Guide détaillé à venir
Nous préparons un guide éducatif complet pour le Layer 2 Fee Comparison Calculator. Revenez bientôt pour des explications étape par étape, des formules, des exemples concrets et des conseils d'experts.
A Layer 2 Fee Comparison Calculator provides real-time and historical cost analysis across major Ethereum Layer 2 scaling networks, enabling users to identify the cheapest chain for their specific transaction type. Layer 2 networks are blockchain protocols built on top of Ethereum (the Layer 1) that process transactions off the main chain while inheriting Ethereum's security guarantees. By batching hundreds or thousands of transactions into a single Ethereum transaction, L2s achieve dramatically lower per-transaction costs while maintaining the trustless security properties that make Ethereum valuable. As of 2025, the Layer 2 ecosystem has grown to encompass over a dozen major networks, each with different architectures, fee structures, and trade-offs. Optimistic rollups (Arbitrum, Optimism, Base) post transaction data to Ethereum and assume transactions are valid unless challenged during a 7-day dispute window. Zero-knowledge rollups (zkSync, Scroll, Linea, Starknet, Polygon zkEVM) generate mathematical proofs that verify transaction correctness, enabling faster finality but at higher computational cost. Validiums and volitions (some Polygon configurations) store transaction data off-chain for even lower costs but with reduced data availability guarantees. The EIP-4844 upgrade (also called Proto-Danksharding), implemented in March 2024, fundamentally changed L2 fee economics by introducing blob transactions. Before EIP-4844, L2s paid for data posting using expensive Ethereum calldata. After the upgrade, L2s use a separate blob fee market with dramatically lower prices, reducing L2 fees by 10-100x for most operations. A simple token swap that cost $0.50-2.00 on Arbitrum pre-4844 now costs $0.01-0.10. This cost reduction has made L2s competitive with alternative Layer 1 chains (Solana, BNB Chain) for the first time. The calculator decomposes L2 fees into their two components: the L2 execution fee (gas consumed by the L2 sequencer to process the transaction) and the L1 data fee (the cost of posting transaction data to Ethereum for security). Understanding this decomposition is essential because different transaction types have different ratios: a simple ETH transfer is cheap on both components, while a complex DeFi transaction with large calldata is expensive on the L1 data component. The calculator compares these costs across all major L2s for the user's specific transaction type.
L2 Total Fee = L2 Execution Fee + L1 Data Fee L2 Execution Fee = L2 Gas Used x L2 Gas Price (in L2 native gas units) L1 Data Fee = Compressed Calldata Size (bytes) x L1 Blob Fee Per Byte Pre-EIP-4844: L1 Data Fee = Calldata Bytes x (16 gas per nonzero byte) x L1 Gas Price Post-EIP-4844: L1 Data Fee = Blob Data Size x Blob Base Fee (separate fee market) Fee Savings vs L1 = L1 Transaction Cost - L2 Total Fee Savings Ratio = Fee Savings / L1 Transaction Cost x 100 Worked Example: Simple token swap on Uniswap V3. Ethereum L1: 150,000 gas x 25 gwei = 0.00375 ETH = $13.13 (at $3,500/ETH) Arbitrum: L2 execution = 250,000 L2 gas x 0.1 gwei = 0.000025 ETH ($0.088) + L1 data = 256 bytes blob data x $0.000015/byte = $0.004. Total: $0.092 Optimism: L2 execution: $0.095 + L1 data: $0.005. Total: $0.100 Base: L2 execution: $0.035 + L1 data: $0.002. Total: $0.037 zkSync: L2 execution: $0.085 + proof amortization: $0.025. Total: $0.110 Polygon PoS: Execution: $0.005 (separate fee market, not L2 rollup). Total: $0.005 Savings vs L1: Arbitrum saves 99.3%, Base saves 99.7%.
- 1Step 1 - Select the transaction type to compare. Different operations have vastly different gas profiles: a simple ETH transfer uses approximately 21,000 gas on L1, a token transfer uses approximately 65,000 gas, a Uniswap V3 swap uses approximately 150,000 gas, an NFT mint uses approximately 150,000-300,000 gas, and a complex multi-step DeFi operation can use 500,000-2,000,000 gas. Each L2 translates these L1 gas values differently based on its execution environment and compression efficiency. The calculator maintains benchmarks for common transaction types across all supported L2s.
