The Ethereum Foundation Protocol Cluster scheduled the Glamsterdam network upgrade on the Hoodi testnet for epoch 132,352 (slot 4,235,264), which is 17:42:48 UTC on October 26, 2026. The same post leaves mainnet as TBD. Ethereum (ETH) holders do not need to do anything because of this announcement.
That is a client-opt-in clock for Hoodi operators, not a surprise fork of Ethereum mainnet. The Foundation even answers the holder question in the negative: this post covers Hoodi only, and a later announcement will cover mainnet.
The CoinGecko market snapshot on this site put ETH at $2,419.54 (24h −0.85%) and Bitcoin (BTC) at $81,517 at 2026-10-17T00:05:13+00:00. That tape is context. It is not evidence that Glamsterdam already shipped.
What happened
On October 16, 2026, the Protocol Cluster published Glamsterdam Hoodi Testnet Announcement on the Ethereum Foundation blog. Sepolia already activated on October 6, 2026, at epoch 353,024 (slot 11,296,768, 13:53:36 UTC, Unix 1791294816). Hoodi is next: epoch 132,352, Unix timestamp 1793036568. At that Hoodi epoch the gas limit schedule also jumps to 200 million, matching Sepolia.
Glamsterdam is the paired name for Amsterdam on the execution layer and Gloas on the consensus layer. The post’s headliners are enshrined proposer-builder separation (ePBS, EIP-7732) and block-level access lists (BALs, EIP-7928). EIP-7732 puts the builder-proposer split into consensus: a proposer includes a builder’s commitment, the builder later reveals the payload, and a payload timeliness committee attests that the reveal and blob data arrived on time. EIP-7928 records which accounts and storage a block touched, so clients can read disk and check transactions in parallel.
The Foundation also flags gas accounting changes (including EIP-8037 on state-creation cost and EIP-8038 on state-access cost) and tells application developers to test contracts that assume a leftover gas stipend. That is homework for Hoodi and Sepolia, not a mainnet countdown.
The same announcement ships a live client table. Consensus-layer rows filled as of the post: Lodestar 1.50.0, Nimbus 26.10.1, Prysm v7.3.0, Teku 26.10.1. Grandine and Lighthouse were still blank. Execution-layer rows filled: Besu 26.10.0, Ethrex 30.0.0-rc.2, Nethermind 2.2.0. Erigon, go-ethereum, and Reth were still blank. The table says only use a release whose notes confirm the scheduled Hoodi activation, and that rows will be updated as more clients land.
Those versions match what the client repos tagged the same day. Prysm v7.3.0 was published at 2026-10-16T21:08:47Z. Its notes schedule Gloas on Hoodi at epoch 132352 (2026-10-26 17:42:48 UTC) with a 200M gas-limit schedule entry, and they tell Hoodi operators to update both beacon nodes and validator clients before then. On networks that have already forked into Gloas, the V1 GetExecutionPayloadEnvelope and PublishExecutionPayloadEnvelope gRPC endpoints return FAILED_PRECONDITION, so an old validator client cannot sign envelopes against this beacon node. Pre-Gloas networks, including mainnet today, are not in that pairing break. Prysm still recommends that mainnet operators update beacon nodes for the fixes in this tag. That is a patch suggestion, not a mainnet fork epoch.
Nethermind 2.2.0 was published at 2026-10-16T16:26:20Z. Its notes put Amsterdam (Glamsterdam) on Hoodi at timestamp 1793036568, the same Monday 17:42:48 UTC clock, and say Hoodi nodes must run 2.2.0 or later, with a consensus client that also supports the fork. For most nodes it describes an in-place upgrade with no resync.
Context
We already treated Lodestar 1.50.0 as a Hoodi Gloas clock, not a mainnet fire. This Foundation post is the network-level version of that clock: same epoch, same UTC second, now with a public activation table and an explicit “ETH holders do nothing” FAQ. I am not updating that view. I am tightening it. Lodestar was one client. Hoodi is now an EF-scheduled testnet upgrade with four consensus clients and three execution clients named in the table.
The blanks in the table matter as much as the filled rows. Hoodi operators still waiting on Geth, Reth, Erigon, Lighthouse, or Grandine do not have a confirmed row yet. Ethrex is listed as 30.0.0-rc.2, a release candidate, not a final tag. The Foundation tells operators to wait for notes that explicitly confirm Hoodi, not to guess from a marketing blog.
Fee prints can fool people into reading a testnet clock as live throughput. DefiLlama’s Ethereum fee overview (api.llama.fi/overview/fees/ethereum), read at 2026-10-17T00:13:31Z, showed about $16.70 million in Ethereum fees over the prior 24 hours, up 48.97% versus the previous 24-hour window. That is a usage print on today’s chain. It is not Hoodi, and it is not Glamsterdam activating. Show me the fee, sure. Then show me which network those fees cleared on.
Ethereum still uses proof of stake. Upgrades do not “just happen” to holders. Node operators opt in by installing software that speaks the new rules. A node that skips the client cut cannot follow Hoodi after 17:42:48 UTC on October 26. That is why the Foundation repeats the boring instruction: update both the execution client and the consensus client, and if you validate, update the beacon node and the validator client together.
For a longer arc of how Ethereum has shipped forks before, the desk’s Ethereum history page is the hub. For why a 200 million gas limit is a client-and-builder problem as much as a trader headline, see gas fees. Related Hoodi notes from this month sit on our Lodestar 1.50.0 reaction and in Ethereum news.
Our read
Stance: treat October 16’s Foundation post as a Hoodi operator deadline, not as a hidden mainnet date. Spot ETH around $2,400 does not get a Glamsterdam premium from this document. The document says the opposite of urgency for holders.
The pairing risk is real on Hoodi, and it is mechanical. Prysm’s Gloas-network gRPC break is exactly the kind of “I upgraded one binary” failure that takes a validator offline at a fork. Nethermind’s 2.2.0 notes push the same pairing: execution and consensus both have to know the timestamp. If you run Hoodi, the clock is 2026-10-26 17:42:48 UTC, not “sometime this quarter.”
The unfinished client table is also the story. A Foundation announcement with empty Geth and Lighthouse cells is not a completed rollout. It is a schedule plus the clients that have already tagged. Operators on unlisted clients should watch the table, not invent a version from Discord.
Falsifiable claim: Hoodi activates Glamsterdam at epoch 132,352 / Unix 1793036568 (17:42:48 UTC on October 26, 2026), matching the Protocol Cluster post. If Hoodi does not activate at that epoch and timestamp, or if the Ethereum Foundation Protocol Cluster republishes a different Hoodi epoch or UTC time before that instant, this read is wrong.
What to watch next
Watch the Foundation client tables fill. The first useful print is a non-RC go-ethereum, Reth, Erigon, Lighthouse, or Grandine row whose release notes name Hoodi epoch 132,352. Until those notes exist, those stacks are not confirmed on this clock.
Watch Prysm and Nethermind follow-through on Hoodi, not on Twitter. If either project yanks v7.3.0 or 2.2.0, or ships a replacement tag that changes the Hoodi epoch, the pairing instructions change with it.
Watch for a separate mainnet announcement. This post is explicit: mainnet date TBD, holders do nothing, operators on mainnet do nothing because of Hoodi. A later Protocol Cluster post with a mainnet epoch would be new evidence. This one is not that post.
Keep one eye on ETH fees only as a usage check, timestamped, after you know which chain produced them. A loud fee day in October 2026 still is not Glamsterdam on mainnet.