Ethereum developers activated the Glamsterdam network upgrade on the Sepolia testnet on Tuesday at 13:53 UTC, epoch 353024, according to the Ethereum Foundation’s blog and live chain data. The fork is one of the final public rehearsals before a mainnet launch, and it carries the boldest capacity experiment yet: Sepolia’s gas limit jumped from roughly 60 million to 200 million, more than tripling how much work each block can hold.
What Glamsterdam changes
The upgrade has two headliner features. Enshrined proposer-builder separation, or ePBS, moves the handoff between block proposers and block builders into the protocol itself. In the Foundation’s words, “A proposer includes a builder’s commitment to an execution payload, and the builder subsequently reveals that payload.” Separating the two roles in code, rather than in a sidecar agreement such as the MEV-boost relay used today, widens the window validators get to verify blocks from about 2 seconds to roughly 9 seconds. A payload timeliness committee also attests whether the builder revealed the payload and whether its blob data arrived on time, so validators have a fuller picture before attesting.
The second change is block-level access lists, known as BALs, under EIP-7928. Each block now records which accounts and storage locations it touches, along with post-transaction state changes. The Foundation said the lists let clients “read state from disk in parallel, validate transactions in parallel, and compute state roots more efficiently.” The upgrade also reprices gas so that costs better reflect the load execution and state growth put on node operators, including separate metering for creating new state under EIP-8037 and updated state-access costs under EIP-8038. Together, the two changes are meant to raise capacity while keeping the hardware requirements of running a full node manageable, which the Foundation has framed as the main constraint on how far the gas limit can eventually go.
A last-minute validator fix
Hours before the fork, the Prysm validator client shipped an update to protect the experiment itself. Prysm 7.2.0 supported the Sepolia fork but defaulted validators to a 60 million gas limit after activation, and its release notes said the command-line flag –suggested-gas-limit would have no effect after Gloas, the consensus half of the upgrade. Validators who wanted bigger blocks had to configure it separately through version-2 proposer settings or the keymanager API. The fix, version 7.2.1, released late Monday, makes validators on Sepolia propose 200 million gas blocks automatically from the fork, removing the manual step that could have skewed the test.
“Without the update or a manual configuration change, some validators would have continued producing blocks at Sepolia’s previous 60 million gas limit, undermining the capacity test,” CoinDesk reported in its coverage of the fix. The point of the exercise is to show whether clients can reliably handle blocks more than three times larger before developers decide how much capacity mainnet should actually take.
Why 200 million gas matters
A gas limit is the ceiling on how much computation a single block can contain. Raising it raises the number of transactions per block, but it also raises the bandwidth, disk and CPU costs of running a full node. Ethereum’s current mainnet limit sits near 60 million, and the Foundation set 200 million as the floor for Glamsterdam testing. Devnet 11, a private development network, already ran the transition while raising the gas limit from 60 million to 200 million, using roughly 84,000 validators across several Ethereum clients before Sepolia went live.
Validators on Sepolia can now test client behavior at block sizes closer to what a scaled mainnet might see. How the fork behaves over coming days will shape the Hoodi testnet activation, which the Foundation’s EIP-7773 tracking table lists as TBD, and the eventual mainnet date, also not set. Community contributor Pooja Ranjan reported the fork went live “with 18 Core + 7 Other EIPs.” Activation finalized at 14:24 UTC, and the focus now shifts to whether validators can keep the new, heavier blocks stable in practice over the coming weeks before the next testnetz stage begins.
Where the fork sits in Ethereum’s timeline
Glamsterdam, the name combines two previously separate upgrade labels, Gloas for consensus and Amsterdam for execution, and it follows last year’s Fusaka hard fork. Sepolia is the second of three public testnets to activate, with Hoodi expected around October 26-27 per community reporting and mainnet currently targeted for December 2, 2026, though the Foundation has not confirmed a date. Decisions on the final gas limit configuration land in the last weeks before Hoodi, and the Foundation’s official blog has been explicit that the Sepolia announcement covers Sepolia only. Once Glamsterdam reaches mainnet, developers plan to begin work on the next upgrade, Hegotá, whose scoping has already started, with 66 proposals from various teams under review since August, several of which focus on bringing more privacy capabilities into the protocol itself.
| Network | Status | Gas limit change |
|---|---|---|
| Devnet 11 | Completed, about 84,000 validators | 60M to 200M |
| Sepolia | Activated Oct. 6, 13:53 UTC | 60M to 200M |
| Hoodi | Scheduled, date not set | Expected to match |
| Mainnet | Targeted around December, no date confirmed | Undecided |
One more consequence is worth watching. Gas repricing will change assumptions some applications have relied on for years. The Foundation’s announcement flags the separate metering of state creation in particular, and the testnet guidance asks application developers to review EIP-7773 and re-test gas estimation, logs, opcodes and client interfaces after the fork. Contracts that assumed old, cheap state writes could see their running costs rise once mainnet takes the new rules, and some projects may need to re-audit their gas paths before the upgrade reaches production networks running real funds rather than testnet tokens of no value.
Node operators on mainnet have no action to take yet, and the Foundation’s rollout guidance makes that explicit: a separate announcement will cover Glamsterdam’s activation on mainnet once client teams decide on a date. For now the visible effects are on Sepolia only, where validators, block builders on testnets and client developers are the ones watching block sizes and payload delivery in real time.