Ethereum price prediction: Glamsterdam forks Hoodi on October 26 with a 200 million gas limit, what to watch

Here is what is documented: the activation table in EIP-7773, the meta EIP covering Glamsterdam, lists epoch 132,352 and timestamp 1793036568 for the Hoodi testnet. That is Monday, October 26, 2026, at 17:42:48 UTC. The same epoch raises the gas limit on Hoodi to 200 million. The mainnet row in that same table is empty; it will only be filled once the client teams set a date. Our assessment follows further down: for the price, the fork on Hoodi is an intermediate step and not an event.

Ether trades at $2,406.36 on Thursday morning, October 22, which is 1.96 percent below its level 24 hours earlier (CoinGecko, retrieved at 08:43 UTC). The daily high was $2,518.86. Where the price can point on a horizon of months is covered by our Ethereum price prediction. This piece sticks to what can be verified up to October 26.

Glamsterdam forks Hoodi on October 26 at 17:42:48 UTC

Hoodi is one of Ethereum's public testnets. It replicates the mainnet with real validators, only with worthless tokens, and serves as the final rehearsal before a fork on mainnet. A fork, also called a hard fork, is a rule change that every node has to adopt at the same moment. Anyone still running the old software drops out of consensus and sees a different blockchain from the rest of the network from the fork epoch onwards.

The Ethereum Foundation announced the Hoodi date on October 16. The numbers behind it are public in EIP-7773: epoch 132,352, slot 4,235,264, Unix timestamp 1793036568. To check the arithmetic you need no tool from us, only the table in the hard fork meta for Glamsterdam. Beside it sits the activation on Sepolia that has already run, at epoch 353,024, on October 6, 2026 at 13:53:36 UTC.

For holders, nothing follows from this for now. A testnet fork moves no real balances, triggers no transaction and creates no new token. If you read a report over the coming days about a supposed Glamsterdam token or a distribution to ETH holders, you are reading a fake: none of the 18 proposals that EIP-7773 lists under "Scheduled for Inclusion" contains a payout.

Gas limit of 200 million: Hoodi catches up with Sepolia

Gas is the unit of account in which Ethereum measures the effort of a transaction. A block's gas limit caps how much work fits into a block. Raise the limit and more transactions fit into the same block, and at unchanged demand the price per transaction falls. On Hoodi that limit jumps to 200 million with the fork epoch, reaching the value Sepolia has been running since October 6.

The jump is not automatically the jump the number promises. Sepolia showed on October 6 that a higher protocol limit on its own achieves little as long as the validator software builds smaller blocks. There, Prysm version 7.2.1 was needed at short notice before validators produced blocks with 200 million gas at all instead of staying at the old ceiling. A limit is a ceiling, not an instruction.

For mainnet later, that means this in practice: more room per block only pushes fees down if demand stays the same and the client teams actually use the higher ceiling. If you want to move ETH from an exchange into your own custody today, work with current fees rather than those after a fork without a date. Which trading venues pass on which network fee at withdrawal is shown by our crypto exchange comparison.

EIP-7732 moves the block auction into the consensus protocol

The largest piece of Glamsterdam is EIP-7732, enshrined proposer-builder separation, or ePBS. Today the division of labour between the validator that proposes a block and the specialised builder that assembles it runs through software outside the protocol. The validator has to trust the builder and a relay to deliver the block it promised.

EIP-7732 turns that trust into a protocol rule. A dedicated committee, the payload timeliness committee, attests whether the builder disclosed its payload including blob data in time. If that attestation comes out negative, the protocol's rule applies instead of an arrangement between two firms. The full text is held by EIP-7732 in the official register.

Three things about it are tangible for investors. First, dependence on a handful of relays falls, which makes censoring individual transactions harder. Second, the way revenue from block ordering reaches the validator changes. Third, every node operator gets new software, and new software is the most common reason validators collect penalties on a fork day.

EIP-7928 records every storage slot a block touches

The second large piece is EIP-7928, block-level access lists. Such a list records which accounts and which storage slots a block touched and which state changes resulted. The benefit is technical: anyone verifying a block can load the necessary data in advance and process several transactions side by side instead of working through them one after another.

Alongside it, EIP-7773 holds a further 16 proposals that get less attention and still show up in everyday use. EIP-7708 makes every ETH transfer emit a log, which simplifies accounting for wallets and tax tools. EIP-7954 raises the maximum contract size. EIP-8061 increases the churn limit for validator exits and consolidations, which is precisely the figure that decides the length of the staking queue. EIP-8045 bars slashed validators from proposing.

The meta EIP's status reads "Review". In the language of the register that means the list is settled enough for a testnet and open enough that something can still change before mainnet.

