# Glamsterdam: Ethereum to test 200 million gas blocks on October 6

> Ethereum activates Glamsterdam on Sepolia on October 6 with ePBS and 200 million gas blocks. What changes and who it affects
> By José Maldonado · Ethereum · 2026-10-02

Some tests decide the future of a network, and the one on October 6 is one of them. That day, at 13:53:36 UTC, the Ethereum Foundation will activate the Glamsterdam upgrade on Sepolia, its public test network, with an unprecedented experiment: coordinating validators toward 200 million gas blocks, more than triple the current capacity, according to the [Foundation's official announcement](https://blog.ethereum.org/2026/09/17/glamsterdam-testnet-announcement).

If you use Ethereum, think of Sepolia as the closed circuit where engineers test the car before taking it onto the highway. Nothing that happens on October 6 will move your ether or change the fees you pay today. But the mainnet schedule, where your money actually lives, depends on that test, so it is worth understanding what is being tested, who has homework and what can go wrong.

The key is that Glamsterdam is not a simple tweak of the nuts and bolts. The package builds into the protocol the separation between those who propose blocks and those who build them, adds per-block access lists and redesigns gas pricing so it resembles the real cost of using the network. Three fundamental changes in a single test, with the mainnet watching from the sidelines and no confirmed date.

**[MetaMask Staking withdraws its validators from Lido due to a security flaw](https://observatorioblockchain.com/en/cybersecurity/metamask-retira-validadores-de-lido-por-un-incidente-de-seguridad/)**

## What gets activated on October 6

The appointment has surgically precise coordinates. Activation will occur at epoch 353,024 and block 11,296,768 of Sepolia, on Tuesday, October 6 at 13:53:36 UTC, and will affect only that test network. Neither Hoodi, the validator-focused network, nor the mainnet has a date yet, a nuance worth repeating because every upgrade stirs the same misunderstanding.

Anyone running a node on Sepolia does have pending work. Each operator must update both of their programs, the execution client and the consensus client, before Tuesday to versions that explicitly declare support for the fork: **Geth 1.17.6, Erigon 3.7.0, Nethermind 2.0.0 or Reth 2.7.0 on execution, and Prysm 7.2.0 or Teku 26.9.1 on consensus, among others**. Anyone who arrives late is left out of the test network, with no drama for their funds, but out of the test.

After all, the most interesting nuance is that the 200 million are not imposed. Proposal EIP-8261 introduces a preference schedule that coordinates clients, but each validator keeps its own judgment and blocks remain valid whichever option they choose. Prysm and Teku will keep 60 million by default after activation, so only operators that expressly configure 200 million will propose large blocks. Sepolia will measure, block by block, how many sign up for the experiment.

## Three changes in a single package

The first change has an acronym for a name, ePBS, and a powerful idea behind it. Today, the split between who builds a block and who proposes it happens outside the protocol, in the hands of trusted external intermediaries. Glamsterdam brings that separation inside Ethereum's own code, so the network takes on a task it currently delegates to third parties, as explained by the [merca2 analysis](https://www.merca2.es/2026/09/29/glamsterdam-ethereum-sepolia-epbs-testnet-2464393/). Less middleware, fewer blind trust points.

The second pillar is block-level access lists, EIP-7928. Each block will record the accounts and storage slots it has touched during execution, and with that inventory nodes will know which operations do not clash with each other so they can order them better. The practical result is more efficient use of available computing capacity, something like moving from a single checkout line to several open registers at the supermarket.

### New gas pricing

The new gas pricing completes the trio. Proposals EIP-8037 and EIP-8038 raise and separately measure the cost of creating state and of accessing it, so that the price paid resembles the real cost of storing and reading data on the network. The Foundation warns of the side effect: smart contracts that hardcode gas could need changes, because an assumption that holds today may stop holding. Teams have months ahead, but the warning is explicit and should be taken seriously.

Even so, the package has its flip side and it is worth facing squarely. Bringing block building into the protocol reduces dependence on external intermediaries, but it can concentrate that function in a handful of highly specialized actors with the capacity to invest in infrastructure the rest cannot afford. The decentralization gained on one side can be lost on the other, the same tension that has accompanied Ethereum since professional validators replaced home miners.