- 2Step 2 - Fetch current fee parameters from each L2 network. The calculator queries the RPC endpoints of each L2 to obtain current gas prices, L1 data costs, and any network-specific fee parameters. Arbitrum uses an adaptive gas price mechanism that adjusts to congestion; Optimism and Base use EIP-1559-style fee markets with base fees and priority fees; zkSync has a separate fee model that includes proof generation costs; and Starknet uses a unique fee market denominated in STRK. These parameters change continuously, so the calculator provides real-time snapshots with timestamp annotations.
- 3Step 3 - Decompose the fee into L2 execution and L1 data components for each network. This decomposition reveals important structural differences. On Arbitrum and Optimism, the L1 data component was historically 80-95% of the total fee, meaning Ethereum mainnet gas prices dominated L2 costs. After EIP-4844, the L1 data component dropped to 10-30% of the total, making the L2 execution component more significant. On ZK rollups, there is an additional proof amortization component: the cost of generating the zero-knowledge proof, distributed across all transactions in the batch. This component is relatively fixed per batch, so it decreases per transaction as the L2 processes more transactions.
- 4Step 4 - Calculate the effective cost in USD for the selected transaction type on each L2. The calculator converts native gas unit costs (ETH, MATIC, STRK) to USD using current market prices. Some L2s have independent fee tokens (Starknet uses STRK, Polygon PoS uses MATIC) while others use ETH as the gas token (Arbitrum, Optimism, Base, zkSync). The USD conversion allows direct comparison across chains that use different gas tokens, though it introduces exchange rate variability as an additional factor.
- 5Step 5 - Present historical fee trends showing how costs have evolved over time. The calculator displays 7-day, 30-day, and 90-day fee charts for each L2, highlighting events that caused fee spikes (airdrop claims, NFT mints, network congestion, L1 gas spikes). Historical data reveals that L2 fees are most volatile during Ethereum L1 congestion events (which increase the L1 data component) and during L2-specific congestion events (popular airdrops, memecoin launches). The historical view helps users understand whether current fees are typical or anomalous.
- 6Step 6 - Compare additional factors beyond raw fee cost. The calculator evaluates: finality time (how quickly the transaction is considered irreversible), ecosystem depth (number of DeFi protocols, DEX liquidity, available tokens), withdrawal time to L1 (7 days for optimistic rollups, hours for ZK rollups), bridge availability and cost, and decentralization status (whether the sequencer is centralized, which affects censorship resistance). A chain might have the lowest fees but lack the DeFi protocols the user needs, making a slightly more expensive chain with better ecosystem support the practical choice.
- 7Step 7 - Generate a ranked recommendation based on the user's stated priorities (lowest cost, fastest finality, best ecosystem, most decentralized). The calculator produces a weighted score for each L2 based on the user's priority weights, with the default weighting favoring cost (40%), ecosystem depth (30%), finality speed (15%), and decentralization (15%). The recommendation includes specific routing advice: for example, suggesting that a user perform token swaps on Base (cheapest) while keeping lending positions on Arbitrum (deepest liquidity) and using zkSync for time-sensitive transfers (faster L1 finality).
For a simple ETH transfer, all L2s are dramatically cheaper than L1, with savings exceeding 97% across the board. Base offers the lowest cost at $0.012, benefiting from Coinbase's sequencer optimization and high transaction throughput that efficiently amortizes L1 data costs. Polygon PoS is even cheaper at $0.002 but operates as a sidechain rather than a true L2 rollup, meaning it does not inherit Ethereum's full security guarantees. For a simple transfer where the recipient is on a specific chain, the fee difference between L2s is negligible (fractions of a cent), and the choice should be driven by where the recipient can receive funds.
DEX swaps are the most common L2 transaction type, and fee differences become meaningful for frequent traders. A trader executing 10 swaps per day saves $157 daily by using Base instead of Ethereum L1, or $57,000 annually. Even among L2s, Base's $0.038 per swap is 60% cheaper than Arbitrum's $0.092, which can matter for high-frequency arbitrage and market-making strategies. However, Arbitrum has significantly deeper DEX liquidity (Uniswap V3 TVL of $2B+ versus Base's $500M+), meaning large swaps may experience less slippage on Arbitrum despite higher gas costs.
