Ethereum Gas Fees Drop Below 5 Gwei as Dencun Upgrade Impact Deepens

Ethereum Gas Fees Hit Historic Lows

Ethereum gas fees have fallen below 5 gwei on a sustained basis for the first time, with the 7-day average sitting at 4.3 gwei as of February 4, 2026. A simple ETH transfer now costs roughly $0.25, while a token swap on Uniswap runs about $1.50. These figures represent a 95% decline from the 80-120 gwei fees that were common during peak congestion periods in 2021 and early 2022.

The fee reduction is not the result of declining network usage. Ethereum mainnet processes approximately 1.1 million transactions per day, consistent with levels seen throughout 2025. The difference is that much of the activity that previously congested the mainnet has migrated to Layer 2 rollup networks, which now handle the bulk of user-facing transactions while only posting compressed data summaries back to Ethereum.

This shift represents exactly the outcome Ethereum's roadmap envisioned. The network is transitioning from a platform where all activity happens on a single congested chain to a modular architecture where the mainnet serves primarily as a secure data availability and settlement layer, while Layer 2s handle execution at scale.

How Dencun and Blob Transactions Changed the Fee Structure

The Dencun upgrade, which went live in March 2024, introduced EIP-4844 (also known as proto-danksharding). This created a new transaction type called "blobs" that provides dedicated data space for Layer 2 rollups at a fraction of the cost of regular calldata. Before Dencun, rollups had to compete with all other Ethereum transactions for block space, driving up costs for everyone.

Blob transactions operate in a separate fee market from regular transactions. Each Ethereum block can include up to 6 blobs (approximately 750 KB of data), and blob fees adjust independently based on demand for blob space specifically. This separation means that a surge in NFT minting on the mainnet does not affect the cost of posting rollup data, and vice versa.

The impact has been dramatic. Arbitrum, the largest Ethereum rollup by transaction volume, saw its data posting costs drop from roughly $50,000 per day pre-Dencun to under $1,000 per day. These savings flow directly to users in the form of lower transaction fees. A token swap on Arbitrum now costs less than $0.01 in most cases.

Layer 2 Migration Accelerates

The fee economics created by Dencun have accelerated user migration to Layer 2 networks. Combined daily transactions across Arbitrum, Optimism, Base, zkSync, and other Ethereum rollups now exceed 12 million, compared to 1.1 million on mainnet. This 11:1 ratio has grown steadily from roughly 3:1 when Dencun launched.

Base, Coinbase's Layer 2 network built on the OP Stack, has seen particularly explosive growth. It now processes over 4 million daily transactions, driven by a combination of DeFi activity, social applications, and onchain gaming. The network's deep integration with Coinbase's 100 million+ user base provides a natural onramp that competing L2s lack.

Total value locked (TVL) across Ethereum Layer 2s has surpassed $48 billion, representing roughly 35% of Ethereum's total ecosystem TVL. A year ago, that figure was closer to 18%. The data suggests that users are not just experimenting with L2s but permanently shifting their primary activity away from mainnet for cost-sensitive operations.

Impact on ETH Burn Rate and Supply Dynamics

Lower gas fees have a direct impact on Ethereum's monetary policy. Under EIP-1559, implemented in August 2021, a portion of every transaction fee (the base fee) is permanently burned, removing ETH from circulation. When gas fees were high, more ETH was burned than issued to validators, making the token deflationary. At current low fee levels, the burn rate has fallen below the issuance rate.

In January 2026, approximately 18,500 ETH was burned through transaction fees, while validator issuance added roughly 27,000 new ETH to the supply. This means Ethereum's supply grew by about 8,500 ETH during the month, an annualized inflation rate of approximately 0.085%. While this is negligible in absolute terms, it represents a shift from the deflationary periods that ETH holders had come to expect.

This dynamic has reignited debates within the Ethereum community about the network's long-term economic model. Some argue that the "ultrasound money" narrative requires consistently high mainnet fees to maintain deflation. Others contend that Ethereum's value proposition should be measured by the total economic activity it secures (including Layer 2s) rather than the gas fees generated on mainnet alone.

What Users Pay for Common Transactions

For everyday users, the current fee environment makes Ethereum mainnet viable for transactions that were previously cost-prohibitive. Sending ETH to another wallet costs about a quarter, comparable to what Layer 2s charged just a year ago. Complex DeFi operations that once required $50-$100 in gas can now be executed for $2-$5.

The reduced costs particularly benefit smaller users who were priced out during high-fee periods. A user wanting to stake 0.1 ETH or make a $50 swap no longer faces gas fees that eat a significant percentage of their transaction value. This accessibility improvement could help Ethereum retain users who might otherwise have migrated permanently to alternative Layer 1 blockchains like Solana or Avalanche.

Still, Layer 2s remain substantially cheaper for most operations. Users who perform many transactions per day, such as active DeFi traders or gaming participants, still benefit from operating on Arbitrum, Base, or Optimism where fees are measured in fractions of a cent rather than fractions of a dollar.

Road Ahead: Pectra and Full Danksharding

The current fee environment is likely to persist and potentially improve further. Ethereum's next major upgrade, Pectra, is scheduled for mid-2026 and will increase the target number of blobs per block from 3 to 6 (with a maximum of 9). This effectively doubles the data throughput available to Layer 2s, putting further downward pressure on both blob fees and, indirectly, mainnet congestion.

Looking further ahead, full danksharding (the complete version of what Dencun's proto-danksharding introduced) will expand blob capacity by another order of magnitude. When fully implemented, Ethereum could support 64 blobs per block, providing enough data space for hundreds of rollups to operate simultaneously at near-zero data posting costs.

The combination of cheap mainnet transactions and ultra-cheap Layer 2 transactions positions Ethereum's ecosystem to handle mainstream-scale applications without the fee spikes that historically deterred adoption. Whether this translates to sustained user growth and increased demand for ETH as a settlement asset remains the central question for Ethereum's next phase.