Ethereum developers have scheduled the Glamsterdam upgrade to activate on the Sepolia testnet on Oct. 6, a milestone on the road to the network’s next mainnet hard fork, according to the Ethereum Foundation. Node operators must update both execution-layer and consensus-layer clients before the activation date, and client teams set Sept. 29 as the deadline for compliant software releases.
What Glamsterdam changes
The upgrade’s headline feature is enshrined proposer-builder separation, a change that moves the handoff between block builders and validators into the protocol itself. Today that relationship runs through out-of-protocol software called MEV-Boost, which most validators use to sell their block-building rights to specialized builders. Enshrining it removes a layer of trusted middleware and makes the auction for block space a first-class part of the chain, with the censorship-resistance and accountability guarantees that come with protocol-level treatment.
The second major piece is block-level access lists. Each block will carry a record of every account and storage location it touches, which lets execution happen in parallel rather than in the current strict sequence. That is the main speed lever: transactions that do not conflict with each other can be processed at the same time, raising throughput without changing the gas limit. Client teams have described this as the largest execution-layer performance change since the merge of the execution and consensus chains.
Gas pricing also changes. Glamsterdam adjusts the cost model so that creating and accessing data is priced more accurately than before. Contracts that rely on fixed gas stipends or hardcode gas limits may need modifications, and client teams have flagged this as the most likely source of breakage for older smart contracts. Auditing firms have begun publishing compatibility notes for major protocols, and several large DeFi deployments have already queued patches.
The upgrade also raises the block gas limit. In a private rehearsal, the limit reached 200 million gas, a multiple of the current mainnet figure, though the mainnet target at launch remains subject to final testing. A higher limit multiplies the benefit of parallel execution, since more transactions per block means more opportunities for non-conflicting work to run side by side.
Testnet risks developers are watching
Developers have warned that the Sepolia trial could face disruption from so-called fake builders. Because testnet ether is free and builder identities are disposable, attackers can flood the block auction with junk bids, spam the relay network or otherwise interfere with the rehearsal without cost. The concern is not chain failure but noisy data: a messy test makes it harder to judge whether the new code performs as intended, and a false alarm can delay a mainnet date that was otherwise on track.
The scope of the fork is frozen but not immutable. The meta EIP governing Glamsterdam remains in draft, and the ethereum.org roadmap page lists associated proposals including EIP-7688, EIP-7778, EIP-7843, EIP-7976, EIP-8024, EIP-8246 and EIP-8282. Changes before mainnet are possible if testing surfaces problems, and the Foundation’s Forkcast tracker remains the authoritative source for activation epochs and scope changes.
Where it sits on the roadmap
Glamsterdam is the network’s next scheduled hard fork after Fusaka, and it is the fork where the roadmap’s three stated goals converge: faster transaction processing through parallelization, better resilience under load, and a cleaner separation of roles between builders, proposers and users.
After Glamsterdam comes Hegotá, targeted for 2027. Ethereum co-founder Vitalik Buterin said on Sunday that Hegotá could be the network’s last conventional fork before development shifts toward recursive STARK proofs, automated formal verification, optimized consensus and quantum-safe cryptography. In the same remarks he described Ethereum in 2030 as a cryptographic world computer rather than a plain blockchain, blending those techniques into the base layer rather than treating them as add-on layers.
That framing matters for how the market reads the upgrade. Glamsterdam is infrastructure work, not a token-price event, but the parallel-execution changes it brings are the kind of capacity expansion that application developers plan around. Layer 2 rollups in particular benefit from cheaper and more predictable data access at the base layer, since their posting costs fall when data-heavy operations are priced more accurately.
Market reaction and operator checklist
ETH traded near $2,680 on Tuesday, up modestly on the day but down more than 2% over the week, moving with the broader market as bitcoin held near $83,000. Derivatives positioning shows no unusual buildup ahead of the test date, and the upgrade has been priced as a routine development since the scope froze in August.
The practical checklist for operators is concrete: update execution-layer clients, update consensus-layer clients, and verify that deployed contracts do not depend on fixed gas stipends that the new pricing model invalidates. Exchanges and infrastructure providers typically run Sepolia replicas in the weeks before a testnet fork, so the Oct. 6 date functions as a dry run for the mainnet activation that would follow, likely in the first half of 2027 if testing stays on schedule.
After Sepolia, the usual sequence is a second public testnet before mainnet. Dates for those steps have not been announced. For institutional users the more consequential item is the access-list format, which custody providers and block explorers will need to support, and which changes how transaction simulation is displayed in tooling. Those integrations typically lag the fork by weeks, so early mainnet deployments of the new format should be watched for quirks rather than assumed clean.
The Oct. 6 rehearsal also gives researchers their first large-scale look at how the enshrined auction behaves under adversarial conditions, which is exactly the kind of evidence the client teams need before committing to a mainnet epoch. If the test passes without incident, the next date on the calendar will be a second public testnet, and after that the mainnet activation that ends the middleware era for Ethereum block building.