The Sepolia date slipped by eight days

This is the point most reports on the fork leave out, and it can be measured against our own archive. On August 21 we reported under the headline Glamsterdam has a date: Ethereum forks Sepolia on September 28, 2026 on the Sepolia date announced at the time, including the time of 14:44:48 UTC. The activation on Sepolia then ran on October 6 at 13:53:36 UTC.

Eight days therefore lay between the date first announced and the actual one. That is no mishap but the normal course of an upgrade that depends on stable devnets. For expectations about mainnet it remains the most important number of the day all the same: anyone deriving a November mainnet date from October 26 is counting on a reliability this process has not shown so far.

ETH at $2,406, $37 below its level of October 20

In our Ethereum price prediction of October 20 ETH stood at $2,443, while 1,155,535 ETH were waiting for a staking slot and the wait came to 20 days. Since then the price has moved down by roughly $37, or 1.5 percent. The $2,500 mark that was in play at the time did not hold: the daily high of $2,518.86 was above it, but the price fell back over the course of the day.

Three figures frame the situation. To the upside, the daily high of $2,518.86 is the next level that can be evidenced, because that is where sellers last appeared. To the downside sits the level of October 19 at $2,396, where the price turned up once before. Between them stands October 20 at $2,443. Against Bitcoin, one ETH currently costs around 0.0294 BTC, with Bitcoin at $81,880 at the same retrieval time.

There is also the fund side. US spot ether funds recorded combined net outflows of $129.4 million on October 19 and 20. A testnet fork does nothing against flows of that kind; it changes nothing about the money going into or out of the product each day.

Staking, client versions and where holders stand

If you stake ETH yourself, meaning you run your own validator, Glamsterdam affects you directly once mainnet's turn comes. The same mechanics apply before every fork: the execution client and the consensus client have to be raised to a version that knows the new rules. Miss that and you miss attestations and lose yield for the day.

Two points count with a staking provider

For investors who stake through an exchange, a pool or a liquid staking service, the provider handles the update. Asking is still worthwhile, on two things: whether the provider pauses deposits and withdrawals during the mainnet fork, and how it books the revenue from block ordering after ePBS. Neither is usually in the standard terms, and both can be settled before the date. Which services report which yields and which lock-up periods is shown by our overview of staking platforms.

If you only hold ETH

For pure holders, a fork changes nothing about the amount in the wallet. There is no action, no confirmation and no migration required of the user. Any demand to "synchronise" a wallet before an upgrade, to enter a seed phrase or to "migrate" tokens is an attempted fraud. Experience shows this trick clusters in the days before an announced fork date.

Tax: staking income counts in the year it is received

For investors in Germany, staking carries a tax question that the upgrade does not make go away. Income from staking is recorded at the moment it is received, at its value on that day, and not only at a later sale. If you receive income in October, document the price on the day it arrived, whether or not you keep the coins.

Selling the coins themselves is a separate matter. How long the holding period applies to staked holdings, and what has shifted on that recently, is tracked by our page on the status of the crypto holding period. The fork changes nothing about the period: a protocol change is not a disposal, and a testnet fork creates no receipt of income.

Our assessment of the situation before the fork

From the newsroom's point of view, October 26 is a technical date and not a price date. Three pieces of evidence support that. The mainnet row in EIP-7773 is empty. The Sepolia date slipped by eight days between announcement and activation. And fund flows pointed the other way on October 19 and 20, with $129.4 million leaving.

What argues against it is that Glamsterdam carries two changes, ePBS and the access lists, that touch the fee market and validator revenue structurally. If the test on Hoodi comes out clean, a mainnet announcement in the fourth quarter is plausible, and a date would be the first real occasion this upgrade delivers to the market. No buy recommendation follows from that; crypto investments can fall to a total loss.

Glamsterdam fork: without a mainnet date, $2,518 stays the hurdle

The fork on October 26 is a test that leaves little in the price as long as the next row of the table stays empty. Three steps put it in order:

1. Note October 26, 17:42:48 UTC, as a testnet date and look at the mainnet row in EIP-7773 again afterwards. Only a date there is an occasion. If you want to build or reduce holdings in the meantime, the fee models are in the exchange comparison.
2. Ask your staking provider two things: whether it pauses deposits and withdrawals during the mainnet fork, and how it reports revenue after ePBS. The terms are in the staking overview.
3. Keep an eye on the levels: $2,518.86 to the upside, $2,396 to the downside. Between them the fork supplies no impulse of its own.

(As of October 22, 2026. This article is not investment advice. Prices and fee structures change; check the terms with the provider before you buy.)