**[Vitalik Buterin explains why Ethereum will be much more than a blockchain](https://observatorioblockchain.com/en/ethereum/ethereum-mas-que-blockchain-vitalik-buterin/)**

## Who has homework and who can keep sleeping?

The good news is for most people. If you hold ether on an exchange, in your wallet or in delegated staking, you have absolutely nothing to do: no moving funds, no claiming anything, no pressing any button before Tuesday. The test takes place on a parallel network with no economic value and your money will not even notice, a point scammers exploit with every upgrade to slip in fake airdrops and urgent migrations that do not exist.

Sepolia node operators, by contrast, are working against the clock. Updating both clients, reviewing the new validator duties that come with integrated separation and adapting block-building tools are part of their homework this week. The Foundation has also opened Glamsterdam's specifications to its bug bounty program, inviting hunters to find flaws with an economic incentive before the code touches the real network.

It is worth noting that application developers carry the quietest part of the work. Testing contracts under the new gas rules, reviewing hardcoded assumptions and simulating behavior with per-block access lists is lab work that does not make headlines but decides whether or not the mainnet will suffer on launch day. In this way, October 6 evaluates both the code and those who use it.

## From gas to the user's pocket

The question everyone asks is whether this will lower fees, and the honest answer has nuances. A 200 million limit does not speed up blocks or magically make each operation cheaper: it creates more aggregate room for more transactions when demand is high, which eases congestion and pressure on fees at peak usage. It is a wider highway, not a faster car.

The road to here helps calibrate the ambition. The gas limit had been frozen at 30 million since the arrival of proof of stake, rose toward 36 million in February 2025, then climbed to 45 million and the Fusaka versions adopted 60 million by default. Glamsterdam more than triples that bar in the test environment, although Fusaka already safeguarded the distribution with a cap of 16.7 million per individual transaction so that no single operation hogs the block.

## And after Sepolia, what comes next?

If the test goes well, the next testing ground will be Hoodi, though without a formal date. Developer notes mention October 27 as a tentative target for that validator network, with the final decision pending observation of Sepolia, according to [DiarioBitcoin](https://www.diariobitcoin.com/blockchain/ethereum-apunta-a-probar-glamsterdam-en-sepolia-el-6-de-octubre-sujeto-a-las-pruebas/). Tentative means exactly that: an intention, not a commitment, and treating it as a firm date would be the first misreading.

The mainnet, where the ETH that trades operates, still has no schedule. The official roadmap places Glamsterdam in the fourth quarter of 2026 with the date expressly unconfirmed, and experience teaches that weeks or months of analysis lie between Sepolia and mainnet. Anyone promising you a mainnet activation date today, October 1, is selling smoke with a technical letterhead.

The long view points even further ahead. Developers place Hegotá, the upgrade that will follow Glamsterdam, in 2027, and Vitalik Buterin suggested in September that it could be Ethereum's last conventional fork before the era of recursive STARKs, automated formal verification and quantum-resistant cryptography. Sepolia is, in that story, the penultimate test of the Ethereum we know.

Upgrade

Date

What it brought

The Merge

Sep 2022

Goodbye mining, hello proof of stake

Shanghai

Apr 2023

Staking withdrawals unlocked

Dencun

Mar 2024

Blobs and minimum fees on rollups

Fusaka

2026

60M gas cap and per-transaction limit

Glamsterdam

Sepolia Oct 6 2026

ePBS, access lists and new gas

Hegota

2027 planned

FOCIL and frame transactions, perhaps the last classic one

## The test that allows no rush

Glamsterdam condenses the philosophy that made Ethereum great: test in public, measure without self-deception and do not promise dates until the code deserves them. The 200 million gas is an ambition, not an order; the mainnet, a consequence, not a premiere. In between are validators voting with their blocks, developers reviewing contracts and a Foundation that prefers a boring test to a heroic launch.

We will have to wait until Tuesday to know whether Sepolia swallows the whole package or spits out the first bite. If something breaks in the closed circuit, it gets fixed without any user losing a euro, which is exactly what it is for. And if everything works, the mainnet will have taken a silent step toward blocks built without trusted intermediaries, nodes that order their work better and gas that costs what it costs.
