Ethereum Layer 2 transaction fees have fallen sharply since Ethereum’s Dencun upgrade, with many L2 networks charging pennies or less for simple transfers and meaningfully lower fees for swaps. The main drivers are EIP-4844 “blob” space, improved batching and compression, and greater competition among rollups.

The decline is not uniform across all chains or applications. Fees can still rise during congestion, complex swaps remain more expensive than simple transfers, and blob prices can spike when demand for limited blob space increases.

Lower L2 fees matter because they make smaller on-chain actions more practical, including routine transfers, gaming transactions, social applications, and low-value DeFi interactions that were often uneconomical when data posting relied on calldata. The tradeoff is that users still need to compare security assumptions, liquidity, bridges, and application availability across networks rather than choosing only by the lowest fee quote.

Ethereum Layer 2 fees fell roughly 90% to 95% because Dencun introduced EIP-4844, which created a cheaper way for rollups to store data than legacy calldata. Most major L2s moved their batch posting to blobs and continued improving compression and batching. Sequencer competition and better user-experience infrastructure also helped lower end-user fees.

EIP-4844 created a separate market for rollup data, reducing a major cost item for L2 networks, according to Ethereum documentation from the Ethereum Foundation and the EIP-4844 specification. Wallets and sequencers now batch more transactions with stronger compression, lowering the per-user share of data costs. Some L2s also subsidize or optimize fees to attract users and developers. Even so, prices still vary by application type, time of day, and demand for blob capacity.

What Changed After EIP-4844

Before March 2024, most rollups posted transaction data to Ethereum as calldata. That approach worked, but it was costly because calldata competed with all other activity in Ethereum’s main gas market. With the Dencun upgrade, Ethereum added EIP-4844, allowing L2s to use temporary data blobs instead of placing rollup batches into calldata.

Blobs are cheaper because they are separate from normal gas, optimized for rollup data, and pruned after a fixed period. That change removed a large cost driver for L2 networks. As rollups moved from calldata to blobs, the per-transaction cost of posting data to Ethereum fell substantially.

Users did not need to take any specific action for this change to apply. Sequencers handle batch posting behind the scenes, and wallets began displaying lower fees because the data-posting component became cheaper. Ethereum’s documentation provides the core details on blob pricing and mechanics through the EIP-4844 specification and Ethereum Foundation materials.

A separate market for blob space also means rollups no longer compete as directly with DeFi traders and other L1 users for the same gas. When demand for L1 blocks rises, L2s are affected less than before, though not completely insulated.

How Rollups Convert Blob Space Into Lower User Fees

Rollups bundle many user transactions into a batch, compress the batch, and then post the compressed data to Ethereum for data availability. With blobs, the posting step is much less expensive. The more users a sequencer can pack into a single blob, the lower each user’s share of the cost.

This means fees can sometimes fall further during busier periods on a given L2 if the sequencer is able to batch transactions efficiently. More activity can lower per-user costs when batching is effective.

Compression has also improved. Teams have adjusted how they encode state differences and calldata, reducing the number of bytes per transaction before the data reaches a blob. These engineering improvements compound: slimmer batches reduce each user’s share of the total posting bill.

Competition is another factor. Multiple L2 networks are competing for the same users and developers. Beyond cheaper data posting, teams are improving mempool filters, batcher timing, and fee estimators to avoid overpaying for blob space. Some also run promotions or subsidize certain transaction flows, which can result in lower displayed fees even when the underlying ETH price rises.

Calldata, Blobs, and Alternative Data Availability Layers

The way a rollup stores its data affects both costs and security assumptions.

Most general-purpose Ethereum L2s adopted blobs soon after Dencun because blobs allowed them to remain within Ethereum’s security framework while reducing costs. Some ecosystems also experiment with external data availability to lower costs further or scale specific workloads. These architectural choices help explain the fee differences users see across chains and applications.

Why Fees Still Differ Between L2s and Applications

Not all transactions require the same resources. A simple transfer is small. A decentralized exchange swap touches multiple contracts and state slots. That additional state access and computation inside the rollup makes swaps more expensive even if the data-posting method is the same. Wallets typically reflect this difference: transfers are usually cheapest, while mints, multi-hop swaps, and similar actions cost more.

Each L2 also has its own software stack and proof system. Optimistic rollups and ZK rollups differ in how they compress data and submit proofs. Some sequencers have more advanced batchers. Some networks share infrastructure through OP Stack, which can help standardize cost curves, while others operate more customized pipelines. These differences create real fee variation.

Live fee boards such as L2Fees track common actions across major L2s and can help users compare current quotes.

Timing also matters. Blob space has its own base fee, which rises and falls with demand. Around popular NFT mints, airdrops, or broader market spikes, blob prices can rise. When that happens, fee estimates across blob-using rollups can increase together before declining as demand cools.

If a fee quote suddenly rises, waiting a minute or two may produce a different estimate because blob base fees update block by block.

Other Factors That Reduced Costs

EIP-4844 was the largest protocol-level change, but engineering improvements also contributed. Many teams shipped more efficient state differences, better compression dictionaries, and smarter batcher timing. If a rollup can fit 5% to 10% more transactions into the same blob, that reduces per-user costs.