During Ethereum L1 congestion events (popular NFT mints, airdrops, memecoin launches), L1 gas prices can spike 4-10x above normal. Pre-EIP-4844, this would cause L2 fees to spike proportionally because the L1 data cost dominated. Post-EIP-4844, the L1 blob fee market is mostly independent of L1 execution gas, so L2 fee spikes during L1 congestion are much more moderate. In this example, L1 gas quadrupled from 25 to 100 gwei, but L2 fees only increased 2-3x because the blob market remained relatively stable. This decoupling is one of the most significant benefits of EIP-4844 for L2 users.
Complex DeFi operations involving multiple protocol interactions benefit the most from L2 pricing because the absolute gas savings scale with transaction complexity. A multi-step strategy that costs $70 on L1 costs less than $0.50 on most L2s, a 99.4%+ savings. This cost reduction has enabled entirely new DeFi strategies that would be uneconomical on L1: frequently rebalancing positions, actively managing concentrated liquidity, executing multi-leg options strategies, and performing small-amount dollar-cost averaging. The DeFi innovation happening on L2s is directly enabled by the fee economics that make experimentation virtually free.
DeFi protocol developers use L2 fee comparisons to decide which chains to deploy on. A protocol like Aave or Uniswap evaluates the fee economics of each L2 alongside TVL potential, user base, ecosystem incentives, and technical compatibility. Lower fees directly increase protocol usage because users execute more transactions when costs are negligible. Uniswap V3 on Base processes more transactions per day than on Ethereum mainnet despite Base having a smaller TVL, because the near-zero fee environment encourages more frequent trading. This demonstrates how L2 fee reduction does not just save money on existing transactions but fundamentally enables new transaction patterns.
Cryptocurrency wallets and portfolio management apps integrate L2 fee comparison data to provide automatic routing recommendations. When a user initiates a token swap through a wallet like MetaMask, Rainbow, or Coinbase Wallet, the wallet queries fees across available chains and suggests the cheapest execution route. If the user holds USDC on Ethereum and wants to swap for ETH, the wallet might recommend bridging to Base first (at $3 bridge cost) and then swapping on Base (at $0.03 gas) for a total of $3.03, versus swapping directly on Ethereum at $15. This routing optimization saves users money automatically and drives adoption of the most cost-efficient L2s.
Layer 2 network operators (Arbitrum Foundation, OP Labs, Matter Labs, Coinbase for Base) use competitive fee analysis to calibrate their sequencer pricing and attract users. When one L2 reduces fees (through technical optimization or pricing policy), others respond to maintain competitiveness. This competitive dynamic has driven L2 fees down rapidly since 2023, benefiting all users. L2 operators also use fee revenue data to project their sustainability: sequencer fee revenue must eventually cover operating costs (data posting, proof generation, sequencer infrastructure) without relying on token subsidies or foundation grants.
Academic researchers and policymakers studying blockchain scalability use L2 fee data to evaluate whether the Layer 2 roadmap is successfully scaling Ethereum. Vitalik Buterin and the Ethereum Foundation have repeatedly stated that Ethereum's scalability strategy centers on L2 rollups rather than increasing L1 capacity. The dramatic fee reduction from EIP-4844 (10-100x cheaper L2 transactions) validated this approach, and further improvements from full Danksharding (expected 2026-2028) are projected to reduce fees by another 10-100x. Researchers track whether fee reduction translates into proportional usage growth and whether the security properties of L2s are maintained as costs decrease.
