Ethereum Upgrade Glamsterdam: What Stakers Need to Check on Their Client Now

Ethereum is getting its next major upgrade, and for the first time it has a date: on October 6, 2026, Glamsterdam is due to be activated on the Sepolia test network. If you stake yourself, that is the point from which your node needs a software version that does not exist today as a stable release. On September 11, 2026 we counted the published version levels of all ten major Ethereum clients to show how far preparation has actually come.

The short answer first: not one of the ten clients checked carried a stable version on that day with the upgrade in its name. The work is happening in pre-releases and nightly test builds. For you that means nothing needs doing today. What matters more is knowing when the moment arrives.

Glamsterdam, Gloas and Sepolia: What Lies Behind the Names of the Ethereum Upgrade

Three terms turn up in every report on this upgrade, and they mean different things.

Glamsterdam is the collective name for the next change to the rules of the Ethereum protocol. It is assembled from the names of the two layers Ethereum is changing at once: Gloas for the consensus layer and Amsterdam for the execution layer. Anyone reading about Gloas in meeting minutes is therefore reading about the same process, simply the other half of it.

Sepolia is a test network. It runs the same software as the main network, but with worthless tokens. New protocol rules go live there before they touch real money. A bug on Sepolia costs developer time; a bug on the main network costs investors money. Hence the sequence.

A hard fork is a rule change that makes old and new software incompatible: a node that does not know the new rules follows a different chain from the rest of the network from the switchover onwards. That is exactly why a client update before a fork is a precondition for taking part rather than a recommendation.

Why October 6, 2026 Is in the Calendar for Sepolia

The date comes from the 186th consensus layer developer call, held on September 3, 2026. There the client teams settled on epoch 351232 as the switchover point, which corresponds to October 6, 2026 at 13:53 UTC. The Ethereum Foundation's roadmap page has listed that date as the next milestone ever since, and the status of the upgrade as testing on devnets.

For the main network there is no date. The same page says Q4 2026, with the explicit addition that the date is not yet confirmed. A December slot has been discussed in developer circles, and nobody is bound by it. Anyone quoting you a main network date today is quoting an expectation.

The Sepolia date itself is subject to a reservation too. The agenda of the call records that Lido and Optimism had asked for at least one stable day of operation on a test network before a fork is scheduled. That condition was not yet met on September 11.

Devnet-11 on September 14: the Test the Sepolia Date Hangs On

A devnet is a short-lived developer network set up specifically for an upgrade test and discarded afterwards. It is the stage before the public test network.

Glamsterdam-Devnet-10 was originally meant to be the last of these networks before Sepolia. That plan has changed. The developers have scheduled a further network, Devnet-11, due to start on Monday, September 14, 2026. Tests on Devnet-8 continue in parallel, and on Devnet-9 clients are being measured individually.

The position before October 6 can therefore be described precisely: 22 days lie between the planned start of Devnet-11 and the Sepolia date. In that time the client teams have to run a network stably, fix the bugs found and build release versions out of the result. If Devnet-11 does not hold up, the Sepolia date moves. That is precisely what these networks are for.

Client Readiness Checked: Ten Ethereum Clients and Their Versions on September 11

This survey was carried out by cryptoticker.io on September 11, 2026. Method: through GitHub's public interface we retrieved the release lists of the ten open-source Ethereum clients and noted the most recent stable version per client along with its publication date; pre-releases are recorded separately. Ten objects were checked, at a reference time of 18:00 UTC.

What we could not check belongs to the survey as well: we evaluated the version names, publication dates and the pre-release marking, not the full text of every release description. A version can contain Glamsterdam rules without carrying the word in its name. Also not captured are special builds for individual devnets that the teams exchange among themselves without publishing them as releases.

What the Figures Say About the Glamsterdam Timetable

Three findings can be read out of the table.

First: none of the ten stable versions carries Glamsterdam, Gloas or Sepolia in its version name or label. That matches the state of development as the roadmap page describes it.

Second: the release rhythms diverge widely. Four clients built their most recent stable version in September, Erigon as recently as two days ago. Another four are more than 30 days behind, led by Geth at 46 days and Prysm at 44. No verdict on readiness follows from that, though an expectation does: at those teams the release is still to come.

Third: the visible movement is happening in pre-releases. Nethermind published a 2.0.0-rc pre-release on September 2, Lodestar published a pre-release for 1.47.0 on August 27, and Nimbus filed a nightly test build on the morning of September 11. That is where the work is going on right now.

The conclusion for you is unspectacular and useful for that reason: there is nothing to install at the moment. The version you need for the fork does not yet exist as a stable release. Updating in a hurry now gains you nothing.

Execution Client and Consensus Client: Why You Have to Update Both

Since the transition to proof of stake, an Ethereum node consists of two programs that run side by side and talk to each other.

The execution client executes transactions and manages the state of the chain. Geth, Nethermind, Besu, Erigon and Reth belong to this group. The consensus client determines which block counts next and looks after staking. Prysm, Lighthouse, Teku, Nimbus and Lodestar handle that job.