Wallets and account-abstraction tools also changed real-world fee experiences. With ERC-4337-style tooling and paymasters, some applications can sponsor gas or allow users to pay fees in tokens they already hold. The transaction still consumes resources, but the out-of-pocket experience can be lower, and fragmented signature calls can be bundled into a single action.

Growth strategies have also affected visible fees. Sequencers compete, and some L2s temporarily subsidize selected flows or keep minimums low to attract developers and users. This can appear as consistently cheaper quotes for everyday transactions, although such subsidies are not guaranteed to continue.

Why L2 Fees Could Rise Again

Blobs are cheaper than calldata, but they are not unlimited. If demand exceeds blob capacity for extended periods, prices can rise until usage cools. This can occur during major on-chain events.

ETH’s dollar price also matters. Even when a fee remains small in ETH terms, a higher ETH price increases its dollar value. Costs can also rise if a rollup has not fully migrated to blobs or temporarily falls back to calldata during incidents.

Competition is another variable. If one L2 gains a dominant position and reduces subsidies, users could see higher fees. On Ethereum’s roadmap, full danksharding is intended to expand data capacity well beyond the current blob system. If that expansion arrives later than expected, the system could face longer periods of higher utilization. Ethereum roadmap updates are available from the Ethereum Foundation.

Choosing an L2 for Transfers or Swaps

Fees are only one factor when selecting an L2. Users commonly compare current quotes on a fee board such as L2Fees, verify that the intended application is available on the target L2, and check whether sufficient liquidity exists there.

Bridge routes and withdrawal times also matter. Native bridges are generally slower, while third-party bridges may be faster but introduce additional risk. Users may also review a network’s status page or social channels for incidents or upgrades and conduct a small test transaction before bridging significant funds.

For contract deployment, the decision involves a wider set of variables, including proof type, data path, ecosystem support, expected throughput, and batching behavior. ZK rollups may offer faster proof finality in some cases, while optimistic rollups often provide mature tooling and shared standards through frameworks such as OP Stack. A project’s cost curve depends on expected traffic and batching efficiency, not only the current quoted fee.

Common Operational Issues

Bridging solely to chase a small fee difference can be inefficient because bridge costs and waiting time may outweigh savings. The application and liquidity needed should exist on the target L2 before funds are moved.

Blob congestion can also affect estimates. If an NFT mint or airdrop is driving heavy activity, quotes can rise quickly, and wallet estimators may be inaccurate during surges. Waiting a few blocks or transacting off-peak can change the displayed cost.

Some users still rely on L1 for small transfers out of habit. After EIP-4844, many day-to-day transfers are substantially cheaper on L2 while still using Ethereum for data availability through blobs.

Approvals can add additional calls and costs when signing swaps. Reusing allowances where appropriate or batching supported actions inside an application can reduce separate transaction steps.

Bridge and RPC selection remains important. Unknown bridges or unofficial RPC endpoints can introduce risk, and fake bridges are a common source of loss.

Frequently Asked Questions

Did the Dencun upgrade cut Ethereum L1 fees?

Not directly. Dencun’s main user-facing impact was EIP-4844 blob space for L2 data. That moved most rollup posting away from calldata and out of Ethereum’s main gas market. L1 fees still depend on L1 demand. Users may see fewer indirect spikes from rollup activity, but general L1 usage patterns still determine L1 prices.

Are ZK rollups always cheaper than optimistic rollups?

No. ZK proofs can reduce some costs and improve finality, but total fees depend on batch size, compression, sequencer policy, and demand for blob space. Optimistic rollups using efficient stacks and blobs can sometimes match or beat ZK rollups for everyday transactions. Live quotes vary over time.

Why does an L2 swap cost more than a simple transfer?

Swaps touch more contracts and modify more state. That adds bytes to the batch and requires more work inside the rollup, which the sequencer reflects in the fee. Transfers are lighter and cheaper, while mints, multi-hop swaps, and NFT listings are usually more expensive.

How low can fees go before full danksharding?

Blobs have already reduced costs substantially. Additional gains are expected to come from better compression, fuller blobs, and smarter scheduling. Without another protocol-level expansion in data capacity, further reductions are more likely to be incremental than another 95% decline.

Do alternative data availability layers make transactions less secure?

They introduce different trust and operational assumptions. Posting data to Ethereum through calldata or blobs relies on Ethereum L1 security. External DA systems can be cheaper or more flexible, but they rely on their own validator sets or restaked security models. The appropriate choice depends on the application’s requirements and risk assumptions.

Why does a wallet show a small base fee but a large “network” line item?

Wallets often break fees into components. On L2, users may see the execution cost plus a share of the data-posting cost. During blob congestion, that data share can become larger. Refreshing the quote after a few blocks or checking another wallet can produce a different estimate.

Can users pay gas in stablecoins on L2?

Some applications on certain L2s use paymasters or sponsored transactions to let users cover fees with tokens they already hold. This is not universal, and limits or higher effective costs may apply. Where supported, it can reduce the need to top up native gas for small transactions.

Disclaimer: This article is provided for informational purposes only. It is not offered or intended to be used as legal, tax, investment, financial, or other advice.