Ethereum Scalability and Gas Fees: How It Works in 2026

Scalability describes how many transactions a blockchain can process, how quickly they confirm, and what each one costs, while the network stays secure and decentralised. For Ethereum this is a long-standing tension. The base layer is designed so that a wide range of people can run a node and check every block themselves. That design caps how much work each block can contain.

When people say Ethereum needs to scale, they usually mean one of three things:

• Higher throughput, meaning more transactions per second across the ecosystem.
• Lower fees per transaction, especially for everyday actions such as swaps, transfers and lending.
• Faster finality, so users know a transaction is settled without long waits.

Ethereum's answer has been layered. The base layer focuses on security, data availability and final settlement. Rollups and other networks handle most of the execution. This approach trades some simplicity for more capacity, and it means users now choose a network as well as an app. Understanding that split is the first step to understanding why your fees vary.

Every Ethereum transaction consumes gas, a unit that measures computational work. A simple token transfer uses far less gas than a swap that touches several liquidity pools or a lending action that updates multiple positions. The total fee is the amount of gas used multiplied by the gas price, which is quoted in gwei, a small fraction of one ether.

Several factors push fees up:

• Demand: many users trying to transact in the same period compete for limited space in each block.
• Complexity: contracts that run more steps consume more gas.
• Urgency: users who want fast inclusion can bid higher priority fees.
• Market events: launches, liquidations and volatile price moves often create sudden spikes.

Because conditions change quickly, a fee you saw an hour ago may not hold. Wallets usually show an estimate, but the estimate is only as good as the recent block data. If a transaction is not urgent, waiting for quieter periods can lower the cost. If it is urgent, paying more may be the only way to get it included in a reasonable time.

In 2021 Ethereum introduced EIP-1559, which changed how the fee market works. Each block now has a base fee that adjusts up or down depending on how full the previous block was. The base fee is burned, meaning it is removed from circulation rather than paid to validators. Users also add a priority fee, often called a tip, which goes to the validator who includes the transaction.

For users, the practical effects are these:

• Wallets can estimate a maximum fee and a tip, and the network refunds any part of the maximum that is not used.
• The base fee moves gradually, so sudden jumps usually reflect a sustained run of full blocks.
• Setting a tip that is too low can leave a transaction pending for a long time, while setting one too high wastes money.

The burn mechanism links fee demand to the supply of ether, which is why fee levels and ether issuance are often discussed together. The core lesson for everyday users is simpler: you pay a base amount that tracks demand, plus whatever tip you choose. If a transaction is stuck, check the network's status and the fee settings in your wallet before resubmitting.

A rollup executes transactions in its own environment, then posts compressed transaction data and a proof or claim back to Ethereum. Ethereum checks the claim and keeps the final record. Because many transactions share the cost of one posting to the base layer, each user pays much less than they would on mainnet.

There are two main designs:

• Optimistic rollups assume transactions are valid unless someone proves otherwise within a challenge window. Withdrawals back to Ethereum often take roughly a week under standard settings, although faster third-party bridges exist.
• ZK rollups publish validity proofs that show the state changes are correct. They can offer faster withdrawals in many designs, but they are more complex to build.

Rollups are not identical in security or decentralisation. Some still use a single sequencer that orders transactions, and some have upgrade keys controlled by a small group. Before you rely on a rollup, read its documentation on sequencers, withdrawals, upgrades and what happens if the operator goes offline. Lending and DEX protocols such as Aave, Uniswap and Curve deploy on several rollups, so the same app may behave slightly differently on each chain.

A major step came with EIP-4844, often called proto-danksharding, which shipped in the Dencun upgrade in March 2024. It added a new transaction format that carries large chunks of data called blobs. Blobs are cheaper for rollups than the older method of posting data as calldata, and they are pruned from nodes after a short period rather than kept forever.

Why this matters:

• Rollups post their data more cheaply, which lowers the fees users pay on those networks.
• Blobs have their own fee market, so rollup posting costs do not compete directly with ordinary mainnet transactions.
• Nodes only need to store blob data briefly, which keeps the base layer manageable.

The design is a stepping stone. Full danksharding would expand blob capacity much further and use sampling so that nodes can verify data availability without downloading everything. Current blob capacity and fee levels change over time, so check the official Ethereum documentation or a live blob tracker for present values rather than relying on older figures. For users the key point is that cheaper rollup data is a structural improvement, not a one-off discount.

Ethereum's long-term plan includes further scaling of execution and data, plus improvements to decentralisation and resilience. Early discussions often used the word sharding, which described splitting the network into parallel chains. The current direction is more focused: keep one secure settlement layer, expand data capacity through blobs and danksharding, and improve execution through rollups and gas-limit increases where validators agree.

