Ethereum’s next major upgrade, Glamsterdam, has completed its first formal test after going live on the Sepolia testnet on Oct. 6. Roughly 11 hours after the fork, Sepolia’s gas limit moved from 60 million to the 200 million target set for the upgrade. Mainnet is still aimed at the fourth quarter of 2026, but the more immediate checkpoint is whether the Hoodi testnet can activate on or around Oct. 27.
The rollout has already seen delays. Glamsterdam had originally been planned for the first half of 2026, and the Sepolia fork itself was pushed back from Aug. 3 to Oct. 6.
Sepolia delivers the first live test
Glamsterdam combines the consensus-layer changes known as Gloas with the execution-layer changes called Amsterdam. The upgrade went live on Sepolia on Oct. 6, and the chain continued finalizing blocks normally through the transition without major failures. That made Sepolia the first formal live test of the package.
The first Glamsterdam block on Sepolia was produced with a 60 million gas limit, matching Ethereum mainnet’s current limit. From there, the testnet climbed to 200 million gas in about 11 hours after the fork and has stayed at that level.
At the design level, Glamsterdam targets a gas limit of about 200 million. Against the current 60 million level, that works out to an increase of roughly 3.3x. For comparison, Ethereum mainnet stayed at a 30 million gas limit for years before validators raised it to 60 million in three steps in 2025. Even the 2021 London upgrade, which doubled maximum block capacity, kept the target at 15 million.
The fork does not instantly triple Ethereum’s capacity the moment it activates. What it changes is the way blocks are produced, making much larger blocks safe to produce. Validators still decide whether to raise the limit.
Researchers call it a major capacity shift
Ethlabs researcher Barnabé Monnot said after the Sepolia activation: “A big deal. Shipping this means at least a 3x increase in L1 scale, a 2.5x increase in contract size limits, and a lot of really useful new features.”
Ethereum Foundation researcher Toni Wahrstätter made a similar point, calling Glamsterdam Ethereum’s “largest throughput increase since launch.”
The three proposals doing most of the work
The fork includes 18 Ethereum Improvement Proposals in total, but three of them carry most of the weight.
ePBS gives validators more time
The first is enshrined proposer-builder separation, or ePBS, under EIP-7732.
Today, most Ethereum blocks are built by specialized builders and then passed to validators through trusted intermediaries known as relays. Under EIP-7732, proposers first commit to signed bids from builders, and the full block is revealed later.
That changes the timing pressure around block inspection. Instead of forcing transaction checks into an attestation window of about four seconds, the proposal would give most validators about nine seconds.
Block-level access lists open the door to parallel work
The second major piece is block-level access lists, or BALs, under EIP-7928.
Most Ethereum clients currently process transactions in a block one by one because they do not know in advance which parts of state each transaction will touch. EIP-7928 adds a block-level map that lists all touched state and the resulting values after execution.
With that information, clients can prefetch data and verify transactions in parallel. The report compares the setup to factory operations: ePBS gives quality control more time, while BALs make sure the materials are lined up in advance so multiple assembly lines can run at once.
Larger contracts and higher costs for new state
The third cluster of changes touches contract size and gas pricing.
EIP-7954 raises the contract size cap from 24 KiB to 64 KiB, easing a long-standing limit that has forced larger applications to split code across multiple contracts.
At the same time, the fork reprices new state. Under EIP-8037, the cost of creating a new storage slot rises from 20,000 gas to 97,920 gas. By contrast, under EIP-2780, a standard ETH transfer to an existing account remains at 21,000 gas.
That means activities that create large amounts of new state, including airdrops, mints, and new wallet-related actions, become materially more expensive. Routine transfers do not see the same change.
Early Sepolia data shows faster processing for comparable blocks
Wahrstätter later shared early data suggesting the changes are already having an effect. Compared with pre-fork conditions, each client was processing gas faster. In one example, Geth needed about 115 milliseconds to execute a 40 million gas block before the fork and about 30 milliseconds after it.
The report notes that this is still early data, covering one node per client and less than a day of observations.
There is also context behind the 11-hour climb to 200 million gas. A proposer can only adjust the gas limit by about 0.1% per block. If every proposer had increased it every time, Sepolia could have reached the target in about four hours. The actual move took more than twice as long. Part of the reason, according to the report, may be that two major consensus clients, Prysm and Teku, kept validators at a 60 million limit after the fork unless operators explicitly chose 200 million.
Hoodi is the next gate before mainnet
After Sepolia, the next stop for Glamsterdam is Ethereum’s Hoodi testnet. The proposed activation date is Oct. 27, but that date has not been finalized. Mainnet remains targeted for some point in the fourth quarter of 2026.
Given the delays already seen, the timetable could still change again. That is why Hoodi’s ability to activate on schedule has become the next practical metric to watch.
Tools will need updates, and some onchain actions get pricier
Beyond throughput, Glamsterdam will affect infrastructure and user activity.
The Ethereum Foundation has warned that wallets, indexers, and gas estimation tools that assume a fixed maximum gas limit must update their logic to avoid problems.
The repricing of state growth has a separate effect on user behavior. Onchain actions that create many new accounts or storage entries, such as airdrops, mints, and new wallet creation, will cost more. Standard transfers will not be repriced in the same way.
The main shift is a roomier Ethereum L1
The broadest change in the report is straightforward: Ethereum’s L1 is being prepared to handle much larger bursts of activity.
As one example, the article notes that Fake World Assets briefly became the biggest consumer of gas on mainnet during its most active day in August. Under Glamsterdam, the report argues that traffic spikes like that would have much more room to play out without pushing fees higher and crowding everyone else out as quickly.
For now, the next real test is Hoodi. Whether that testnet activates around Oct. 27 will shape the path toward a mainnet launch still penciled in for late 2026.