During the Base memecoin mania of March 2024, Base experienced its first major
During the Base memecoin mania of March 2024, Base experienced its first major congestion event when dozens of memecoin launches drove transaction demand far beyond the sequencer's capacity. L2 gas prices on Base spiked 100x from their normal levels, with swap fees briefly exceeding $5-10, comparable to Ethereum mainnet during moderate congestion. This event demonstrated that L2s are not immune to congestion-driven fee spikes and that the fee advantage of L2s over L1 diminishes during extreme demand surges. However, the spike was short-lived (hours rather than days), and the Coinbase team implemented sequencer capacity upgrades that reduced the frequency of similar events. The concept of super-transactions or intent-based execution is emerging as a way to further reduce effective L2 fees. Instead of the user submitting a transaction directly to the L2 sequencer, they submit an intent (a description of what they want to achieve) to a network of solvers who compete to execute the intent at the lowest cost. The solver might batch the user's operation with other compatible operations, execute across multiple L2s to capture the best pricing, or use MEV-aware routing to avoid sandwich attacks. Projects like UniswapX, CoW Protocol, and 1inch Fusion use this approach, effectively reducing the user's gas cost to zero (the solver pays gas and recovers the cost from execution optimization). The upcoming Pectra and Fusaka Ethereum upgrades are expected to further reduce L2 fees by increasing the blob target from 3 to 6 per block and eventually to much higher numbers with full Danksharding. This would double the available data space for L2s, reducing the L1 data component of L2 fees by roughly 50%. Combined with ongoing improvements in L2 compression techniques (Arbitrum Nitro, Optimism Bedrock, Scroll's EVM-equivalent optimizations), the trajectory points toward L2 fees of $0.001-0.01 for most operations within 2-3 years, making blockchain transactions cost-competitive with traditional payment networks for the first time.
| L2 Network | Type | ETH Transfer | Token Swap | NFT Mint | Complex DeFi | L2->L1 Withdrawal Time |
|---|---|---|---|---|---|---|
| Arbitrum One | Optimistic Rollup | $0.03-0.05 | $0.08-0.15 | $0.10-0.25 | $0.30-0.60 | 7 days |
| Optimism | Optimistic Rollup | $0.03-0.06 | $0.08-0.18 | $0.12-0.30 | $0.35-0.70 | 7 days |
| Base | Optimistic Rollup | $0.005-0.02 | $0.02-0.06 | $0.04-0.12 | $0.10-0.30 | 7 days |
| zkSync Era | ZK Rollup | $0.03-0.08 | $0.08-0.20 | $0.12-0.35 | $0.40-0.80 | ~24 hours |
| Polygon PoS | Sidechain | $0.001-0.005 | $0.003-0.01 | $0.005-0.02 | $0.01-0.05 | ~30 minutes |
| Scroll | ZK Rollup | $0.04-0.10 | $0.10-0.25 | $0.15-0.40 | $0.50-1.00 | ~4 hours |
| Linea | ZK Rollup | $0.03-0.08 | $0.08-0.18 | $0.12-0.30 | $0.35-0.75 | ~8 hours |
| Starknet | ZK Rollup (STARK) | $0.02-0.05 | $0.05-0.12 | $0.08-0.20 | $0.20-0.50 | ~12 hours |
Why are Base and Polygon so much cheaper than Arbitrum and Optimism?
Base benefits from aggressive sequencer optimization by Coinbase and high transaction throughput that efficiently amortizes L1 data costs across more transactions. Polygon PoS is fundamentally different from rollups: it is a sidechain with its own validator set rather than an L2 that posts data to Ethereum. This means Polygon does not pay L1 data fees at all, making it inherently cheaper but with weaker security guarantees (Polygon's security depends on its own validators rather than Ethereum's). Among true rollups, fee differences are driven primarily by sequencer efficiency, compression techniques, and pricing policies, with margins of 2-5x between the cheapest and most expensive.
What is EIP-4844 and how did it reduce L2 fees?
EIP-4844, implemented in March 2024, introduced a new transaction type called blob-carrying transactions. L2 rollups post their transaction data to Ethereum using these blobs, which are priced in a separate fee market from regular Ethereum transactions. Before EIP-4844, L2s competed with all Ethereum users for calldata space, making L2 data posting expensive during congestion. After EIP-4844, blob space has its own supply and demand dynamics, with a target of 3 blobs per block and a maximum of 6. The separate market has kept blob fees 10-100x cheaper than calldata fees because blob demand has not yet consistently exceeded supply. Future upgrades (PeerDAS, full Danksharding) will further increase blob capacity.
