
key-trends
THE REAL FRONT-RUNNING THREAT
Executive Summary
Sandwich attacks are a visible symptom of a deeper, systemic MEV problem that distorts market fairness and extracts billions from users.
01
The Problem: Arbitrage is a Red Herring
Focusing on sandwich attacks misses the larger, more insidious MEV landscape. Generalized front-running and time-bandit attacks on PoS chains pose existential risks to consensus. The real cost is systemic inefficiency, not just stolen pennies.
- Long-tail MEV (liquidation bots, NFT sniping) is harder to quantify but equally predatory.
- Consensus-layer MEV threatens chain reorgs, undermining finality for profit.
>90%
of MEV Research
$1B+
Extracted Annually
02
The Solution: MEV-Aware Protocol Design
Protocols must be built from first principles to be MEV-resistant, not just MEV-aware. This requires architectural shifts, not just better detection.
- Batch Auctions (like CowSwap) and intent-based systems (like UniswapX) remove the granular transaction ordering that bots exploit.
- Encrypted Mempools (e.g., Shutter Network) and commit-reveal schemes hide information until it's too late to front-run.
~99%
Attack Surface Reduced
0 Latency
Advantage
03
The Metric: Total Extractable Value (TEV)
We must move beyond quantifying stolen value to measuring distorted value. TEV encompasses all value leakage caused by MEV's existence, including worse prices for all users and increased infrastructure costs.
- L1s/L2s must be benchmarked on their TEV/Total Value Locked (TVL) ratio.
- This shifts the focus from punishing searchers to evaluating protocol-level economic security.
10-100x
vs. Measured MEV
Key KPI
For VCs
04
The Entity: Flashbots & the PBS Future
Proposer-Builder Separation (PBS) is the only credible path to MEV democratization. Flashbots SUAVE aims to be a neutral, cross-chain block-building marketplace, but its success hinges on credible neutrality and widespread validator adoption.
- Centralization risk shifts from searchers to a few dominant builders.
- True solution requires decentralized PBS with permissionless builder sets.
~90%
Ethereum PBS Adoption
Multi-Chain
SUAVE Goal
thesis-statement
THE DATA
Thesis: The MEV Iceberg
Sandwich attack revenue is a visible but shrinking fraction of the total value extracted from users by MEV.
Sandwich attacks are surface-level data. They are the most visible and quantifiable form of MEV, but they represent a declining share of total extraction as protocols like UniswapX and CowSwap implement intent-based architectures that bypass public mempools.
The bulk is sub-surface arbitrage. The majority of MEV is now cross-domain arbitrage between centralized exchanges and DEXs, and liquidations on lending protocols like Aave and Compound, which are harder to track but extract more consistent value.
The foundation is infrastructure rent. The largest, most entrenched MEV is the searcher-to-builder payment (PBS) and the builder-to-proposer payment, a tax paid to validators for block space priority, institutionalized by Flashbots' MEV-Boost and relay networks.
Evidence: In 2023, sandwich attacks generated ~$120M, while total on-chain MEV exceeded $1.5B. The rest was cross-DEX arbitrage, liquidations, and infrastructure payments.
WHY SANDWICH ATTACKS ARE JUST THE TIP
The MEV Landscape: A Quantitative Breakdown
Quantitative comparison of MEV extraction vectors, their impact, and mitigation maturity. Data sourced from Flashbots, EigenPhi, and Chainalysis.
| Extraction Vector | Annualized Value (USD) | Detection Maturity | User Impact | Mitigation Status |
|---|---|---|---|---|
Sandwich Attacks | $300M+ | High (EigenPhi) | Slippage & Price Impact | Partial (MEV-Share, SUAVE) |
Liquidations | $200M+ | Medium (Gauntlet) | Collateral Seizure | Mature (Keeper DAOs, FLA) |
Arbitrage | $1B+ | Low (Decentralized) | Positive (Price Efficiency) | Not Required |
NFT Frontrunning | $50M+ | Low (Heuristic) | Failed Mint Gas | Nascent (Revert Protection) |
Long-tail JIT | $150M+ | Very Low | LP Dilution | Theoretical (MEV-Share) |
Cross-domain (LayerZero) | Unquantified | Nascent | Bridge Delay Exploit | Active Research |
deep-dive
THE DATA
Deep Dive: Beneath the Surface
Sandwich attack quantification reveals a systemic failure in transaction ordering, exposing deeper architectural flaws.
