Ethereum’s next major protocol upgrade has a concrete public-network milestone.
The Ethereum roadmap now lists October 6, 2026 for the Glamsterdam fork on Sepolia. Mainnet is still targeted for the fourth quarter, but no mainnet date has been confirmed.
That turns Glamsterdam from a bundle of accepted engineering proposals into something operators can start preparing to test against a real public testnet schedule.
Optimisus previously covered the architecture of the upgrade when the key proposals were being assembled. The new milestone matters because Glamsterdam changes several of Ethereum’s most basic assumptions at once: who builds a block, how nodes understand transaction dependencies, how state growth is priced and how quickly validators can exit.
Block Building Moves Into the Protocol
The headline change remains enshrined proposer-builder separation, or ePBS, specified in EIP-7732.
Today, much of Ethereum’s block-building market relies on off-protocol software and trusted relays between block builders and validators. Glamsterdam moves the core handoff into Ethereum itself, including trustless builder payments and separate timing for the consensus block and execution payload.
Ethereum.org says the design expands the effective propagation window for payload data from about two seconds to roughly nine seconds. That is not just a cleaner MEV architecture. More propagation time gives the network room to handle larger payloads without forcing every validator into the same narrow timing bottleneck.
Optimisus’s earlier deep dive, Ethereum Is Taking Block Building Back From the Relays and Learning to Run Transactions Side by Side, explains why ePBS is a scaling change as much as a trust change.
Ethereum Is Preparing to Stop Treating Every Transaction as Sequential
The second major piece is Block-Level Access Lists, or EIP-7928.
Ethereum currently discovers much of a transaction’s state access by executing it. That makes safe parallel processing difficult because a node cannot know in advance whether two transactions will touch the same account or storage.
Block-Level Access Lists provide an explicit map of state accesses and post-execution values for the block. The network can identify non-overlapping work, perform parallel disk reads and eventually process more work concurrently rather than treating execution as one long queue.
The companion networking change, EIP-8159, gives execution clients a way to exchange those lists directly. Together, the proposals also enable forms of executionless state updating, reducing how much historical work a syncing node must replay.
Scaling Now Includes Charging Correctly for State
Higher throughput only helps if the database underneath Ethereum can survive it.
Glamsterdam therefore includes changes that raise and restructure the cost of creating and accessing permanent state. Ethereum.org says EIP-8037 introduces a state-cost model targeting roughly 120 GiB of annual database growth, while EIP-8038 reprices state-access operations that have become too cheap relative to the hardware work they require.
This is the less glamorous half of scaling. A network can raise gas limits quickly and still become harder to run if every block permanently expands storage requirements. Glamsterdam is trying to increase capacity without quietly pushing ordinary node operators out of the network.
Validator Exits Get Faster Too
EIP-8061 changes validator churn rules so exits can scale with the total amount of ETH staked rather than remaining constrained by the same fixed structure.
At current staking levels, ethereum.org estimates exit capacity would rise by roughly four times and consolidation capacity by roughly two times. That matters in a network where one in three ETH has recently been staked and exit queues can become part of the asset’s liquidity profile.
Optimisus has already tracked the economic side of that problem in its analysis of record Ethereum staking participation and falling validator rewards. Glamsterdam attacks a different part of the same system: how quickly stake can move.
The Testnet Date Is the Point Where Operators Have to Care
For normal ETH holders, the October 6 Sepolia fork requires no token conversion. Ethereum.org explicitly warns that users do not need to ‘upgrade’ their ETH, and anyone asking them to migrate coins because of Glamsterdam should be treated as a scam.
Node operators and validators have a different responsibility. Both execution and consensus clients must support the fork, and operators will need compatible releases before participating on upgraded networks.
The mainnet date remains unconfirmed, so October 6 should not be treated as a countdown to a guaranteed production fork. It is a test milestone.
But it is an important one. Glamsterdam is attempting to change Ethereum’s block market, prepare execution for parallelism, make state growth economically sustainable, speed validator exits and lower the cost of simple ETH transfers in one release.
The upgrade has spent months being described as Ethereum’s next scaling architecture. Sepolia is where that architecture begins proving it can run.
This is not financial advice.
Sources
- Ethereum.org — Glamsterdam roadmap — Primary upgrade status, Sepolia date, scheduled EIPs and implementation explanations.
- EIP-7732 — Enshrined Proposer-Builder Separation — Technical specification for ePBS.
- EIP-7928 — Block-Level Access Lists — Technical specification for block-level state access lists.
- EIP-8061 — Increase exit and consolidation churn — Technical specification for validator exit and consolidation capacity.
