After Dencun in March 2024, Pectra in May 2025 and Fusaka in December 2025, Glamsterdam is the next scheduled hard fork. Unlike its predecessors it is aimed not primarily at Layer 2 rollups but at the base layer itself.
The name
Glamsterdam is a portmanteau of two upgrade names. Amsterdam denotes the execution-layer component, named after a previous Devconnect location. Gloas denotes the consensus-layer component and is a star name. Ethereum traditionally names hard forks by pairing a star name with the city that hosted the most recent Devconnect developer conference.
The two core changes
Two proposals are treated as headliners.
EIP-7732, Enshrined Proposer-Builder Separation (ePBS). Until now the separation between those who build blocks and those who propose them has run through external relays and MEV-Boost setups outside the protocol. ePBS moves that relationship, including commitments and payments, directly into the protocol. This removes a trust assumption that currently sits outside the consensus rules.
EIP-7928, Block-Level Access Lists (BALs). Blocks will have to declare their data dependencies upfront. Only then can the network process multiple transactions simultaneously rather than strictly one after another. BALs are the precondition for genuine parallel execution.
Alongside the two headliners, further proposals are being tested in the devnets, including EIP-7610, EIP-7688, EIP-7778, EIP-7843, EIP-7976, EIP-7981, EIP-8024, EIP-8246 and EIP-8282.
What this is meant to achieve
The practical purpose is capacity. According to ethereum.org, BALs and ePBS together make it safe to raise the block gas limit from 60 million today toward roughly 200 million. That leaves more L1 capacity for applications.
For a sense of the pace: the limit sat at 30 million for four years, then rose within months to 36 million and on to 60 million.
What matters equally is what Glamsterdam is not. It is not a turn away from the rollup strategy. BALs are transparent to L2 sequencers; no sequencer-side changes are required.
The timeline — where the sources diverge
This is where most Glamsterdam coverage becomes imprecise, so it is worth stating explicitly.
The official roadmap on ethereum.org lists Glamsterdam for Q4 2026 with the status "in development". Elsewhere ethereum.org cites the second half of 2026.
Other sources give earlier dates: one staking provider reported end of August 2026 as an internal working target, expressly described as aspirational and subject to devnet readiness and cross-client testing. Older pieces from spring 2026 mentioned May or June.
For classification the primary source governs: the figure on ethereum.org — Q4 2026 — is authoritative. Earlier mentions come from interim stages of developer discussion and are either superseded or were never binding.
Ethereum has a history of delayed upgrades. The Merge was pushed back repeatedly over several years. The on-time delivery of Pectra and Fusaka suggests an improved cadence, but Glamsterdam's scope is larger, and the interaction between ePBS and BALs has not been tested in this form at mainnet scale.
Glamsterdam is followed by Hegotá, currently listed on the official roadmap for 2027.
What investors need to do
For plain ETH holders: nothing. A hard fork of this kind requires no action from holders, no migration and no token swap. Anyone asked to "migrate" tokens or "verify" a wallet is dealing with an attempted fraud. Such campaigns cluster around every announced upgrade.
For stakers and node operators: both the consensus-layer and execution-layer clients must be updated before mainnet activation. Anyone staking through a provider should await that provider's announcement; anyone running their own setup should track client release cycles.
On costs: the L1 fee regime has already changed substantially ahead of Glamsterdam. According to ethereum.org, Etherscan's gas tracker showed standard gas around 0.15 gwei on 5 May 2026, with daily averages near 0.5 gwei through April. A basic ETH transfer cost under a cent at that level. Anyone whose mental model of Ethereum fees formed between 2021 and 2023 is working from outdated assumptions.
What can go wrong
Two risks are named openly in the developer discussion.
First, scope creep: more than 25 non-headliner EIPs were under consideration for inclusion. One engineering team publicly warned that additional inclusions would jeopardise the timeline.
Second, coordination: a hard fork of this size requires synchronised updates from validators, exchanges and infrastructure providers. Without that alignment, the effect is delayed regardless of whether the code is finished.
This article is for information purposes and does not constitute investment advice.