Quantifying MEV is reactive. Measuring sandwich attacks on Uniswap or Curve provides a damage report, not a solution. The core failure is the permissionless frontrunning enabled by public mempools and naive block building.
The real cost is trust. The measurable extractable value is a symptom of user experience degradation. Every user who pays a higher slippage tolerance or uses a private RPC like Flashbots Protect is paying a trust tax to avoid the system's default state.
Intent-based architectures solve this. Protocols like UniswapX and CowSwap shift the paradigm from transaction execution to outcome fulfillment. Users submit desired results, and solvers compete to fulfill them, eliminating the adversarial latency race at its source.
Evidence: Over $1.2B in MEV was extracted from Ethereum users in 2023, with sandwich attacks constituting a dominant share. This directly funds the infrastructure, like specialized block builders from builders like Flashbots and bloXroute, that perpetuates the cycle.
risk-analysis
BEYOND SANDWICHES
Systemic Risks of the Hidden MEV
Quantifying front-running is easy; the real danger lies in the opaque, systemic risks that erode blockchain's core guarantees.
01
The Problem: Consensus-Level MEV
Validators reorder or censor blocks for profit, compromising liveness and neutrality. This is a protocol-level failure that no user-level solution can fix.\n- Threat: $100M+ in extracted value from Ethereum block reorgs.\n- Impact: Undermines time-bandit attacks and proposer-builder separation (PBS) assumptions.
$100M+
Extracted Value
>50%
Relay Censorship
02
The Problem: Cross-Chain Arbitrage Loops
MEV bots exploit price discrepancies across Layer 2s and alt-L1s, creating systemic instability. This isn't simple arbitrage; it's a liquidity attack vector.\n- Mechanism: Bots drain liquidity from one chain's DEX to profit on another, causing cascading liquidations.\n- Amplifier: Fast bridges like LayerZero and Wormhole enable near-instant capital movement.
~500ms
Attack Window
10+ Chains
Surface Area
03
The Solution: Encrypted Mempools & SUAVE
Hide transaction content from searchers until execution. Flashbots' SUAVE aims to be a decentralized, cross-chain block builder for encrypted intents.\n- Core Tech: Threshold decryption and secure enclaves.\n- Limitation: Requires universal adoption to be effective; currently a coordination nightmare.
0%
Front-run Leakage
TBD
Adoption Rate
04
The Solution: Intent-Based Architectures
Shift from specifying transactions to declaring desired outcomes. Let specialized solvers (UniswapX, CowSwap, Across) compete on fulfillment.\n- Benefit: Users get MEV-free best execution without understanding the mechanics.\n- Risk: Centralizes power in a few solver networks, creating new trust assumptions.
+20%
Better Prices
~5 Solvers
Market Concentration
05
The Problem: Oracle Manipulation MEV
The most dangerous and underrated vector. Bots can profitably manipulate Chainlink price feeds to trigger or prevent on-chain liquidations worth billions.\n- Scale: A $50M manipulation can cause $500M+ in downstream liquidation value.\n- Defense: Requires decentralized oracle networks and time-weighted average prices (TWAPs).
500M+
Liquidation Value
1-2%
Feed Deviation
06
The Systemic Fix: MEV-Aware Protocol Design
Future L1s and L2s must bake MEV resistance into their core. This means native encrypted mempools, fair ordering rules, and MEV redistribution mechanisms.\n- Examples: Aztec with full privacy, Fuel with parallel execution reducing contention.\n- Trade-off: Often sacrifices some throughput or composability for robustness.
10x
Complexity Increase
New L1s
Primary Vectors
counter-argument
THE DEEPER THREAT
Counter-Argument: "But Sandwiches Are the Gateway Drug"
Sandwich attacks are a visible symptom, but the underlying market structure enabling them is the systemic risk.
Sandwiches reveal structural failure. They are the most visible exploit of a broken public mempool. The real threat is the entire class of latency arbitrage that extracts value from all users.
The iceberg is generalized MEV. Front-running, back-running, and time-bandit attacks on rollups are the submerged mass. Protocols like Flashbots SUAVE and CowSwap exist to solve this broader problem.
Quantifying sandwiches is insufficient. It measures a single tactic, not the total value leakage from users to searchers/validators. The systemic cost includes protocol design constraints and wasted gas.
