Ethereum’s next major upgrade, Glamsterdam, went live on the Sepolia testnet on Oct. 6, with the chain completing the transition and continuing to produce blocks without major faults.
Glamsterdam combines Gloas on the consensus layer and Amsterdam on the execution layer. Ethlabs researcher Barnabé Monnot called the milestone significant, saying, “This is huge! Shipping this means at least a 3x increase in L1 scale, 2.5x larger smart contracts, and more useful features.” Ethereum Foundation researcher Toni Wahrstätter also wrote on X that Glamsterdam opens the door to Ethereum’s “largest throughput increase since launch.”
A 200 million gas target
By the numbers, Glamsterdam aims to raise Ethereum’s gas limit to 200 million. That is about 3.3x the current 60 million level on mainnet. For years, Ethereum mainnet stayed at a 30 million gas limit, only reaching 60 million in 2025 after three separate increases. Even the 2021 London upgrade, which doubled maximum block capacity, had a target gas limit of just 15 million.
The fork does not expand Ethereum by 3x in one step. What it does first is change the block construction process so very large blocks can be generated safely. After that, validators can decide whether to move the gas limit higher.
Three changes sit at the core of the fork
The upgrade package includes seven EIPs in total, with three doing most of the heavy lifting.
The first is enshrined proposer-builder separation, or ePBS, under EIP-7732. Today, most Ethereum blocks are assembled by professional block builders and then passed to validators through trusted intermediaries known as relays. Under ePBS, proposers first commit to signed bids from builders, and builders reveal the full block afterward.
That changes the timing for validators. Instead of having to rush through block transaction verification inside a proof window of roughly four seconds, most validators would get about nine seconds under the EIP design.
The second major change is Block-level Access Lists, or BAL, under EIP-7928. Right now, Ethereum clients need to execute transactions one by one because they cannot know in advance which state each transaction will read or modify. BAL requires each block to include a mapping of all state the block will touch, along with the post-execution values. With that list in hand, clients can prefetch data and verify multiple transactions in parallel.
The third key change is support for larger contracts alongside gas repricing. EIP-7954 raises the maximum contract size from 24KB to 64KB, addressing a long-running constraint that forced larger applications to split up code.
At the same time, the fork raises the gas cost of scarce resources, meaning newly added storage state. Under EIP-8037, the gas fee for creating a new storage slot rises from 20,000 to 97,920. EIP-2780 keeps ETH transfers to existing accounts unchanged at 21,000 gas.
Sepolia reached the target after about 11 hours
The first Glamsterdam block on Sepolia still carried a 60 million gas limit, matching Ethereum mainnet’s current level. About 11 hours after the fork, Sepolia’s gas limit climbed to the 200 million target and then held there.
Each block proposer can adjust the gas limit by only about 0.1% at a time. In theory, if every proposer raised the limit in every round, the move could finish in four hours. In practice, it took more than twice that long. One reason was that Prysm and Teku, two major consensus clients, kept the default at 60 million after the fork, meaning validators had to opt in before enabling the 200 million ceiling.
Wahrstätter later said the adjustment mechanism was already showing results. In his testing, client processing speed per unit of gas improved sharply after the upgrade. For example, a 40 million gas block that previously took Geth about 115 milliseconds to execute now took about 30 milliseconds.
Those figures, however, came from a single node and less than a day of testing.
Hoodi is next, with mainnet targeted for Q4 2026
With the Sepolia deployment complete, the next step is Glamsterdam on the Hoodi testnet. Activation is tentatively set for Oct. 27, though that date has not been finalized. The current mainnet target is the fourth quarter of 2026.
The schedule has already moved before. Glamsterdam had originally been planned for the first half of 2026, and the Sepolia fork itself was delayed from Aug. 3 to Oct. 6, so the broader timeline could still change again.
Infrastructure updates and higher costs for some onchain actions
The rollout also creates follow-up work for parts of Ethereum’s infrastructure. The Ethereum Foundation said wallets, indexers, and gas estimation tools that still rely on old hard gas limit assumptions need to update or they may run into issues.
With storage repricing, onchain activity that creates many new accounts or adds new storage will become more expensive. The article points to airdrops, NFT minting, and new wallet creation as examples. Ordinary transfers are not affected.
The main shift is more usable L1 block space
The central change is a much larger amount of usable space on Ethereum L1. The article cites Fake World Assets, which at one point during a sharp market rally in August consumed a large share of mainnet gas. After Glamsterdam, bursts of traffic like that would have more room to be absorbed instead of pushing gas fees higher and crowding out other users.
In that framing, Ethereum would gain more room for more ideas, more projects, and more onchain activity. How the community uses that added capacity comes next.