Commentary in 2025 and early 2026 described a five-year scaling plan with near-term gains in execution capacity and longer-term changes to the protocol. Roadmaps change, and promised timelines often slip, so treat any specific milestone as provisional until it appears in a mainnet release. The Ethereum Foundation's public research and blog posts, and the release notes for each client, are the most reliable places to see what has actually been deployed.

When reading about the roadmap, separate three questions: what is live, what is scheduled, and what is still research. Articles often blend these. A feature that has passed testnets and entered a mainnet fork is live. One described in a research post is not yet guaranteed. This distinction matters more than any headline capacity number.

Scalability does not mean every network suits every task. The right choice depends on how much you transact, how fast you need confirmation, and how much you rely on specific assets or apps. The table below compares the broad categories rather than naming exact fee levels, which change often.

Option	Typical strength	Typical trade-off
Ethereum mainnet	Deepest liquidity, strongest settlement and security	Higher and more volatile fees
Ethereum rollups	Lower fees, same Ethereum-style wallets and contracts	Bridge and sequencer risks, withdrawal delays on some designs
Alternative Layer 1 chains	Very different performance profiles and native apps	Separate security assumptions and smaller ecosystems in some cases

For many users, a practical approach is to keep large holdings on mainnet or a well-established rollup, use a rollup for frequent small actions, and check the same protocol on each chain before committing funds. Uniswap, Aave and Curve all run on multiple networks, so the app you know may already be available where fees are lower. Always confirm you are using the official deployment listed in the project's own documentation.

Moving assets between chains is often the most expensive and risky part of using cheaper networks. Bridges lock or burn assets on one side and release or mint them on the other, and some rely on messaging layers or third-party validators. Each design adds code and trust assumptions beyond the base chain.

The main options include:

• Canonical bridges, such as the Arbitrum Bridge for moving ETH and tokens between Ethereum and Arbitrum. These are usually the most conservative route, though withdrawals may be slower.
• Third-party bridges and intent-based protocols, such as Across, which aim for faster transfers using liquidity providers or solvers.
• Cross-chain messaging infrastructure, such as LayerZero V2, which apps use to pass messages and tokens between many networks.

Practical safety steps apply to all of them. Confirm the bridge address from official documentation, not from a search ad or a direct message. Send a small test amount first. Read what each approval permits, and revoke approvals you no longer need. Keep records of transfer transaction hashes in case you need support. Remember that a fast transfer is not the same as a safe one, and that bridge exploits have caused large losses in the past.

Staking is a separate cost and yield question, but it shares the same underlying network, so it helps to understand both. Ethereum secures its chain through proof-of-stake, which replaced proof-of-work in September 2022 in an event known as The Merge. Participants lock ether to run validators, and the protocol rewards honest behaviour while penalising some kinds of misconduct.

You do not have to run a validator yourself. Liquid staking protocols pool user deposits and issue a receipt token that represents the staked position. Lido issues stETH, and Rocket Pool issues rETH, and each has its own fee structure, node operator set and risk profile. Receipt tokens can usually be used in other DeFi apps, but using them adds smart contract and market risk on top of staking risk.

Before staking, read the protocol's documentation on fees, withdrawal mechanics, slashing rules and where the validators are run. Yields change with network conditions, so check the live figures on the protocol site or a reputable data dashboard rather than trusting a fixed number in any article.

Scalability news is easy to misread, so a few habits help. Start with the primary source. Protocol research posts, client release notes and official upgrade documents describe what changed. Secondary articles often blend live features with plans, so check the status of any feature before acting on it.

For costs, a simple routine works well:

• Compare the same action on two or three networks using a live fee estimate.
• Include bridge fees, withdrawal waiting times and the cost of the exit trip.
• Check whether the app you need is officially deployed on the network you chose.
• Look at the protocol's total value locked and activity over time using an analytics site such as DefiLlama, keeping in mind that these figures change daily.
• Record your transactions so you can review costs later.

As a reader, you should also be sceptical of any claim of guaranteed returns, fixed fees or a particular transaction-per-second figure presented without a source. Networks change, and what is true in one month may not hold in the next. Use the principles in this guide as a checklist, and verify the specifics before you commit funds.

You can lower costs without taking unnecessary risk by changing how and when you transact. Start with timing: fee levels often fall during quieter periods, which can be checked on a gas tracker. Set your wallet's fee settings carefully, because an overly high tip wastes money and an overly low one can strand a transaction.

Then consider the network. Doing routine activity on a major rollup can cut costs substantially, provided you understand its withdrawal rules. Batch actions where the app supports it, and avoid repeated small transfers that each pay a full fee. Revoke token approvals you no longer need, since leaving broad approvals open adds risk and sometimes gas for later cleanup.

Finally, keep your security practices consistent. Store recovery phrases offline, never share them, check every signature request, and bookmark official app addresses rather than following links from messages. Cheaper transactions are only a saving if the funds arrive safely and remain under your control. Gambling-style apps and high-yield products carry added risk, and you should check local laws and only use services you are legally permitted to access.