Glamsterdam changes rules on both layers. An update on one side alone is therefore not enough. Anyone who updates the consensus client and forgets the execution client ends up, after the switchover, with a node whose halves disagree. In practice that means missed attestations and, if you are a validator, forgone rewards. How much adds up depends on how long the condition goes unnoticed.

If a service stakes your ETH for you, that service carries the task. Which providers take it on and on what terms is set out in our comparison of staking platforms.

What Solo Stakers Should Check on Their Node Before the Fork

Four points are worth doing already, even with no release available.

Make a note of which versions are running on your machine, separately for both clients. A surprising number of operators do not know this by heart when it counts, and then go looking under time pressure.

Check whether your installation pulls releases automatically or whether you update by hand. For a fork the automatic route is the safer one, as long as it pulls stable releases and not pre-releases.

Look at how much disk space is free. An upgrade that fails on a full disk is an avoidable outage.

Subscribe to the release notifications of both your clients. That is the channel through which the teams will announce the fork date for the main network, and it is faster than any news site.

EIP-8061 and the Exit Queue: What Changes for Stakers

Among the proposals earmarked for Glamsterdam, one bears on staking directly. EIP-8061 raises what is known as the churn, the number of validators allowed to exit the network or be consolidated per unit of time.

Why that counts: anyone pulling their ETH out of staking joins a queue. Its length follows from this ceiling. How long it can become at times is something we described in August using the figures of the day; the details are in our analysis of the staking queue. If the churn rises, the wait shortens at unchanged demand.

Two further proposals shape the substance of the upgrade. EIP-7732 anchors the separation of block proposal and block building in the protocol. EIP-7928 introduces block-level access lists, which tell processing in advance which parts of the state a block touches. Both aim at speed and even treatment in block building, and not at any change to your balances.

The scope of the upgrade counts as frozen, but it can still change before the main network, because the overarching proposal formally sits in draft status. The full list is carried by the Ethereum Foundation's roadmap page.

Staking Through a Provider: What You Can Check Without Running a Node

The larger part of staked ETH sits with exchanges, staking services and liquid staking protocols, only a small part on private machines. If that describes you, the client update is not your job. Two things are still worth doing.

Check whether your provider keeps a status page or an announcement channel, and whether anything about Glamsterdam appears there. Providers that handle network upgrades silently give you no way to make sense of an outage.

Second, check how your provider handles maintenance windows. Around a main network fork many services suspend deposits and withdrawals for a few hours. That is customary and mostly harmless, but it is a poor moment for a planned transaction.

What the Upgrade Means for the ETH in Your Wallet, and What It Does Not

Here is the reassurance that belongs with this subject: if you simply hold ETH, in a wallet of your own or with a provider, Glamsterdam requires nothing of you. There is no swap, no migration, no deadline that costs you money. Your addresses stay the same, your balances stay the same.

An upgrade of this kind changes the rules by which the network builds and checks blocks. It does not change the holdings in your wallet. A hardware wallet needs no particular preparation for it either; which devices cover which functions we have set side by side in our hardware wallet comparison.

Caution is warranted elsewhere. Network upgrades are a proven occasion for attempted fraud. Around a fork, messages regularly appear urging a supposed migration or asking for a seed phrase. No genuine protocol change ever demands that. Anyone asking you to enter your recovery words wants your money.

Assessment: How to Tell Whether the Date Holds

Three signals over the coming weeks say more than any headline.

The first is Devnet-11 from September 14. If it runs for several days without breaking, October 6 is plausible. If finalisation breaks off again, the date moves.

The second is the releases from the client teams. Once several clients publish stable versions with Sepolia support within a few days of each other, the fork is genuinely prepared. That sequence never runs the other way round.

The third is the roadmap page itself. It carries the next milestone with its date, and the entry is maintained. We described the original decision in August in our piece on the Glamsterdam date for Sepolia; the reservation from September is not yet included there. The minutes of the developer calls are summarised by Christine Kim in her notes on the 245th execution layer developer call.

Glamsterdam Date and Client Update: What to Take Away

1. Note both of your client versions today and subscribe to their release notifications. There is nothing to install: on September 11 not one of the ten clients checked carried a stable version with any Glamsterdam reference. If you would rather not run a node at all, our comparison of staking platforms is the quicker route.
2. Watch Devnet-11 from September 14 rather than the calendar. The Sepolia date of October 6 hangs on that test, and for the main network there is in any case only the unconfirmed indication of the fourth quarter of 2026. If you hold your own keys along the way, our hardware wallet comparison helps with choosing a device.
3. Treat every prompt to migrate as attempted fraud. A protocol upgrade never asks for your seed phrase. Keep your records clean instead, so that staking rewards stay traceable; suitable tools are in our comparison of crypto tax tools.

(As of September 11, 2026. This article is not investment advice. Prices and fee structures change; check the terms with the provider before you buy.)