News

Ethereum's upcoming Glamsterdam upgrade alters gas charges

Glamsterdam's proposed gas changes separate state creation costs. Wallets and contracts with fixed gas assumptions may need updates before activation.

Category byline: Megan Sutherland Updated October 7, 2026 5 min read

This byline is assigned by coverage category. It does not confirm personal authorship or review. Editorial policy.

Ethereum's upcoming Glamsterdam upgrade alters gas charges
Illustrative photograph. Photo by Pok Rie. Pexels. Pexels License.

Ethereum’s planned Glamsterdam upgrade changes how the network measures work done by transactions. Wallets, contracts and infrastructure may need to adjust their gas assumptions before activation. The changes concern gas units, rather than a guaranteed transaction price in dollars.

One familiar number needs particular care: 21,000 gas for a simple ETH transfer. The proposed rules distinguish between updating an existing account and creating a new one. Treating every transfer as the same operation can produce an inadequate estimate.

A second gas dimension

EIP-8037 proposes separate measurement of execution gas and state gas. Execution gas covers transaction processing. State gas measures the creation of new account records, contract storage or deployed code.

Users would pay for both kinds of work. The separate accounting is intended to price state growth more consistently as network capacity increases. It is not a second token or a separate fee currency.

The related EIP-2780 specification retains 21,000 execution gas for a plain transfer to an existing account without code or delegation. A transfer creating a new account also incurs state gas. Transfers involving contract code or delegated accounts can require additional work.

That distinction matters because a newly generated wallet address is not necessarily an existing account in Ethereum’s state. Generating an address alone does not add a record to the network.

Why creating state costs more

Ethereum’s state records account balances, account counters called nonces, contract code and contract storage. Full nodes need this information to check transactions and calculate the resulting state. Adding new records increases the data they must maintain.

Updating an existing balance and creating a new account therefore place different demands on the system. Contract deployment and new storage entries can also increase the database. EIP-8037 aims to make their charges better reflect the amount of state created.

That does not mean Ethereum keeps every address ever touched forever. EIP-161 already provides rules for removing touched empty accounts. The scaling concern is sustained growth in the state nodes must maintain, rather than an absolute ban on deletion.

Gas is a unit used to measure transaction work. An opcode is an individual instruction run by Ethereum’s execution engine. Repricing an opcode changes its gas charge; it does not directly set a fixed dollar price.

As Ethereum’s gas documentation explains, the ETH fee depends on gas used and the effective price per unit. The dollar value also depends on ETH’s market price. A protocol change alone cannot guarantee that everyday transfers become cheaper.

What the developer warning actually covers

The Foundation’s August 24 repricing impact guide identifies risks from fixed gas assumptions. It discusses EIP-8037 alongside EIP-8038, which updates charges for accessing state.

Testing replayed historical mainnet transactions under the proposed schedule. The large majority had identical outcomes and execution. Some succeeded with changed gas usage, while others needed a higher gas limit. A smaller group still failed when the limit was substantially increased.

The distinction rules out a blanket claim that every wallet or contract will break. Fixed limits, fixed gas amounts inside contract calls and logic based on remaining gas can require review. Simply raising the outer transaction limit will not fix every affected contract.

Wallets and gas estimation services also need to account for the new rules. A cached constant can underestimate the gas a transaction requires. An insufficient limit can cause failure, even when the user is willing to pay the market gas price.

How Ethereum tests an upgrade

Developers first use development networks, or devnets, to test client implementations together. These are testing environments rather than evidence that the same rules are active on mainnet. Public testnets give application and infrastructure teams another environment for integration checks.

Sepolia and Hoodi are established Ethereum testnets. Their test ETH is used for testing rather than representing mainnet funds. Mainnet activation requires its own coordinated upgrade and compatible client releases.

Testing can uncover problems that change the schedule or implementation. A published target date is therefore different from confirmation that an upgrade has activated successfully.

What the rollout announcements establish

The August impact guide identifies Platåberget as glam-devnet-8 and directs developers there to test the new schedule. That is a development network used for Glamsterdam testing, distinct from the established Sepolia and Hoodi testnets.

The Foundation’s September testnet announcement scheduled Sepolia activation for October 6, 2026. In that announcement, Hoodi and mainnet activation dates were still undecided. The scheduled Sepolia date alone does not confirm that activation occurred or establish a mainnet deadline.

The proposals and release announcements are the relevant references for teams preparing integrations. They describe the rules and client requirements more precisely than an assumed rollout calendar.

Other changes planned for Glamsterdam

Enshrined proposer builder separation, or ePBS, brings the builder and proposer exchange into the protocol. Builders assemble transaction payloads, while proposers commit to a selected payload. Protocol rules handle the exchange and payment, reducing reliance on trusted intermediaries.

The proposal also separates the timing of consensus checks from execution checks. That gives validators more time to check execution payloads. It does not establish that builder concentration, transaction ordering concerns or manipulation disappear.

Block level access lists record the accounts and storage locations used during a block. Knowing those dependencies helps clients read data and validate work in parallel. Parallel processing means doing compatible work at the same time, while respecting dependencies between transactions.

That can support greater capacity, but the lists alone do not guarantee a particular increase in throughput. Client implementation and the rest of the upgrade still matter.

EIP-7954 proposes higher limits for deployed contract code and initialization code. Initialization code is the code used while creating a contract. A larger permitted size does not mean deployment has become free or inexpensive.

What it means for everyday users

For routine users, updated wallets and infrastructure are expected to handle the accounting changes. The Foundation’s impact guide says regular users do not need to take special action. That guidance is not a promise of identical fees for every transaction.

Sending ETH to an account that does not yet exist is a particularly relevant case for gas estimation. Exchanges, bridges and other services sending funds to fresh addresses need to test that path. A fresh address is not itself a fault; the estimator must account for the state the transfer creates.

The practical aim is accurate gas estimation under the rules that actually activate. Teams should test affected integrations before relying on old constants in production. Planned changes still need to be distinguished from mainnet rules already in force.

Sources