Geth v1.17.10 does not fork Ethereum mainnet. The go-ethereum team shipped the tag on 23 October 2026 at 12:07 UTC to cancel the Amsterdam execution-layer activation on the Hoodi testnet, previously set for 26 October 2026 at 17:42:48 UTC. Operators still running v1.17.9 keep that clock. Everyone else can ignore the calendar.
That is the whole claim. Consensus clients already published Hoodi postponement builds. Execution-layer software was the remaining gap. Geth is the client a lot of Hoodi nodes still run, so the gap mattered.
Ether itself is not the story. Our CoinGecko snapshot at has ETH at $2,384.81, down 0.01% over 24 hours, with Bitcoin at $81,099 (−0.20%) and Solana at $102.73 (−0.80%). Global market cap sits near $2.73 trillion and Bitcoin dominance near 59.51%. Quiet tape. Loud client notes.
What happened
The Geth v1.17.10 release notes (Soil Reducer) say this is a quick release that postpones the Amsterdam fork on Hoodi. Hoodi node operators on v1.17.9 must upgrade before 26 October 2026, 17:42:48 UTC. A new date will be set in a future release. Other networks are unaffected.
Two days earlier, Geth v1.17.9 (21 October 2026, 15:08 UTC) had done the opposite. It enabled Amsterdam on Hoodi at block timestamp 1793036568, which is that same 26 October 17:42:48 UTC instant. If you leave a Hoodi Geth process on v1.17.9 through Monday afternoon, it will still try to activate. v1.17.10 is the cancel button for that timestamp.
The Ethereum Foundation Protocol Cluster posted the policy change on 22 October 2026: Glamsterdam on Hoodi postponed. Core developers delayed the upgrade on ACDE #248 to look harder at how EIP-8037 interacts with mainnet smart contracts. The delay removes the Amsterdam execution-layer activation, the Gloas consensus-layer activation, and the Hoodi gas-limit step that would have moved the testnet to 200 million gas at the same epoch. The Foundation says this rolls back configuration for a future upgrade. It does not revert Hoodi chain history. No replacement epoch or timestamp is in the updated config.
The same Foundation post now lists go-ethereum v1.17.10 in its execution-layer Hoodi table. Consensus-layer rows already named Lighthouse v8.3.2, Nimbus v26.10.2, Prysm v7.3.1, Teku 26.10.2, and Lodestar v1.50.1. We covered the Lighthouse consensus build yesterday. Today’s Geth tag is the execution-layer counterpart those notes were waiting on.
Context
Hoodi is a public Ethereum testnet. Glamsterdam is the upgrade name that pairs Amsterdam on the execution layer with Gloas on the consensus layer. Testnets exist so client teams can break things before mainnet users notice. When a testnet fork is pulled this close to the scheduled slot, the boring reason is usually: a client still had the old timestamp compiled in.
That is why v1.17.9 and v1.17.10 have to be read as a pair. v1.17.9 was the “turn it on” binary. v1.17.10 is the “turn it off” binary. The Foundation was explicit: do not treat older client tables as proof that those versions include the postponement. Follow the notes that say the 26 October activation was removed.
Mainnet users do not have a Geth homework assignment here. The Foundation says the announcement does not set a Glamsterdam mainnet date. ETH holders do not need to move funds because a testnet clock changed. Traders who use Hoodi to dry-run contracts, including anything sensitive to EIP-8037 gas repricing, do have homework: do not plan a 26 October Hoodi rehearsal around an activation that is no longer scheduled.
Fee and TVL prints still belong in an Ethereum note, because they show what mainnet is actually doing while the testnet calendar slips. DefiLlama’s Ethereum fee overview, pulled at from api.llama.fi/overview/fees/ethereum, shows about $16.17 million in 24-hour fee revenue, up 60.22% versus the prior 24 hours. The daily chart print for the UTC day starting 23 October 2026 is about $13.06 million. Ethereum chain TVL on the same pull is about $49.62 billion. Show me the fee, not the roadmap. The fee is live. The Hoodi fork is not.
We already wrote that Lighthouse v8.3.2 turns Hoodi Gloas off. Consensus nodes that patched and execution nodes that did not would have met at 17:42 UTC on 26 October with different fork views. That is how you manufacture a messy testnet split without meaning to. Geth v1.17.10 is how Geth operators avoid being the execution side of that mismatch.
For a longer view of how Ethereum ships upgrades, the desk’s Ethereum history page still holds. Category coverage lives under Ethereum news. None of those pages change because Hoodi’s clock was pulled.
Our read
My stance: treat Geth v1.17.10 as a Hoodi-only cancel of timestamp 1793036568, not as a mainnet Glamsterdam signal, and not as optional if you still run Geth on Hoodi.
The market will try to sell this as “Ethereum delayed the upgrade.” The primary texts do not say that about mainnet. They say Hoodi is postponed, the old epoch is gone, and a replacement date is not set. That is a client-coordination event on a testnet, with a hard local deadline for one class of operators: anyone whose Geth binary still thinks 26 October 17:42:48 UTC is fork time.
I would rather over-upgrade a testnet node than debate whether “quick release” means you can wait until next month. The notes say Hoodi operators on v1.17.9 must upgrade before the old instant. I am taking that sentence as written.
Falsifiable claim: by 26 October 2026, 17:45 UTC, Geth v1.17.10 still activates Amsterdam on Hoodi at timestamp 1793036568, or go-ethereum’s v1.17.10 notes (or an official follow-up tag) restore that timestamp as the live Hoodi fork time. If either happens, this read is wrong. If Hoodi stays quiet through that minute on patched Geth, and the notes still say other networks are unaffected, the cancel held.
What to watch next
First, the old instant itself. 26 October 2026, 17:42:48 UTC is still useful as a negative check. Hoodi should not fork Amsterdam on patched Geth at that second. If explorers or client discords report an Amsterdam activation on Hoodi at that timestamp after v1.17.10, the postpone failed in production.
Second, the replacement schedule. The Foundation says it will announce a new Hoodi activation later. Until that post names an epoch and clients ship matching tags, any “Glamsterdam is next week” thread is noise. Watch the same Foundation page; it is where they are appending client rows.
Third, mixed-client Hoodi. Consensus patches are already out. Execution patches are landing on different clocks (Geth v1.17.10 at 12:07 UTC on 23 October; Erigon also published a Hoodi unschedule later that day). A leftover v1.17.9 Geth next to a patched Lighthouse is the failure mode that still has a deadline.
Fourth, EIP-8037 work. The delay exists so developers can keep studying how that gas-repricing EIP hits mainnet contracts. That investigation can change Hoodi’s next date, and it can change what gets tested. It still does not, by itself, set mainnet. Keep mainnet fee prints in view so a testnet calendar slip does not get mistaken for a usage crash.
If you run Hoodi Geth, upgrade before Monday’s old timestamp. If you hold ETH on mainnet, this tag is not your sell or buy cue. The range on ether tonight is quiet. The client note is the only clock that actually moved.