Evidence: Research from Flashbots and Chainalysis shows sandwich volume is a fraction of total extracted MEV, which itself is a proxy for the market's inefficiency tax.
takeaways
MEV DEFENSE IN DEPTH
Key Takeaways for Builders
Sandwich attacks are just the visible symptom; the real challenge is the systemic, multi-layered MEV supply chain that extracts value across the entire transaction lifecycle.
01
The Problem: MEV is a Protocol-Level Tax
Sandwiching is the retail face of a ~$1B+ annual extractable value industry. The real drain is the opaque, multi-layered supply chain from searchers to builders to proposers. This acts as a direct tax on user transactions and creates systemic risks like chain reorgs.
- Latency Arms Race: Validators outsource block building to maximize MEV, centralizing power.
- Protocol Inefficiency: Value intended for LPs or users is siphoned by intermediaries.
- Unstable Base Layer: MEV incentives can lead to time-bandit attacks and consensus instability.
$1B+
Annual Extract
>80%
Blocks MEV-Boost
02
The Solution: Architect for Intents, Not Transactions
Move users from submitting vulnerable transactions to declaring desired outcomes. Let specialized solvers (like UniswapX or CowSwap) compete to fulfill the intent optimally, baking MEV protection into the UX.
- MEV Absorption: Solvers internalize arbitrage, converting extractable value into better prices for users.
- Privacy: Encrypted mempools or SUAVE-like systems prevent frontrunning.
- Expressiveness: Users define complex, cross-chain actions without managing gas or slippage.
~100%
Sandwich Proof
1-5%
Price Improvement
03
The Problem: In-App MEV is the Next Battleground
As on-chain activity moves into app-specific chains and rollups (dYdX, Aevo), MEV doesn't disappear—it morphs. The centralized sequencer model common to many rollups creates a single, privileged extractor.
- Sequencer Monopoly: The sole sequencer has perfect frontrunning insight on user order flow.
- Fragmented Liquidity: Cross-rollup arbitrage becomes a new, complex MEV category.
- Opaque Extraction: Users have zero visibility into value captured by the app's own infrastructure.
1
Central Point
New Vector
Cross-Rollup
04
The Solution: Enforce Fair Sequencing & Proposer-Builder Separation (PBS)
Design your chain's stack to separate transaction ordering from block building. Implement verifiable, fair ordering rules (like first-come-first-served) at the sequencer level to neutralize in-app MEV.
- PBS for Rollups: Adopt a decentralized sequencer set or a shared sequencer network (Espresso, Astria).
- Cryptographic Fairness: Use Timelock Encryption (like Shutter Network) to hide transaction content until block inclusion.
- Transparent Auctions: If MEV exists, make its capture permissionless and verifiable via auctions.
0ms
Frontrun Window
Decentralized
Sequencer Set
05
The Problem: MEV Leaks Across Bridge Vaults
Cross-chain messaging protocols (LayerZero, Axelar, Wormhole) and liquidity bridges (Across, Stargate) hold billions in vaults. These are prime targets for latency arbitrage and oracle manipulation during the attestation period.
- Slow Finality = Risk Window: The 10-20 minute delay for attestations is an arbitrage playground.
- Oracle MEV: Manipulating the price feed that determines cross-chain swap amounts.
- Liquidity Fragmentation: Inefficient capital deployment across chains creates arbitrage gaps.
$10B+
Bridge TVL
10-20min
Risk Window
06
The Solution: Build with Native Yield & Verifiable Delay
Design bridge economics so liquidity providers earn yield from inevitable arbitrage, not just fees. Use Verifiable Delay Functions (VDFs) or threshold signatures to create a forced, unbiased time delay for settlement, removing the latency advantage.
- MEV as LP Yield: Capture cross-chain arbitrage in vaults and distribute it to LPs (see Across model).
- Deterministic Finality: Use VDFs to make the attestation delay predictable and ungameable.
- Unified Liquidity Layers: Leverage shared liquidity networks (Circle CCTP, Chainlink CCIP) to reduce fragmentation.
VDF
Delay Enforced
LP Yield
MEV Recycled
ENQUIRY
Get In Touch
today.
Our experts will offer a free quote and a 30min call to discuss your project.
NDA Protected
Confidentiality24h Response
Time to ValueDirectly to Engineering Team
No Sales Fluff10+
Protocols Shipped
$20M+
TVL Overall