Are L2 fees fixed or do they fluctuate?
L2 fees fluctuate based on multiple factors. The L2 execution component varies with L2 network congestion (more demand raises the L2 gas price). The L1 data component varies with Ethereum blob fee market pricing (when many L2s compete for blob space simultaneously, blob fees increase). The most volatile fee component is the L2 gas price during usage spikes: during a popular airdrop claim on Arbitrum, L2 gas prices can spike 10-50x above normal for minutes to hours. Base experienced similar spikes during memecoin launches in 2024. Outside of congestion events, L2 fees are relatively stable day-to-day, with gradual trends driven by L2 adoption growth and Ethereum protocol upgrades.
Should I use the cheapest L2 for everything?
Not necessarily. The cheapest L2 may lack the DeFi protocols, liquidity, or specific tokens you need. If you want to trade a token that only has deep liquidity on Arbitrum, using Base (cheaper gas) would result in higher slippage that far exceeds the gas savings. For simple transfers and common token swaps, the cheapest L2 is usually the best choice. For complex DeFi strategies involving lending, liquidity provision, or derivatives, choose the L2 with the deepest ecosystem for your specific use case. Many advanced users maintain active wallets on 3-4 L2s and route each transaction to the optimal chain based on the specific operation.
What are ZK rollup fees and how do they compare to optimistic rollups?
ZK rollups (zkSync, Scroll, Linea, Starknet, Polygon zkEVM) incur an additional fee component for proof generation that optimistic rollups do not have. The zero-knowledge proof must be computed for every batch of transactions, and this computational cost is amortized across all transactions in the batch. Currently, proof generation costs make ZK rollups slightly more expensive than optimistic rollups for the same transaction type. However, ZK rollups offer faster finality (proofs can be verified on L1 in hours rather than the 7-day challenge period for optimistic rollups), which has value for users and protocols that need faster settlement. As proof generation technology improves and ZK rollup adoption increases, the proof cost per transaction is expected to decrease.
How do L2 fees compare to alternative L1 chains like Solana?
Post-EIP-4844, L2 fees are broadly competitive with alternative L1 chains. Solana fees are approximately $0.001-0.01 for simple transactions, comparable to Base and Polygon. However, during Solana congestion events, priority fees can spike to $1-10, similar to L2 congestion spikes. The key difference is that L2s inherit Ethereum's security and liquidity ecosystem, while alternative L1s have independent security models and smaller DeFi ecosystems. For users already in the Ethereum ecosystem (holding ETH, ERC-20 tokens, or NFTs), L2s are almost always more cost-effective than bridging to an alternative L1 and back.
Do L2 transactions have the same security as Ethereum mainnet?
Rollup L2s (Arbitrum, Optimism, zkSync, Base) inherit Ethereum's security because all transaction data is posted to Ethereum L1 and can be verified by anyone. If the L2 sequencer acts maliciously, users can force-exit their funds back to L1 using the data available on Ethereum. However, most L2s currently operate with centralized sequencers that could theoretically censor transactions temporarily (though not steal funds). The decentralization of sequencers is an active area of development for all major L2s. Sidechains like Polygon PoS have independent security that depends on their own validator set, which is generally considered weaker than Ethereum's security.
Conseil Pro
For maximum fee savings, time your L2 transactions during off-peak hours (weekday nights and weekends in US time zones) when both L1 blob fees and L2 sequencer congestion are lowest. Monitor the blob fee market using tools like blobscan.com to identify windows when blob utilization is below the target (3 blobs per block), as L1 data fees drop to minimum levels during these periods. For recurring transactions like weekly DCA purchases, set up automated execution during known low-fee windows rather than executing manually during peak hours.
Le saviez-vous?
Before EIP-4844, the most expensive single transaction ever executed on Ethereum Layer 2 was a complex DeFi operation on Arbitrum during the March 2023 gas crisis that cost the user $47 in L2 fees. After EIP-4844, the same operation would cost approximately $0.30. The 99.4% fee reduction from a single protocol upgrade demonstrates the extraordinary leverage that Ethereum's modular scaling architecture provides: improvements to the base layer multiply benefits across all Layer 2 networks simultaneously.